Content and licensing view original scan buy a printed copy

Geology of the Ayr district. Sheet description of the British Geological Survey, 1:50 000 series sheet 14W and part of 13 Ayr (Scotland)

Bibliographical reference: Smith, R A, and Monaghan, A. 2013. Geology of Ayr district. Sheet description of the British Geological Survey, 1:50 000 Series Sheet 14W and part of 13 Ayr (Scotland).

British Geological Survey

Authors: R A Smith and A A Monaghan

Contributors: S L B Arkley, J D Floyd, E R Phillips, M T Dean, M Williams, P Williamson and M Stephenson

Copyright in materials derived from the British Geological Survey's work is owned by the Natural Environment Research Council (NERC) and /or the authority that commissioned the work. You may not copy or adapt this publication without first obtaining permission. Contact the BGS Intellectual Property Rights Section, British Geological Survey, Keyworth, email ipr@bgs.ac.uk. You may quote extracts of a reasonable length without prior permission, provided a full acknowledgement is given of the source of the extract.

Your use of any information provided by the British Geological Survey (BGS) is at your own risk. Neither BGS nor the Natural Environment Research Council gives any warranty, condition or representation as to the quality, accuracy or completeness of the information or its suitability for any use or purpose. All implied conditions relating to the quality or suitability of the information, and all liabilities arising from the supply of the information (including any liability arising in negligence) are excluded to the fullest extent permitted by law.

The National Grid and other Ordnance Survey data © Crown Copyright and database rights 2013. Ordnance Survey Licence No. 100021290.

(Front cover) Culzean Castle (National Trust for Scotland) partly constructed of local Siluro-Devonian Swanshaw Sandstone Formation. (P601213).

(Rear cover)

Acknowledgements

The accompanying Bedrock and Superficial Deposits and Simplified Bedrock 1:50 000 geological maps were compiled from the component 1:10 000 maps produced by S L B Arkley, J D Floyd, A A Monaghan and R A Smith.

In addition to the above, published work on the Ordovician to Silurian inliers has utilised the published work of L R M Cocks, P T Toghill and D A T Harper.

Many individuals in the BGS have contributed to the Midland Valley Project. This assistance has been received at all stages of the study. In addition to the collection of data, many individuals have freely given their advice. Key staff contributed to the project and we would particularly like to thank M C Akhurst, E R Phillips, M Stephenson, M Williams, M T Dean, S Molyneux, N S Jones, N Breward, P Williamson, P Stone and other participants at Midland Valley Project meetings.

Contributions to the Sheet Description were provided by E R Phillips (Petrology), P Williamson (Concealed Geology) and M Williams (Lower Palaeozoic stratigraphical palaeontology). It was edited by D Stephenson, M C Akhurst and M A E Browne. The Upper Devonian, Carboniferous and lower Permian palaeontology was edited by M T Dean. A W A Rushton (BGS & Natural History Museum, London) provided taxonomic data.

Notes

The term 'Ayr district', or simply 'district', is used in this description to refer to the area covered by Sheet 14W and part of 13 (Ayr) of the 1:50 000 geological map of Scotland. There are separate Bedrock, and Superficial Deposits and Simplified Bedrock editions of the sheet.

British National Grid references in the text are given in the form [NS 295 096] as the whole district lies within 100 km square NS. The few grid references outwith the NS grid square, are prefixed NX, NR, NT or NO, depending on which grid square the locality referred to lies within.

The codes in round brackets after lithostratigraphical names are the same as those used on the 1:50 000 geological maps.

Numbers referred to in plate captions are BGS Imagebase registration numbers. Numbers preceded by S or N refer to the Scottish BGS sliced rock/thin section collection held at BGS, Edinburgh. Numbers preceded by MPA refer to palynological samples held by BGS, Keyworth. Numbers such as (NS31NE/9) refer to borehole record numbers registered at BGS, Edinburgh. Numbers preceded by SE refer to site investigation reports held by BGS, Edinburgh.

BGS products and services relating to the district are listed in the Information Sources.

Summary

This account of the geology of the Ayr district, in the south-west of the Midland Valley of Scotland, covers the area from around Monkton in the north, to Dailly in the south. The description of the bedrock geology includes field observations recorded up to 2002, together with earlier published and unpublished work on the area. The Quaternary section updates field records and reviews earlier published work.

The district covers the growing town and administrative centre of Ayr and its neighbour Prestwick, extending south into rural farmland and scenic hill country. Farther south is the market town of Maybole and the former coalfield around Dailly. It also covers some dramatic coastal scenery and sweeping bays, with tourist attractions stretching from the golf courses south of Troon, past the ruined castles of Greenan and Dunure, to Culzean Castle overlooking Culzean Bay and onto Maidens and Turnberry.

The bulk of the rocks are sedimentary and Palaeozoic in age, with a succession extending from Ordovician to early Permian, punctuated by several volcanic and intrusive igneous episodes.

In Mid Ordovician times, about 470 Ma ago, the district lay close to the southern edge of the Laurentian continent. Oceanic crust was thrust up when a volcanic arc collided with the microcontinental segment that formed the Midland Valley Terrane as it docked against Laurentia. Upper Ordovician to lower Silurian sediments were deposited in a forearc basin on the southern margin of this terrane and were deformed as the Southern Uplands accretionary prism was pushed up from the south during the Caledonian Orogeny. The Lanark Group was later deposited, in Siluro-Devonian times, in a sequence of sandstones and conglomerates lain down in a semi-arid environment, prior to the Early Devonian calc-alkaline magmatic event which produced shallow intrusions and eruptions of predominantly basaltic andesite. By Mid Devonian times the Lanark Group had been weakly deformed and uplifted as a far-field effect of the late stage Acadian deformation event.

The Upper Devonian to Carboniferous strata were then deposited unconformably in a widespread but fractured basin, with inherited Caledonian structures still responding to weak left lateral and subsequent right lateral displacements. Fluviatile red-bed sandstones (Stratheden Group) were deposited in the Upper Devonian followed by semi-arid terrestrial sandstones with pedogenic carbonates (Kinnesswood Formation) in the early Carboniferous. The overlying mudstones, thin-bedded limestones and sandstones (Ballagan Formation) were deposited in a marginal marine and/or restricted saline lagoonal environments, followed by a return to semi-arid fluvial sandstone deposition (Clyde Sandstone Formation).

Visean volcanic activity, prevalent farther north in the Midland Valley, is represented in this district by the Heads of Ayr volcanic vent and accompanying Greenan Castle pyroclastic deposit. After a brief hiatus, lower Carboniferous deposition continued with fluviatile to fluviodeltaic sands, muds and peats (Lawmuir Formation). The first lower Carboniferous marine limestone transgressions occurred during deposition of the Lower Limestone Formation. This formation and the succeeding Limestone Coal Formation are fairly well developed in the Dailly Sub-basin, but around Ayr, a prevailing palaeohigh prevented much of the lower Carboniferous succession from accumulating or being preserved. Later fluviatile sandy deposition (Passage Formation) was interrupted by the outpouring of basaltic lavas (Troon Volcanic Member).

Later in the upper Carboniferous, in a wetter, tropical climate, the Scottish Coal Measures Group was deposited cyclically on broad deltaic flats. Towards the close of the Carboniferous Period, the Variscan Orogeny tightened and faulted the coal basins which had undergone variable growth over the inherited Caledonoid and strike-slip structures. In early Permian times rifting on north-westerly trends accompanied a pulse of basic alkaline magmatism, including the eruption of the Mauchline Volcanic Formation.

In Palaeogene times, as the Atlantic Ocean opened and during activity of the North Atlantic Igneous Superprovince, the Prestwick-Mauchline Sill was intruded. Subsequently alkali and tholeiitic mafic dykes filled fractures along mainly north-westerly trends associated with the igneous centres at Arran and Mull.

In Quaternary times, that is the last 2.6 million years, several glaciations have affected the district but only the last major Pleistocene (Devensian) glaciation (about 30 000 years ago) has left its mark. At that time an ice-sheet flowing south-eastwards, onshore from the Firth of Clyde, merged with ice flowing north from the Southern Uplands. Beneath this ice a layer of glacial till was widely spread on the low ground; locally moulded into drumlins. The higher ground was commonly scoured of superficial deposits. As the ice melted away or stagnated, hummocky glacial deposits were left in the Maybole region and reworked sand and gravel was deposited in eskers or in small spreads beside watercourses.

Holocene times (postglacial) began about 11 700 years ago, when the Gulf Stream ocean current was re-established, providing a warmer climate over the British Isles. Rivers and streams actively reworked superficial sediment to create floodplains, local terraces and alluvial fans. Vegetation recolonised the land, but peat has been preserved in this district only in poorly drained upland hollows. Along the coast, postglacial isostatic rebound is evident in the development of raised beaches and rock platforms. More recently belts of blown sand have been deposited by prevailing westerly winds inland of the high water mark, mainly in the Prestwick and Turnberry areas.

Chapter 1 Introduction

The Ayr district lies on the south-western edge of the Midland Valley of Scotland. The area covered by this sheet description comprises the 1:50 000 Geological Sheet 14W and the adjacent part of Sheet 13, together with the offshore geology. It includes the western part of the Ayrshire Coalfield and the hills to the south formed by mainly Siluro-Devonian sedimentary and volcanic rocks, with inliers of Ordovician to Silurian rocks near Craighead and Straiton. The formerly productive Dailly Coalfield lies in the south of the district, and lower Carboniferous strata occupy much of the intervening low ground towards Straiton. Away from the thriving county town and port of Ayr and its northern neighbour, Prestwick, the district is largely rural. Recreational venues along the coast south of Ayr lie mainly beside sandy bays separated by headlands formed by resistant igneous rocks, which are commonly well exposed. South of Ayr, the remains of Greenan Castle are perched on a crag of lower Carboniferous pyroclastic rock, and a volcanic vent of similar age forms the landmark Heads of Ayr. The Heads of Ayr Vent is a Geological Conservation Review (GCR) site (Williamson and Monaghan, 2003).

Culzean Castle, the prime attraction of the National Trust for Scotland, is situated above prominent cliffs formed of Carrick Volcanic Formation, and Culzean Harbour is a GCR site (Durant, 1999). Scenic coastal exposures extend from the Little Craigs, south of Troon, to Dipple, south of Turnberry. The best coastal geological sections include the Maidens to Doonfoot SSSI, and both the Port Schuhan to Dunure section and the Turnberry Lighthouse to Port Murray sections are designated GCR sites (Durant, 1999).

Inland, the Carrick Volcanic Formation forms the broad ridge culminating in Brown Carrick Hill at a height of 287 m. South-east of Straiton the same formation forms a series of prominent terraces rising to 354 m on Kildoach Hill.

Most of the topography inland has been moulded beneath glacial ice in Pleistocene times; the harder igneous rocks, such as Mochrum Hill, remain standing out from the cover of till which is locally formed into drumlin fields with mounds mainly oriented east–west. Moundy morainic deposits and sand and gravel, laid down as the Devensian ice melted, are widespread around Maybole, extending south-east along the Water of Girvan. South of Straiton, the latter river runs within a scenic U-shaped valley north from Carrick. The rivers Ayr and Doon form scenic valleys meandering westwards to the sea near Ayr, the beauty of the latter has been extolled since Robert Burns wrote his poetry in the 18th century.

Along the coast, the raised beaches and more recent blown sand form a platform on which most of Ayr and Prestwick developed, including the international airport. Ayr racecourse lies on the upper (late Devensian) raised beach to the east of the town. The sand dunes have attracted the attention of golfers with links courses stretching from south of Troon to Turnberry. Several holiday camps are situated beside the sandy beaches along the coast. The varied and generally well-drained soils of south Ayrshire have traditionally favoured potato growing and livestock farming, and Scotland's Rural College campus of Auchincruive is located north-east of Ayr.

Geological history

The oldest rocks in the district occur in the Lower Palaeozoic inliers of Craighead and the extension of the Girvan inlier as far east as Straiton (Figure 1). Ordovician strata exposed within the Craighead Inlier include two GCR sites and two small inliers on the southern margin of the sheet at [NS 395 010] and [NS 330 007]. The Ordovician strata comprise pillow lavas, limestones, sandstones, mudstones and conglomerates, deposited on the southern flank of the Laurentian continent. They were unconformably overlain by Silurian (Llandovery to Wenlock) conglomerates, fossiliferous mudstones and sandstones, some of which are turbidites, deposited on the Laurentian shelf. Following the closure of the Iapetus Ocean and the accretion of the Southern Uplands Terrane to the south, these deposits were deformed and uplifted during the Caledonian Orogeny.

In the semi-arid climate of late Silurian to Devonian times, the uplifted rocks were eroded and overlain by the fluviatile, and locally aeolian, beds of the Lanark Group. Early in Devonian times, widespread volcanic activity produced the andesitic basalts of the Carrick Volcanic Formation and associated shallow intrusions. Acadian uplift during the Mid Devonian was a minor event in this district, but deformed Siluro-Devonian strata and re-activated Caledonian faults, causing the uppermost Devonian and lowermost Carboniferous successions to lie unconformably on the older rocks. Upper Devonian fluviatile red sandstones are succeeded by lower Carboniferous sandstones with calcrete horizons (Kinnesswood Formation), deposited in a semi-arid climate. The succeeding mudstones, siltstones, thin limestones and sandstones (Ballagan Formation) were deposited in lagoons, locally overlain by red sandstones (Clyde Sandstone Formation). Early Carboniferous volcanism produced the Heads of Ayr vent and the associated pyroclastic deposit at Greenan Castle. Subsequent uplift and erosion during the early Carboniferous created a minor unconformity with the overlying clastic Lawmuir Formation, and later lower Carboniferous cyclic sequences began with a widespread marine transgression which deposited the Lower Limestone Formation. This formation and the rest of the succeeding Clackmannan Group strata are condensed in this district and become non-existent over a palaeohigh around Ayr. Deposition of the Passage Formation was interrupted by the outpouring of basaltic lavas forming the Troon Volcanic Formation. The fluviodeltaic cycles of the upper Carboniferous produced the Scottish Coal Measures Group deposits of mudstones, sandstones, seatearths and coals. Following the Variscan Orogeny in the late Carboniferous, a period of rifting caused the lower Permian Mauchline Volcanic Formation to be laid down unconformably on the Carboniferous strata in the Mauchline Basin, in the north-west of the district.

During Palaeogene times, magmatism allied to the North Atlantic Igneous Superprovince and rifting prior to Atlantic sea-floor spreading produced the Prestwick–Mauchline Sill Complex as well as alkali and tholeiitic mafic dykes, mainly with a north-westerly trend.

During the last major Pleistocene glaciation (Devensian), ice flowing east from the Firth of Clyde merged with ice flowing north from the Southern Uplands. Till deposited beneath the ice-sheet covers most of the low ground. As the ice-sheets melted, hummocky glacial deposits and glaciofluvial outwash sand and gravel were deposited. Postglacial rivers reworked the sediments to form alluvial floodplains, fans and terraces. Along the coast, postglacial raised beaches continued to develop in Holocene times as the land rebounded from the removal of the mass of the ice sheet. More recently, loose sand was blown inland from the shore by prevailing westerly winds. Peat is sparse in this district, but accumulated in poorly drained hollows.

History of research

The first official geological survey around Ayr was made in the 1860s by Sir Archibald Geikie, J Geikie and B N Peach at the six-inch scale. The resulting one-inch Sheet 14 (bedrock with postglacial deposits) was published in 1868 and the adjacent Maidens area, mapped by Sir Archibald Geikie, was published as one-inch Sheet 13 in 1867. Accompanying short explanations for sheets 13 and 14 were both published in 1869. A revision of the Lower Palaeozoic rocks was made by geological survey workers Peach and Horne (1899) after the stratigraphical advances of Lapworth (1882). Sheets 13 and 14 were revised by V A Eyles, J B Simpson, J E Richey and A G MacGregor in the 1920s, but only Sheet 14 was published in 1933 (as separate Solid and Drift editions). Economic memoirs covering parts of the district were published in 1930 (Area III, covering Ayr and Prestwick) and 1932 (Area IV, covering Dailly) as part of a series for the Ayrshire Coalfields. A soil texture edition of Sheet 14 was produced by the West of Scotland Agricultural College, and published by the Geological Survey in 1932. The second edition of the explanation of one-inch Sheet 14, Geology of Central Ayrshire (Eyles et al., 1949), also described parts of sheets 13, 8, 15 and 22. During the 1950s and 60s, W Mykura revised the coalfield geology in south-west Ayrshire as coal mining in the area expanded after nationalisation (Mykura, 1967). In 1978, 1:50 000 Solid, and Drift-only editions of Sheet 14W (including the adjacent part of sheet 13) were published after compilation by A Davies that incorporated the mapping of Mykura and the earlier surveyors.

The early geological research of the area around Ayr started with Sir Archibald Geikie, in the 1860s, continuing into the early part of the 20th century, including a brief account of the geology along the Carrick shore published by J Smith in 1895. The latter Ayrshire geologist was a well-known character in the district, and into the 1930s, local naturalists told of their 'racy reminiscences of geological transactions' with him (Tyrrell, 1933). Smith's interests also included the deposition of Quaternary glacial deposits (Smith, 1898b) and the occurrence of agate in Ayrshire (Smith, 1910).

The igneous geology of the district was comprehensively studied by Tyrrell (1912, 1914, 1917, 1920, 1928a, b), Scott (1915) and Eyles et al. (1929); all summarised in Eyles et al., (1949). Subsequent studies on volcanic aspects were made by Whyte (1963) and Kokelaar (1982).

The Lower Palaeozoic inlier at Craighead has long been a classic area for fossil collecting and biostratigraphy (with visits by Sedgwick and Murchison from about 1850; for detailed history see Harper, 1982; Rushton et al., 1999a, b). A comprehensive account of the stratigraphical palaeontology of the Craighead Inlier has been made by M Williams (2001). Prior to this, the initial findings by Sedgwick and Murchison were followed later in the 19th century, by the comprehensive collecting of the Geological Survey and Mrs Eliza Gray of Edinburgh together with her daughters, and active collecting and study has continued up to the present day. Many of the classic Ordovician exposures in the district are currently geological conservation sites (Rushton et al., 1999a, b) and the previous research has been reviewed in relevant sections below. The Lower Palaeozoic biostratigraphical studies have proved important because of the intermingling of shelly and graptolite faunas enabling widespread stratigraphical ties to be made. The more recent important studies, linking the Silurian inliers in the Midland Valley, were carried out by Cocks and Toghill (1973), Dorning (1982), Clarkson et al. (1998) and Molyneux et al. (2008).

Palaeontological research in the younger rocks is limited by the general lack of fossils due to adverse palaeofacies and palaeoclimatic controls. Well preserved trace fossils from Siluro-Devonian strata collected by Smith (1909) have most recently been studied by Walker (1985). The Upper Devonian fossil fish collected from Bracken Bay are described by Miles (1968). Records of the macropalaeontology up to the 1940s is summarised in Eyles et al. (1949), and the subsequent work in the lower Carboniferous (e.g. Wilson, 1989) and upper Carboniferous (e.g. Calver and Wilson in Mykura, 1967; Wilson, 1967) is reviewed by Dean (2002). Early work on the Tournaisian/Mississippian spore floras from the Ballagan Formation, south of Ayr, was published by Lele and Provan (1962), Sullivan (1968) and Neves et al. (1973). The stratigraphical palaeontology of the Westphalian has been reviewed and correlated across Ayrshire (Brand, 1983). A probable early Permian flora collected from a sedimentary intercalation in the Mauchline Volcanic Formation on the River Ayr, near Stairhill, was studied by Wagner (1966; 1983).

As a result of this resurvey, detailed studies of lower Carboniferous palynomorph and ostracod assemblages within the district have been published by Stephenson et al. (2002), Williams et al. (2005, 2006). Fluvial and aeolian deposition in the Siluro-Devonian strata has also been described (Smith et al., 2006).

Chapter 2 Ordovician

In the southern part of the Midland Valley Terrane, Ordovician rocks are exposed in the Ballantrae Ophiolite Complex and the Girvan Lower Palaeozoic inliers. In the Ayr district, they are exposed within the Craighead Inlier and two small inliers on the southern margin of the sheet, one near Genoch at [NS 395 010] and the other south-east of Barony Moss at [NS 330 007], which correlate with successions north of the Stinchar Valley (Floyd, 1999).

The Craighead Inlier includes a lensoid area of Ordovician rocks in the core of an anticline, the south-western extremity of which extends on to the adjacent Carrick Sheet (Floyd, 1999).

This lensoid area is unconformably overlain by the Silurian succession (Llandovery–Wenlock), and the oldest rocks (Table 1) at its core may correlate with the Ballantrae Ophiolite Complex or somewhat younger basaltic lavas (Floyd, 1999). In any case, they formed a basement over which a Caradocian cover sequence was deposited on the Laurentian side of the Iapetus Ocean. This cover is an internationally important conservation site for upper Ordovician fossils. The history of palaeontological collecting has been given by Clarkson (1985) and its biostratigraphical importance by Williams (2001). A summary is given below.

Sedgwick (1851) refers to a 'North Girvan group' of post-Llandeilo–pre-'Old Red Sandstone age', and mentions fossils found in the more earthy beds of the limestone at Craighead and its association with a 'great mass of trap'. Murchison (1851) then applied his knowledge of 'Silurian' rocks to the Craighead inlier. The first definitive work on the stratigraphy of the inlier was published by Lapworth (1882) after his pioneering work on the Moffat Shales in the Southern Uplands and followed up by geological survey workers Peach and Horne (1899, pp.462–463, 522–533). Subsequent revisions were made by Jones (1928), Lamont (1935), Freshney (1959), Williams (1962), Cocks and Toghill (1973), Ingham, (1978), Harper (1981; 1982; 1984) and Floyd et al., (1999). A synopsis of published work on the stratigraphical palaeontology of the inlier was made by Williams (2001) and geological excursion guides produced by Ingham (1992a, b, c). Bluck (2002) described the basin in which the Craighead beds were deposited as the youngest and most northerly of the Ordovician age basins on the south side of the Midland Valley. It contains a high proportion of shelf deposits with sheets of storm-generated sediment. Contemporaneous tectonic activity, on normal faults with a southerly downthrow, is evident in mass-flow deposits and slumps from a shelf edge, which partly accounts for the fine preservation and lack of reworking of many of the fossil assemblages collected from these strata. The Craighead Sub-basin records a significant change in the sedimentation pattern of the finer fill in the distal parts of the overall basin. Deeper water turbidites that infill the sub-basins farther south are less abundant here. This is explained by the location of the Craighead Sub-basin, which was nearer to the southern margin of a shelf extending farther north in the Midland Valley (Bluck, 2002).

The thick and variable marine succession preserves important faunas which can be correlatable not only with the other inliers in the Girvan district, but also with successions elsewhere on the eastern margin of the Laurentian palaeocontinent (Floyd and Williams, 2003). The mixture of graptolitic and shelly faunas has provided useful comparisons of their respective biozonal schemes. The inlier is unique in that it contains parts of the Ordovician succession not exposed elsewhere in the Midland Valley Terrane, particularly the Ashgillian Drummuck Subgroup. The inlier has long been a classic area for fossil collecting (for detailed history, see Harper 1982; Rushton et al., 1999a, b; Williams, 2001). The Laurentian (North American) affinity of the faunas is well established, but in the younger formations they become more cosmopolitan as Avalonia–Baltica converged with the Laurentian continental margin.

Ballantrae Ophiolite Complex

The ophiolite complex at Ballantrae, to the south-west of the Ayr district, had been finally obducted over the southern margin of the Midland Valley Arc terrane late in the Arenig (about 470 Ma; Oliver et al., 2008), and experienced marine and then subaerial erosion.

The pillow basalts forming the Craighead Volcanic Formation, at the base of the succession in the Craighead Inlier, have traditionally been correlated with part of the Ballantrae Ophiolite complex (Balcreuchan Group). However, as Floyd (1999) noted, if the cherts associated with the Craighead basalts are coeval with them (rather than overlying them), then the basalts are late Llanvirn or early Caradoc in age (about 458 Ma) and therefore appear to be considerably younger than the lavas at Ballantrae. Several formations in the Northern Belt of the Southern Uplands, for example, the Marchburn Formation, contain pillow basalts of a similar age (Floyd, 1996) and may be related to the Craighead Volcanic Formation.

Craighead Volcanic Formation (CRV)

The Craighead Volcanic Formation (Table 1) comprises mainly basaltic pillow lava and volcaniclastic breccia. The lavas are spilitic basalts apparently associated with bedded red chert. The formation is thought to be over 400 m thick but its base is not exposed. It is unconformably overlain by different formations within the Ardmillan Group; in the east of the inlier, it is the Craighead Limestone Formation which overlies the spilitic lavas as seen in Craighead Quarry [NS 2343 0135]. However, in the west around [NS 216 010] the lavas and cherts are overlain by the Ardwell Farm Formation. Typical exposures of basaltic rock occur in knolls on Craighead Hill [NS 226 014] and in a former quarry [NS 229 016], sheared lava was extracted. Red chert, which is part of the Barneil Member, is interbedded with cherty mudstone and lava at [NS 2225 0147].

The formation is most probably Abereiddian to Caradoc in age (Fortey et al., 2000) as indicated by conodonts found in the red chert of the Barneil Member. The conodont assemblage includes Periodon aculeatus Hadding, a form normally associated with the zonal species Pygodus anserinus Lamont and Lindström, and they indicate a late Llanvirn to early Caradoc age (Lamont and Lindstrom, 1957; Armstrong et al., 1996). If the cherts and lavas are closely related then the Craighead lavas are considerably younger (Floyd, 1999, p.31) than the Arenig Ballantrae Complex with which they had previously been correlated (Ingham, 1992c). It has been suggested that they might correlate with other lavas such as those in the Dalreoch Formation or the Marchburn Formation (Floyd, 1982; 1999).

Ardmillan Group

The Ardmillan Group (Fortey et al., 2000) consists of a sedimentary sequence which is transgressive over the Craighead Volcanic Formation within the Craighead Inlier, where it includes the Ardwell, Whitehouse and Drummuck subgroups (Table 1). The transgressive nature of the base of the group means that its basal formation is not exposed in the inlier. In the main Girvan outcrop to the south of the district, the Ardmillan Group overlies the Barr Group. The Ardmillan Group is unconformably overlain by Silurian rocks in the north-eastern part of the inlier.

Ardwell Subgroup (AWL)

The transgressive Ardwell Subgroup (Caradoc) is the oldest subgroup in the Ardmillan Group. Only a condensed, upper sequence of the subgroup is present in the Craighead Inlier, where it comprises the Craighead Limestone and Ardwell Farm formations, and the latter is faulted out at the top.

Craighead Limestone Formation (CRDL)

The Craighead Limestone Formation includes massive and bedded, nodular limestone with thin alternations of mudstone overlying a basal conglomerate including green spilitic basalt clasts. It is well known for its fossil faunas and some of the thicker graptolitic and shelly mudstone lenses within the limestone have proved to be mid Caradoc in age (low Dicranograptus clingani Zone; Rickards in Tripp, 1980). Conodonts in the limestone suggest the superbus or possibly upper tvaerensis conodont Zone (Bergström, 1990). The formation is about 90 m thick but many exposed sections are faulted as they are situated within the Kerse Loch Fault zone. The disused Craighead Main Quarry at [NS 2345 0142] is the only major outcrop of the formation (Figure 2) and an internationally important type locality (Rushton et al., 1999a). Just east of this quarry, the limestone is overlain by the Ardwell Farm Formation (Plantinhead Member or Flags). Other limestone exposures are present in smaller quarries about 200 m to the south-west around [NS 2327 0116].

Early work on the Craighead Limestone correlated it with the Stinchar Limestone (Lapworth, 1882), and this view was current at the time of the revised mapping by Richey (Richey in Bailey, 1928) and Lamont (1933; 1935) around the Craighead Inlier. Subsequent stratigraphical and palaeontological studies (Reed, 1935; Stubblefield, 1939; Anderson and Pringle, 1946) indicated that the limestone could be correlated with 'Balclatchie Mudstones' and considered a 'reef phase' of limited extent (see Eyles et al., 1949, pp.26–29; 31–32).

Subsequent work (Tripp, 1954; 1980, Williams, 1962; Ingham, 1978; 1992c) described the faunas and the irregular conglomeratic base to the formation, which consists of predominantly transgressive shallow water deposits. A variety of calcareous lithologies are present in the beds above, ranging from brecciated algal (Girvanella) limestones to sandy and rubbly pelmatozoan limestones and calcareous shales, all interdigitating in a complex manner (Ingham, 1978). Two impersistent mudstone members (Figure 2) are known; a lower, Sericoidea Mudstone Member (which contains clingani Zone graptolites) and an upper, Kiln Mudstone Member that lies near the top of the formation (Rushton et al., 1999a). The Sericoidea Mudstone constitutes a deep water interval within the formation interleaved with reef-flank limestones. The Kiln Mudstone Member is more substantial and probably equivalent to part of the highest algal breccia limestones found elsewhere in the Craighead Quarry (Tripp, 1980). The faunas from the Craighead quarries are rich and diverse, including 43 species of brachiopod and 28 of trilobite which have their type material from these localities (Tripp, 1954; 1980; Rushton et al., 1999a). There are marked differences in faunas between the limestone and mudstone, as well as between the mudstone members. The brachiopod and trilobite faunas have a strong Laurentian aspect, and show similarities with faunas found in the Bardahessiagh Formation in Northern Ireland (Mitchell, 2004). There are also bryozoa, bivalves, gastropods, corals, conodonts, cystoids, and cephalopods as well as the graptolites and algae (Tripp, 1980). An ostracod assemblage (Williams and Floyd, 2000) was found to have affinities with Laurentia and Baltica, supporting the suggestion that Iapetus no longer represented a barrier to their migration during mid Caradoc times. Williams (1962) correlated the formation with the upper part of the Ardwell Farm Formation at the Girvan coast. Bluck (2002) envisaged deposition of massive reefoidal limestones built on slight topographical highs in the basaltic basement which were surrounded by bedded inter-reef limestones and shales with abundant crinoids, brachiopods and trilobites.

Ardwell Farm Formation (ARD)

The Ardwell Farm Formation (Fortey et al., 2000) was formerly known as the 'Ardwell Flags' or Ardwell Formation (Floyd, 1999), and comprises a varied succession of grey sandstones, mudstones and conglomerates. The formation is up to 1000 m thick, but is locally faulted out within the Craighead Inlier and unconformably overlain by the Lanark Group on the inlier's north-western flank. Consequently, around 400 m of the formation are exposed in the district.

The succession includes flaggy wacke, fine-grained sandstone, greenish grey siltstone and mudstone, with thick wedges of conglomerate, particularly in the west near Woodhead around [NS 215 012]. Granitic and basic igneous clasts contained within the conglomerate are lithologically similar to those in the older Barr Group, indicating that the same or similar source areas were continuing to supply coarse-grained detritus into the Craighead Basin. This formation includes the basal Plantinhead (Flags/Flagstones) Member (Fortey et al., 2000), but the member is not distinguished on Sheet 14 because outcrops are sparse (see Eyles et al., 1949, pp.28–29). However, cherty sandstone belonging to the formation is exposed at [NS 219 011], and farther east at [NS 2338 0157], a calcareous sandstone bed (0.3 m thick) is interbedded with graptolitic mudstones and siltstones (Plate 1).

Graptolites identified from the formation in the Carrick district (Sheet 8W), range from the wilsoni to upper clingani graptolite biozones within the Caradoc (Floyd, 1999).

Whitehouse Subgroup (WHE)

The Whitehouse Subgroup (formerly of group status; Fortey et al., 2000) can be correlated from the Stinchar Valley and the Girvan coast into the Craighead Inlier. The lower part of the subgroup consists of thin-bedded turbidites, and the upper part comprises mainly mudstones with thin flaggy sandstones. Only the uppermost part of this subgroup is present in the Craighead Inlier (Figure 3); the lower part, which is exposed elsewhere in Ayrshire, is faulted out (Table 1).

Shalloch Formation (SLH)

The Shalloch Formation comprises a monotonous pale grey wacke sandstone, siltstone and mudstone succession over 400 m thick. The formation, extends from the Girvan foreshore into the Stinchar Valley. It was incorporated into the Whitehouse Subgroup by Fortey et al. (2000). The formation crops out towards the core of the Craighead anticline at [NS 253 030] and has been cut by several faults trending east-north-east, roughly parallel to the axis of the anticline. The base of the formation is faulted out at Craighead, so that only the upper part of the formation is exposed in the district, and it appears to be conformably overlain by the Auldthorns Formation (Drummuck Subgroup). This means that the formation here includes younger beds than those known from the Girvan foreshore or Penwhapple Burn (Ingham, 1978). Typical outcrops occur in the burn at [NS 2563 0294]. The succession appears to be poorly fossiliferous here (and was formerly known as the Barren Flagstones), but in the larger outcrop of the formation east of Girvan around [NX 250 980], graptolites from the complanatus to anceps biozones (lower Ashgill) have been recovered (Floyd, 1999).

Drummuck Subgroup (DKG)

The Drummuck Subgroup, of Ashgill age (Table 1), is restricted to the Craighead Inlier and is a marine, mainly argillaceous sequence, about 350 m thick including subordinate arenaceous and rudaceous units. Prior to the stratigraphical revision by Fortey el al. (2000) the Drummuck succession was classed as a group (Harper, 1982; 1984) which excluded the High Mains Formation. The subgroup is richly but not uniformly fossiliferous including trilobite, brachiopod and occasional graptolite and ostracod faunas (Williams, 2001). The stratigraphy and shelly faunas of the subgroup were reviewed and studied in detail by Harper (1982; 1984), who envisaged its deposition in the unstable slope environment of an inner submarine fan. The subgroup has its deeper water correlatives, the Killey Bridge and (overlying) Tirnaskea formations, in Co. Tyrone, Northern Ireland (Mitchell, 2004).

Auldthorns Formation (AUT)

The Auldthorns Formation comprises sandy conglomerate, coarse-grained, greenish sandstone and green mudstone, in a sequence 35 m thick. It is exposed on the south side of Quarrel Hill at [NS 2587 0330] and contains shelly fossils of Cautleyan (Ashgill) age. It has been divided into a lower Escarpment Member and an upper East Brow Member (Harper, 1982; Fortey et al., 2000). Both of these are repeated at [NS 2600 0327] and [NS 2610 0330] by local faulting in the nose of the Craighead anticline (Figure 3). The Escarpment Member, as its name suggests, forms a marked escarpment and is exposed near the ruined farmhouse at [NS 2593 0314]. It comprises greyish green sandstones and conglomerates containing igneous clasts, below thickly bedded, greenish, medium- to coarse-grained sandstone with shelly fossil horizons. The overlying East Brow Member is a sequence of rapidly alternating greyish green, fine- to medium-grained sandstones and mudstones. It differs from the Escarpment Member in having virtually no conglomerate and more and thicker mudstone beds. It also has less abundant and diverse faunal assemblages (Harper, 1982), but both members contain faunas suggestive of a Cautleyan age (Harper, 1984). The top of the East Brow Member is conformable with the overlying Quarrel Hill Formation as exposed at [NS 2523 0260] in the burn south-east of Farden.

Quarrel Hill Formation (QHF)

The Quarrel Hill Formation is 115 m thick, and consists of green- and purple-stained, bedded mudstones with subsidiary thin sandstones and polymictic conglomerates. It is named after the hill at the type area [NS 258 034] and is exposed in a burn to the east at [NS 2621 0340]. The mudstones are locally very fossiliferous. Neither the sediment nor the fauna was considered indigenous to the locality (Ingham, 1978), and was presumed to have derived from an unknown upslope source.

The basal silty mudstones are greyish green, locally fissile with a red-purple coating. Thin green sandstones are interbedded with the mudstones, together with sandy lenses and rare thin 'rottenstones' (Harper, 1982). The fauna in the lower mudstones is abundant and diverse, dominated by brachiopods including dalmanellids, chonetids, strophomenids and leptaenids. In addition, rich gastropod, bivalve and to a lesser extent trilobite faunas have been recovered; sponges, cephalopods, corals, bryozoans, hyoliths and crinoid fragments are also present (Harper, 1982). The middle part of the formation was formerly well exposed in Glenmard Quarry (now backfilled). The section showed fossiliferous conglomerate (Glenmard Conglomerate Bed) and mudflow deposits (Glenmard Member) between bedded mudstones with thin sandstones and the Quarrel Hill Crinoid Bed (Harper, 1982; 1984). The lower and middle parts of the formation are early–middle Cautleyan in age. Late Cautleyan shelly faunas, including trilobites Cryptolithus latus latus Portlock and a species of Tretaspis, are also reported (Ingham, 1992c). The upper part of the formation contains less fossiliferous mudstones and is conformably overlain by more massive, grey silty mudstones of the Lady Burn Formation.

Lady Burn Formation (LBM)

The Lady Burn Formation (Ladyburn Mudstones of Lamont, 1935) comprises an unbedded to locally bedded, locally nodular, pale grey siltstone and greenish grey mudstone succession totalling 140 m thick. The formation crops out in a belt about 300 m wide running through Drummuck Farm, with fairly regular dips of 25–30° to the north or north-west. It is conformably overlaid by the South Threave Formation. Typical exposures are present in Drummuck Burn around [NS 237 031] and Lady Burn near [NS 248 036] and the burn farther east at [NS 2640 0340]. It contains fossils of Rawtheyan age as indicated by the rich assemblage of trilobites, including the trinucleid Tretaspis sortita Reed (Harper, 1984) and also Calymene (sensu lato) drummuckensis Reed and Cybeloides (Paracybeloides) loveni girvanensis (Reed) (Ingham, 1992) near the top of the formation. Low diversity faunas consisting of brachiopods, bivalves and hyoliths have been recovered from localities in both the Drummuck and Lady burns, and lingulid brachiopods and small gastropods are characteristic of certain localities in the formation (Harper, 1982). Brown-weathering silty mudstones containing ostracods in a burn south of Drummuck [NS 2308 0283] were considered part of this formation (Floyd et al., 1999), which necessitated a southwards adjustment of the boundary with the Quarrel Hill Formation as drawn by Harper (1982).

South Threave Formation (STHV)

The South Threave Formation, in the upper part of the Drummuck Group, is essentially an interbedded succession of green mudstone, sandstone and siltstone of upper Rawtheyan (Ashgill) age. In all it is about 60 m thick and includes some unbedded grey silty mudstones. Exposures are found on the northern limb of the Craighead anticline east of South Threave e.g. at [NS 2505 0381]. Overall, exposure is generally poor but detailed investigation (Harper, 1982) established that the formation could be subdivided into three members. The lowest, the Farden Member, contains grey-green mudstones and, in its upper part, thin-bedded, greenish black, fine- to medium-grained sandstones up to about 40 m thick, which includes towards its top, the Lady Burn (or Threave/Thraive Glen) Starfish Beds (Harper, 1982) approximately at [NS 2502 0381]. These include very fossiliferous sandstone beds, dipping 35° to the north near the top of the member, containing dominantly brachiopods but also diverse trilobite, mollusc, echinoderm (crinoids, cystids, starfish, etc.) and hyolith faunas. This world-famous locality (South Threave of Rushton et al., 1999b) contains a Rawtheyan shelly fauna of unrivalled diversity, representing a mixture of relict Laurentian taxa and more cosmopolitan species.

The Starfish Beds were collected by the Gray family from the end of the 19th century into the early decades of the 20th (Reed, 1917; Wood, 1933, Lamont, 1935, Begg, 1939, 1946; Rushton et al., 1999b). Many of the fossils are complete and well preserved, occurring with current orientated graptolites at particular horizons. Recently ostracods were reported from these beds (Floyd et al., 1999) and a new crinoid (Donovan, 1992). It has been suggested (Harper, 1984) that these fossils are well preserved because, after a limited period of post-mortem agitation on the sea floor, the shells were subjected to considerable and sudden down-slope transport and rapid burial.

The Cliff Member lies in the middle of the formation and is about 10 m thick. It consists of fairly massive, bluish grey, soft, nodular, silty mudstones, and crops out in Lady Burn around [NS 2507 0383]. Fossiliferous beds contain few brachiopods but numerous trilobites, particularly Paraproetus girvanensis (Nicholson & Etheridge) and Tretaspis sortita (Ingham, 1992c); many specimens of which are enrolled, suggesting preservation during the process of protective action such as against, for example, sudden turbulence [and] rapid burial (Harper, 1982). They are commonly found in concentric bands around mud 'nodules', suggesting entrainment during downslope catastrophic movement. The graptolite Paraorthograptus pacificus (Ruedemann) collected from this member (Floyd et al., 1999) is indicative of the pacificus Subzone (i.e. the upper subzone of the spiralis Zone).

The uppermost member, the Waterfall Member, is exposed to the east in the Lady Burn section [NS 2514 0385], where it is represented by about 10 m of bedded, poorly fossiliferous, greyish green muddy siltstones. This member is not fully represented here due to the overstepping Mulloch Hill Conglomerate Formation. The member reaches its fullest thickness near High Mains at approximately [NS 2655 0380] but exposure there is poor and faulted. It appears to pass conformably into the overlying Hirnantian age High Mains Formation (Harper, 1981).

High Mains Formation (HMS)

The High Mains Formation was first distinguished as the 'High Mains sandstone', but considered Llandovery in age by Lamont (1935). It comprises a fine- to medium-grained, calcareous sandstone succession, about 30 m thick. The sandstone is pale grey when fresh, but weathers brown. The formation is exposed in a faulted zone west of High Mains Farm at approximately [NS 2643 0376] and in the old quarry in High Mains Wood at [NS 2632 0380]. At the latter locality the thinly bedded sandstones dip about 35° to the north-north-east and are largely devoid of fossils, except for small crinoid ossicles at a few horizons. The formation is unconformably overlain (and largely cut out) by the Silurian Girvan Group.

The stratigraphy and Hirnantian faunas of the formation have been described in detail (Harper, 1979; 1981). In the temporary trench available to Harper (1981) two units were distinguished; a lower one of thinly bedded, grey-green sandstones, and an upper unit of medium-bedded, hard medium-grained sandstone. Both contain horizons with crinoid ossicles and shelly debris, with a brachiopod fauna dominated by small Hindella crassa incipiens (Williams) and Eostropheodonta aff. hirnantensis (M'Coy) in the lower unit, and large H. crassa incipiens and Hirnantia sagittifera (M'Coy) in the upper unit. Trilobites, molluscs, a graptolite fragment and a few bryozoans were also recovered (Harper, 1981). The sparse but relatively diverse relict American trilobite fauna (Ingham, 1992c) from the formation included the new species Hemiarges extremus Owen (Owen, 1986). Crinoid columnal fragments are locally very common on certain bedding planes. A marked change in lithology and fauna is apparent passing up the formation (and within the Hirnantian Stage) which indicates a relative shallowing (Harper et al., 1994) and minor marine regression (accompanying a widespread glacigene event) in this area. The Hirnantia fauna is believed to represent one of a number of colder water ecosystems that became almost globally distributed with the onset of the Ashgillian Glaciation. The High Mains Formation has a significant fauna, containing brachiopod and trilobite genera more typical of the mid Ashgill of the North American Province, as well as the Hirnantia fauna (Rong and Harper, 1988). The Hirnantia fauna of the High Mains Formation is therefore atypical (Harper et al., 1994), and the trilobite fauna lacks Mucronaspis, characteristic of most Hirnantian assemblages (Owen, 1986). The Hirnantia fauna as a whole disappeared at the end of the Ashgill, for complex reasons, probably related to the widespread fall in sea level (regression) as the glaciation ensued, followed by a rapid transgression in anoxic conditions as the ice melted. Elements of the Hirnantia fauna are found in the upper part of the Tirnaskea Formation of Northern Ireland (Harper et al., 1994), but occur in a deep water facies, suggesting a more offshore setting for the Irish deposits.

Barr Group

South of the Straiton Fault, parts of two small Ordovician inliers lie within the district and both contain strata belonging to the Barr Group, which has a wider distribution to the south within the Midland Valley Terrane (Floyd, 1999).

The Barr Group (Fortey et al., 2000) comprises a dominantly conglomeratic succession of Llanvirn to early Caradoc age (Table 2). Its age was determined by identifying fossils from thin limestone interbeds that occur in the relevant type sections on the Carrick (8W) Sheet (Floyd, 1999). The Barr Group succession (Williams, 1962; Ince, 1984) has been interpreted as recording two main phases of fan-delta development (producing the Kirkland and Benan conglomerates), with an intervening phase of shallow water, carbonate-rich sedimentation (the Stinchar Limestone and equivalents). South of the Straiton Fault in the Ayr district, the Barr Group crops out in two faulted inliers, which are in part unconformably overlain by the Siluro-Devonian Lanark Group. Both the inliers expose the Benan Conglomerate Formation and extend southwards into the Carrick district (Floyd, 1999). The Benan Conglomerate lies at the top of the Barr Group and neither its base, nor its conformable passage into the Balclatchie Formation (Ardwell Subgroup), are exposed on Sheet 14W.

Benan Conglomerate Formation (BNC)

The Benan Conglomerate Formation is a distinctive lithostratigraphical unit of Caradoc age within the Barr Group (Table 2). The conglomerate varies greatly in thickness in the Carrick district; but on the Ayr Sheet it appears to be 400–500 m thick and overlain unconformably by the Lanark Group. The variation in thickness is an original feature produced by deposition of the fan system, although some units are interpreted as mass flow deposits, banked against synsedimentary faults throwing to the south-east (Williams, 1962; Ince, 1984; Ingham, 1992a).

In this district, one inlier of Benan Conglomerate Formation occurs east of Genoch at [NS 395 010], and is referred to here as the Genoch Inlier. Another inlier lies west of Glenalla Field at [NS 330 007] and is called The Pilot Inlier, after a hill situated just to the south of the sheet. The bulk of the formation, as exposed on Sheet 14W, comprises a crudely stratified, matrix-supported boulder conglomerate in a grey coarse-grained sandstone matrix.

In the Genoch Inlier, Benan Conglomerate crops out on either side of the Water of Girvan [NS 395 010], but the base of the formation is cut out by faulting. In this area the formation is unconformably overlain by the Greywacke Conglomerate Formation at the base of the Siluro-Devonian Lanark Group. The Benan Conglomerate is generally massive with variably coloured, mainly igneous pebbles to boulders, in a greenish grey, granule to sandstone matrix. However, locally e.g. [NS 3935 0063], the conglomerate is weakly bedded, dipping 42°–50° to the north-west. It is estimated from the dip of the outcrop that the formation is at least 500 m thick in this area. The clasts are generally subrounded to rounded, and are usually matrix-supported. A weak pebble alignment is commonly present but no clear imbrication was observed. In places, beds, 0.4–1 m thick, are clast-supported and mainly contain small pebbles and scattered cobbles. The clasts largely consist of basic igneous rocks but also fine-grained microgranite, medium-grained pink granite, fine-grained reddish brown, siliceous sandstone, pale grey, fine-grained limestone and green cherty rock. The conglomerate has locally been hardened by calcite-veined zones, and some pebbles are cracked and slightly offset within the matrix. The matrix itself is a grey-green, poorly sorted, closely-packed wacke, containing subrounded fragments of similar lithologies to the pebbles, and also includes vein quartz, and amphibole- and epidote-rich clasts.

The Pilot Inlier exposes strata within the lower part of the conglomeratic formation on the southern margin of Sheet 14W [NS 340 006], and includes interbeds of coarse-grained sandstone and granule conglomerate. The lenses of limestone mapped in the previous survey were not confirmed in the field, and appear to be carbonate-cemented intervals. This inlier is unconformably overlain by the Swanshaw Sandstone Formation.

The Benan Conglomerate Formation has been interpreted as a mass-flow deposit, and in the type area north of the Stinchar valley, it has a very variable thickness (60–750 m, Floyd, 1999). Pink granite clasts from the type area (Floyd, 1999) have no known source presently exposed, and have been dated at approximately 470 Ma (Longman et al., 1979). The Benan Conglomerate is known to be Caradoc in age (Floyd, 1999), so the granite clasts are not much older than the conglomerate itself. Floyd (1999) concluded that most of the basic clasts could have been derived from the Ballantrae Ophiolite Complex and the older parts of its cover sequence. However, Rushton and Tripp (1979) found a single boulder of limestone in the type area, containing an early Tremadoc fauna (about 490 Ma) for which there is no known source presently exposed in Scotland.

The lower part of the formation has been interpreted (Ince, 1984) as having accumulated in relatively deep water on a submarine fan-complex, situated offshore from a fan-delta. The upper part of the formation, as seen in the type area, includes thin, in situ algal limestone encrustations and boulders of coral-algal limestone (Floyd, 1999), and has been interpreted as a relatively shallow water deposit (Ince, 1984). No mappable limestones have been identified in the district but the carbonate-cemented beds may represent the shallow-water facies. The granitic clast content of the Benan Conglomerate has been related to source uplift related to granitic magmatism in the area to the north, and the formation's large variation in thickness suggested oblique slip on contemporaneous faults on the northern margin of Iapetus (Ince, 1984).

Chapter 3 Silurian

The Silurian (Llandovery–Wenlock age) strata in the Ayr district form part of the well-known Girvan inliers situated on the south side of the Midland Valley (Bluck, 2002). Nearly the whole of the Craighead Inlier lies within the Ayr district, together with the north-eastern part of the Girvan main outcrop (Cocks and Toghill, 1973) which is truncated to the south by the Straiton Fault, but extends south-west towards Girvan (Floyd, 1999). In the earlier descriptions of the Girvan inliers, Cameron et al. (1986) and Floyd (1999) divided the Silurian beds into Penwhapple South and Penwhapple North groups, but these have been superceded by the Girvan Group (Floyd and Williams, 2003).

Girvan Group

All the Llandovery–Wenlock age strata in the Girvan inliers have been assigned to the Girvan Group (Floyd and Williams, 2003). It has been subdivided into the Newlands, Dailly and Straiton subgroups (Table 1). The Craighead Inlier contains Llandovery strata whereas the Girvan main outcrop includes beds ranging from Llandovery to Wenlock age.

Craighead Inlier

Within the Craighead Inlier (Figure 3), Llandovery strata lie unconformably above Ordovician strata. The inlier forms an internationally important natural depository for Llandovery fossils. Lapworth (1882) was the first to establish the Silurian succession up to the Glenshalloch Shale Formation, and Lamont (1935) revised the stratigraphy and palaeontology. The Silurian crop of the inlier was extended to the north-east when Freshney (1959) recognised younger Llandovery strata, above the Glenshalloch Shale, which had formerly been classed as 'Old Red Sandstone'. The biostratigraphy of the Girvan Silurian outcrops including the Craighead Inlier was studied by Cocks and Toghill (1973) and reviewed by Williams (2001). A subsequent revision of the correlation and nomenclature was made by Floyd and Williams (2003).

Newlands Subgroup (NEWL)

The Newlands Subgroup (Llandovery) was set up in the revised correlation of the Silurian rocks in the Girvan district (Floyd and Williams, 2003). It comprises all the formations in the first major transgressive sequence after the Late Ordovician (Ashgillian) glaciation. Over 1300 m of strata lie between the unconformity on the Ordovician and the base of the next transgression (the Lower Camregan Grits Formation). The clastic sequences show repeated gradations between conglomerates, sandstones and fine-grained graptolitic shales, from which Cocks and Toghill (1973) collected and described differing faunal communities. They related the lithological and faunal differences to water depth and sediment supply.

This subgroup has a pronounced overstep on the Ordovician rocks near the Girvan coast, but this becomes less significant farther east within the Craighead Inlier. However, the unconformity (Table 1) represents a significant time gap, and appears to reflect a period of tectonic deformation (at least open folding) and uplift in early Rhuddanian times.

Mulloch Hill Conglomerate Formation (MLCG)

The Mulloch Hill Conglomerate Formation (Floyd and Williams, 2003) comprises massive, grey-buff, polymict conglomerates with well-rounded pebbles commonly 20–30 mm in diameter, but some boulders are over 0.15 m long. According to Eyles et al. (1949) the clasts include Arenig igneous rocks, chert and quartzite, set in a dull purple sandy matrix. There are subordinate interbedded sandstones which are locally fossiliferous. The whole sequence is estimated to be 103 m thick (Cocks and Toghill, 1973, who referred to it as the Lady Burn Conglomerate). This resistant lithology is exposed on the unnamed hill at [NS 261 039] and near High Mains at [NS 2673 0380]. Its unconformable base above the South Threave Formation is exposed in Lady Burn at [NS 2515 0386]. The formation also crops out at [NS 2600 0300], in a fault lens caught up in the Kerse Loch Fault zone.

A low diversity crinoid and brachiopod fauna (a shallow-water Cryptothyrella community) was collected from the sandstones (Cocks and Toghill, 1973), and indicates a mid Rhuddanian age. This environmental interpretation is supported by the record (Bruce, 1992) of hummocky cross stratification (HCS), indicative of relatively shallow water deposition within storm wave base. The formation has also been interpreted as a channel-fill on a submarine fan (Ward, 1989).

The top of the formation is not exposed but assumed to pass up conformably into the Mulloch Hill Sandstone Formation.

Mulloch Hill Sandstone Formation (MLHS)

The Mulloch Hill Sandstone Formation (defined by Floyd and Williams, 2003) was formerly called the Mulloch Hill Sandstone (Lamont, 1935) or the Mulloch Hill Formation (Cocks and Toghill, 1973). It contains ochreous weathering green and buff sandstones with subordinate interbeds of siltstone and mudstone; in all up to 240 m thick. The formation contains a large and diverse shelly fauna of brachiopods, trilobites, gastropods, echinoids, crinoids and corals. It is described as the most fossiliferous formation in the Craighead Inlier (Aldridge et al., 2000), and well displayed in 'Roughneck Quarry' (sic), the disused quarry 120 m west-south-west of Rough Neuk at [NS 2703 0398]. Cocks and Toghill (1973) suggested an age in the upper half of the Rhuddanian for the formation. The brachiopod assemblages indicate a shallow water Cryptothyrella community for most of the formation, progressing to deeper-water Stricklandia and Clorinda communities towards the top (Cocks and Toghill, 1973). Sandstones in the lower part of the formation are reported (Bruce, 1992) to contain hummocky cross stratification (HCS) supporting the interpretation of a shallow water environment.

Several exposures of the sandstone occur beside the road south of Kirk Hill e.g. at [NS 2648 0414] and in a former quarry south of Craigens Hill at [NS 2593 0423]. The Rough Neuk Sandstone Member, in the upper part of the formation (Ward, 1989), is exposed in the disused Rough Neuk (Roughneck or Mulloch Hill) Quarry [NS 2703 0398] where beds dip about 40° eastwards. A long list of fossils from the quarry was given by Peach and Horne (1899) including corals (Aulacophyllum sp., Favosites sp. and Heliolites sp.), trilobites, gastropods, bivalves, conulariids, tentaculitids, bryozoans, asteroids, crinoids, dendroid graptolites, orthocones and numerous brachiopod species. Cryptothyrella community brachiopods, such as Dalmanella sp. and Mendacella mullockiensis (Davidson) as well as corals were reported from the rich fauna found in the quarry (Cocks and Toghill, 1973). The calcareous alga Mastopora fava (Salter) also indicates a shallow water environment within the photic zone. The quarry is the type locality for several fossil taxa, including the brachiopods Philhedrella mullochensis (Reed), Isorthis prima Walmsley & Boucot, Fardenia (Fardenia) columbana (Reed), Eostropheodonta mullochensis Reed and Rostricellula mullochensis (Reed). This fossiliferous unit includes several trilobites such as Calymene ubiquitosa Howells (Howells, 1982; Ingham, 1992c) and the Rough Neuk Starfish Bed containing dendroid graptolites and starfish in a grey siltstone matrix. Ward (1989) interpreted this as an obrution deposit or 'smothered bottom assemblage'. The discovery of the crinoid Petalocrinus at the quarry extends the palaeobiogeographical range of the genus to the Scottish part of Laurentia (Fearnhead and Harper, 2007). In a transition at top of the formation, at [NS 2703 0445] a Clorinda brachiopod community also contains trilobites and the coral Halysites (Cocks and Toghill, 1973).

The general 'fining up' clastic sequence into the Glenwells Shale Formation, and the interpreted deepening of the depositional environment, are consistent with a mid–late Rhuddanian transgressive event.

Glenwells Shale Formation (GWSH)

The argillaceous Glenwells Shale Formation (Floyd and Williams, 2003) includes hard, pale and dark grey, interlaminated (about 10 mm thick) siltstones and mudstones. They may be turbiditic in origin and represent deposition in relatively open, deep water. Some bluish mudstones are nodular and some contain brachiopods and graptolites e.g. at [NS 2718 0450]. The shales exposed in the burn running through Glenwells Wood at [NS 2733 0418] yielded a sparse graptolite fauna (Climacograptus cf. rectangularis (M'Coy), C. ex gr. scalaris (Hisinger),? Diplograptus modestus Lapworth and various dendroids, together with a single alga, Masapora fava and a brachiopod, Eoplectadonta sp. (Cocks and Toghill, 1973). They concluded that the faunas indicated an upper cyphus Zone age for the formation. The succession is estimated to be 97 m thick (Cocks and Toghill, 1973) and was divided into lower calcareous siltstone, middle unfossiliferous mudstone, and upper pale blue, graptolite-bearing shale members (Ward, 1989). The cyphus Zone (late Rhuddanian) graptolite faunas are the oldest stratigraphically useful markers in the Silurian of the Girvan inliers and allow correlation with the shelly faunas. Shales faulted between the Mulloch Hill Conglomerate and Carboniferous rocks in the burn south-east of Farden at [NS 2534 0255], yielded Monograptus revolutus Kurck and M. gregarious Lapworth to Cocks and Toghill (1973), and were considered by them to correlate with the Glenwells Shale Formation. However, on the recent survey, these shales have been mapped as Shalloch Formation.

Glenwells Conglomerate Formation (GWC)

The Glenwells Conglomerate Formation is thickly bedded and poorly sorted, with pebbles and cobbles over 8 cm long. There are a few interbedded, coarse-grained sandstones in the succession, which is up to 100 m thick. At an outcrop in a disused quarry north of Kirk Hill at [NS 2696 0454], about 15 m of conglomerate and sandstone overlies mudstone belonging to the Glenwells Shale Formation (Cocks and Toghill, 1973). The Glenwells Conglomerate was first recognised and mapped separately by Cocks and Toghill (1973), who stated that it was well exposed in Glenwells Burn below [NS 2738 0409], dipping between 55° and 60° south-east. No fossils are recorded from the formation. It probably represents a channel-fill deposit in a deep water submarine fan. It is considered to form the basal unit to the fining-up cycle, passing up conformably into the Newlands Farm Formation (see Ingham, 1992c; Cocks and Toghill, 1973; Freshney, 1959).

Newlands Farm Formation (NWF)

The Newlands Farm Formation (Floyd and Williams, 2003; formerly Newlands Formation of Cocks and Toghill, 1973) comprises generally blue-grey calcareous siltstone and mudstone, weathering to a yellow-buff colour. Sandstone beds and well-rounded clasts, up to 3 cm in diameter, occur sporadically (Cocks and Toghill, 1973). Typical outcrops are found about 200 m east of Newlands Farm at [NS 2777 0440]. The formation is named after the farm and was commonly referred to as the Newlands Formation (e.g. Ingham, 1992c). It is estimated to be 185 m thick and contains a rich and diverse fauna, including brachiopods, trilobites and algae, representing the deeper shelf Stricklandia and Clorinda communities, but still within the photic zone as indicated by the presence of alga Mastapora fava. The assemblage is considered to be low Aeronian in age and also contains graptolites such as Monograptus sp. and Dictyonema sp., which are non-diagnostic of age. The assemblage is of great interest as it is the only shelly fauna from the middle Llandovery in the Girvan area. Among the brachiopods found, the pentamerids Clorinda undata (J de C Sowerby) and Stricklandia lens intermedia Williams are probably the most common (Ingham, 1992c). At a locality [NS 2771 0457] within 10 m of the top of the formation, Cocks and Toghill (1973) described a bedding plane crowded with the single species Stricklandia lens intermedia. At another locality [NS 2631 0471], towards the base of the formation, they collected the same brachiopod species in a more diverse fauna including crinoid fragments. In the bank to a small burn at [NS 278 046] Ingham (1992c) described this 'classical and well-known Newlands locality', and lists a diverse trilobite fauna including Acernaspis sp., Encrinurus sp., Calymene ubiquitosus and Cyphoproetus externus Reed (Howells, 1982). The latter trilobite is a widespread species at the Rhuddanian–Aeronian boundary (Williams, 2001). The type species of the star-shaped crinoid columnal Floricolumnus girvanensis Donovan & Clark was collected from the Newlands Farm Formation and another species Dimerocrinites sp. A., is also present (Donovan and Clark, 1992). The latter authors suggest that the fauna spans the Rhuddanian/Idwian (Aeronian) boundary rather than being the younger gregarius Zone suggested by Howells (1982). It appears to be conformably overlain by the Glenshalloch Shale Formation.

Glenshalloch Shale Formation (GLSH)

The Glenshalloch Shale Formation (Floyd and Williams, 2003; formerly the Glenshalloch Shale of Cocks and Toghill (1973), contains greyish green or blue-grey, thinly-bedded, soft mudstones and siltstones, weathering orange-brown and pink. Typical outcrops are seen in the Baldrennan Burn at [NS 2827 0431] dipping 75° to the south-east, and on the south-west side of Glenshalloch Hill mudstones dip north-east, on the northern flank of the Craighead anticline. Basal beds are exposed 250 m north-north-east of Newlands Farm [NS 2770 0458] dipping 20° north-east. The formation is 305 m thick and contains several graptolite-bearing intervals, which become more common in the uppermost 75 m (Cocks and Toghill, 1973). Two graptolite localities in the Baldrennan Burn provided well-preserved assemblages. The lower one at [NS 2824 0440] includes a plentiful fauna indicating the magnus Subzone of the gregarius Zone and the upper one, 15 m below the top of the formation at [NS 2827 0433], also yielded a gregarius Zone fauna, but possibly from the younger argenteus/leptotheca Subzone (Cocks and Toghill, 1973). The formation was deposited away from coarser clastic input, probably in deep, open water.

Saugh Hill Grits Formation (SHSG)

The Saugh Hill Grits Formation (Floyd and Williams, 2003) is composed of coarse-grained, massive, greyish green, turbiditic sandstone, with horizons of quartz granule conglomerate. The unit was first recognised within the Craighead Inlier by Freshney (1959), who correlated it with the Saugh Hill Grits of the main Girvan outcrop. The turbidite beds are fine-grained at the base, but mostly medium- to coarse-grained above. Conspicuous sole markings are recorded from the bases of the turbidite beds (Cocks and Toghill, 1973). Typical outcrops are seen in the Baldrennan Burn at [NS 2831 0427] and east of Glenshalloch Hill at [NS 2876 0502]. North of Glenshalloch Hill, grey mudstone is exposed [NS 2844 0520]. The formation is 290 m thick and can be correlated with the upper part of the Saugh Hill Grits in the main Girvan outcrop (Cocks and Toghill, 1973; who referred to it as the Upper Saugh Hill Grits). No fossils have been recorded in this part of the formation and it is assumed to be conformably overlain by the Pencleuch Shale Formation.

Pencleuch Shale Formation (PCH)

The Pencleuch Shale Formation (Floyd and Williams, 2003) comprises a grey and black siltstone and mudstone succession at least 50 m thick. The beds are poorly exposed in a ditch at [NS 2937 0543] about 250 m south-east of Carscallan Cairn. This unit was first recognised in the Craighead Inlier by Freshney (1959) as 'Sedgwicki Shales', and correlated with the main Girvan outcrop. The graptolite fauna is typical of the upper convolutus Zone, but includes sparse Monograptus sedgwickii (Portlock) and so may well be close to the convolutus–sedgwickii (sensu lato) Zone boundary (Cocks and Toghill, 1973). A comparison of faunas collected from this formation and the overlying Lower Camregan Grits Formation (Dailly Subgroup) led Cocks and Toghill to the conclusion that there was a depositional hiatus between them (Table 1).

Dailly Subgroup (DAI)

The Dailly Subgroup (Floyd and Williams, 2003) includes all the units of the second transgressive sequence in the Llandovery–Wenlock strata around Girvan. There appears to be a depositional hiatus between the Dailly Subgroup and the underlying Newlands Subgroup. Within the Craighead Inlier, the Lower Camregan Grits Formation is the only unit of this subgroup exposed.

Lower Camregan Grits Formation (LCRG)

The Lower Camregen Grits Formation consists of red-stained, medium- to thick-bedded turbiditic sandstones, with graded bedding and sporadic sole marks. It is poorly exposed but the lithology is considered to be very similar to that in the Girvan main outcrop (Freshney, 1959). The only permanent exposure of this formation within the inlier is a disused quarry in Craigfin Wood at [NS 2964 0540], where reddened silty sandstone beds are vertical and strike north-eastwards. The locality yielded a diverse fauna including brachiopods of which the most significant is Eocoelia curtisi Ziegler (Cocks and Toghill, 1973). A temporary exposure dug nearby at [NS 2941 0545] produced a different brachiopod-rich community dominated by Pentamerus oblongus J de C Sowerby but still with Eocoelia curtisi (Cocks and Toghill, 1973). The Eocoelia community probably indicates shallow shelf deposition and the Pentamerus community mid-shelf deposits; the fossils could also have been transported but not far, as the shells are mainly intact. The record of Eocoelia curtisi suggests an early Telychian age for the Lower Camregan Grits Formation (Cocks et al., 1984; Floyd and Williams, 2003). However, the record of Stricklandia lens progressa from the Pentamerus community locality indicates an upper Aeronian age, so the assemblage would appear to lie close to the Aeronian–Telychian boundary. The lack of graptolite control at the stratotype boundary in Wales means a definitive classification of age is problematic (but probably of sedgwickii (sensu lato) age). Partly to resolve the problem, the pit in Craigfin Wood was extended (Doyle et al., 1991) into younger strata, and new fauna containing a species of Pentameroides and E. cf. curtisi curtisi Ziegler was identified (and considered Telychian). The E. curtisi recorded by Cocks and Toghill (1973) within the formation were reclassified as E. curtisi immatura Doyle, Höey & Harper, an earlier subspecies (Doyle et al., 1991). However, this presented considerable stratigraphical difficulties, suggesting complex faulting. Floyd and Williams (2003) also point out that the range of E. curtisi immatura needs to be investigated with relation to the graptolite biozones.

Trace fossils and hummocky cross stratification have also been reported from the formation (Bruce, 1992), but its true nature and thickness within the Craighead Inlier cannot be ascertained because it is strongly affected by deformation in the Kerse Loch Fault zone. The full thickness of the Lower Camregan Grits in the main inlier at Penwhapple Glen is only 36.3 m (Cocks and Toghill, 1973), but within the Craighead Inlier, even allowing for the faulting, the nearly vertical exposed succession must be over 50 m thick. The identification of a Pentamerus community within the succession may mean that the overlying Wood Burn Formation in the Girvan main outcrop, is also present in the Craighead Inlier (Cocks and Toghill, 1973). The Lower Camregan Grits Formation appears to be unconformably overlain to the north-west by the Siluro-Devonian Swanshaw Sandstone Formation, which is locally baked where intruded by a Siluro-Devonian quartz-dolerite.

Girvan main outcrop

The north-eastern part of the Girvan main outcrop, lying to the north of the Straiton Fault and south of the Kerse Loch Fault, falls within the district. It consists of the upper part of the Girvan Group (part of the Dailly Subgroup and the overlying Straiton Subgroup) of Llandovery–Wenlock age; the lower part is cut out by the Straiton Fault (see the upper part of (Table 1)).

Dailly Subgroup

The upper two formations in the Dailly Subgroup (Floyd and Williams, 2003) come to crop between the Straiton Fault and the Straiton Subgroup. They are locally overstepped by the Carboniferous Inverclyde Group. These two are the Drumyork Flags Formation succeeded by the Blair Shale Formation. The two formations had previously been part of a now obsolete Straiton Group (Clarkson et al., 1998).

Drumyork Flags Formation (DYF)

The Drumyork Flags Formation is the oldest formation exposed in the main Girvan outcrop around Straiton. The flags are generally thin- to medium-bedded, greyish-green, fine-grained turbiditic sandstones, which in total have a thickness of about 610 m (Cocks and Toghill, 1973), but in this district their base is truncated by the Straiton Fault. The sandstones are commonly micaceous, and in Glenmartin Burn at [NS 3289 0223], thickly bedded. Good exposures of the interbedded grey sandstones, siltstones and mudstones occur in Balsaggart Burn. In the latter burn at [NS 3411 0288] Monograptus dextrorsus Linnarsson has been found within the Drumyork Flags, but no diagnostic faunal assemblage was found in the Blair and Knockgardner area (Cocks and Toghill, 1973). In the adjacent Carrick district (Sheet 8W) to the south, Cocks and Toghill (1973) collected a graptolite assemblage from the formation indicating a greistonensis Zone age. Re-assessment of this fauna (Floyd and Williams, 2003) indicates an age not older than the middle spiralis Zone. Acritarch floras from this formation correlate with those from the North Esk Inlier, which are known to occur in strata of spiralis Zone age (Molyneux et al., 2008). Since the top of the underlying Lauchlan Mudstone Formation and the lower part of the succeeding Blair Shale are also spiralis Zone, the succession must have been deposited very rapidly, during a single graptolite biozone (Floyd and Williams, 2003).

The formation is exposed in a belt from an old quarry south of Knockgardner [NS 3503 0329] to the south-west of Straiton [NS 376 041]. This belt comprises about 200–250 m of greenish-grey turbidites, constituting the upper, relatively thin-bedded part of the formation, abruptly cut out to the south-east by the Straiton Fault. The beds generally dip steeply to the south-east, but younger Silurian formations succeed to the north-west, so that although there is no good younging evidence from within the formation, it appears to be overturned. The flags are exposed in the tributaries to the south of Cawin Burn at [NS 368 037], [NS 371 038] and[NS 376 041]. Mid grey to khaki coloured, very fine- to medium-grained wacke sandstones are interbedded with siltstones and silty mudstones. The beds are generally 0.1–0.4 m thick, and many are calcareous and finely micaceous. Most of the exposures close to the Straiton Fault contain crushed and flexed zones invaded by thin calcite veins. Some alteration has occurred which causes the sandstones to become orange-brown to rusty brown in colour. In places (e.g. at [NS 3765 0414]) crushed silty mudstone beds have developed a crenulation cleavage dipping up to 60° to the south-east. A sample (MPA 49702) of silty mudstone from the tributary south of the Cawin Burn at [NS 3714 0382] yielded an assemblage (Molyneux, 2001; Molyneux et al., 2008) dominated by sphaeromorph acritarchs, Moyeria cabottii (Cramer) Miller & Eames, and Tylotopalla (T. deerlijkianum (Martin) Martin – T. astrifera Kiryanov group). In addition, acritarchs Diexallophasis denticulata (Stockmans & Willière) Loeblich , Domasia limaciformis? (Stockmans & Willière) Cramer, Domasia trispinosa Downie, Dorsennidium europaeum (Stockmans & Willière) Sarjeant & Stancliffe, Micrhystridium stellatum Deflandre and Multiplicisphaeridium sp. are present with rare spores, including cryptospores Laevolancis divellomedia (Chibrikova) Burgess & Richardson and Tetrahedraletes medinensis Strother & Traverse and the miospore Ambisporites?. This assemblage lacks chitinozoans and scolecodonts but is similar to the assemblages found north-west of the Straiton Fault (see below) and in the Reservoir Formation (North Esk inlier). The age of the Reservoir Formation is late Llandovery (spiralis Zone, Bull and Loydell, 1995) and a similar age is therefore suggested for the Drumyork Flags Formation.

Another slice of the Drumyork Flags Formation lies to the north-west of the Straiton Fault, east of Straiton at [NS 399 051]. Turbiditic, fine-grained, khaki-grey and grey-green sandstones, siltstones and silty mudstones dip about 52° south-east and young to the north-west on the evidence of grading in Bouma units 0.1–0.15 m thick. A thin section (N2899) of one of the sandstones shows fine lamination and cross-bedding in quartz, feldspar, biotite, chlorite and muscovite grains cemented by carbonate. The assignment of these beds to the Drumyork Flags Formation is supported by the identification (Molyneux, 2001; Molyneux et al., 2008) of acritarchs, rarer cryptospores, chitinozoans and scolecodonts. The acritarch assemblage (MPA 49700) from this locality [NS 3999 0512] is dominated by sphaeromorph acritarchs and the Tylotopalla deerlijkianum–T. astrifera group. The full assemblage is listed in Molyneux et al. (2008) and includes Moyeria cabottii and Visbysphaera spp. comparable to the assemblage found in the Reservoir Formation of the North Esk Inlier. A similar assemblage (MPA 50941) was recovered from the formation collected from Toddy Burn [NS 3196 0160] where the Tylotopalla deerlijkianum–T. astrifera group microflora was accompanied by Visbysphaera connexa Le Hérissé among other acritarchs but lacking Moyeria. The acritarch V. connexa has its first appearance in the spiralis Zone of Gotland and ranges into the Wenlock (lower part of the Sheinwoodian Stage). The last appearance of Tylotopalla deerlijkianum is close to the Llandovery–Wenlock boundary on Gotland (see discussion in Molyneux et al., 2008) and the distinctive sphaeromorph–Moyeria–Tylotopalla assemblage has potential for correlating the Silurian successions in the Midland Valley.

The narrow belt of Drumyork Flags up thrown between the Bargany and Straiton faults is exposed in the Lindsayston Burn at [NS 292 007], on the adjacent Carrick Sheet, where green fissile mudstone and granule-grade conglomerate were recorded dipping from vertical to 35° south-eastwards. Two thin clays interpreted as metabentonites (Batchelor, 1999) are interbedded within the Drumyork Flags on the Carrick Sheet at [NX 2445 9930], and may correlate with metabentonites of similar age in the Ree Burn Formation of the Hagshaw Hills, and the Wether Law Linn Formation of the Pentland Hills. Grey-green clays within the formation in the Ayr district lie near the Straiton Fault, e.g. [NS 339 051], and were considered to be fault clays, but may in fact have been metabentonites originally.

The graptolite assemblages from the formation were considered (Cocks and Toghill 1973) to range from the greistonensis to the 'crenulata' zones. However, if the formation is of similar age to the Reservoir Formation (see above) then the upper part exposed here is probably middle spiralis Zone and the Blair Shale Formation, upper spiralis Zone (see Bull and Loydell 1995; Loydell in Clarkson et al., 1998, p.63).

According to Clarkson et al. (1998), three lithofacies can be identified in the Drumyork Flags Formation. Only the upper one, Facies 3, is present in the Knockgardner area, comprising massive sandstones, cross-bedded and occasionally graded sandstones, interbedded with sandy shales. Pebble lags were recognised at some levels. This lithofacies was considered relatively deep water, mid to outer shelf in origin, and the coarser sediment was introduced both as non-turbulent gravity slides and by turbidity currents, probably generated by distant storms and palaeoslope instablity (Clarkson et al., 1998). Sphaeromorph-dominated assemblages are generally considered to be characteristic of either nearshore or deep basinal marine environments (Molyneux et al. 2008) and the common occurrence of Moyeria cabottii in nonmarine to nearshore environments of Wenlock age (e.g. Molyneux, 1992; Wellman and Richardson, 1993) would tend to favour the shallow water depositional model. This interpretation is at variance with the relatively deep water environment suggested by Clarkson et al. (1998) above. It is possible that the palynomorph assemblage was transported into the deeper water environment, but a shallow water environment was also inferred for the Reservoir Formation by Bull and Loydell (1995) on the basis of the limited graptolite assemblages.

Blair Shale Formation (BLRS)

The Blair Shale Formation is the uppermost formation in the Dailly Subgroup (Floyd and Williams, 2003; Cocks and Toghill, 1973) and consists of 84 m of mainly thin-bedded, yellow-brown sandstones, siltstones and shales (mudstones). The beds are generally close to vertical or overturned, steeply dipping south-east but they young north-west in graded beds. The formation is exposed mainly in gullies and small quarries south of Blair Farm [NS 324 024]. This is a geological conservation site and the type section (Aldridge et al., 2000), in the gully 70 m south-south-east of Blair Farm [NS 3248 0233], also exposes the transitional base of the overlying Knockgardner Sandstone Formation. The succession is dominated by green or grey mudstones with thin carbonaceous beds containing graptolites. On Badminnie Knowe [NS 3269 0238], brown to grey-green, fine-grained micaceous sandstone is also exposed. An old quarry 180 m east of Blair Farm [NS 3265 0240] has yielded a few specimens of Monoclimacis sp. (Peach and Horne, 1899).

At the Blair Farm locality, the uppermost 27 m of the formation contain three graptolite horizons (Cocks and Toghill, 1973). The lowest and highest of these horizons contain abundant vomerinids, and the middle horizon abundant spiral monograptids (for full assemblage see Cocks and Toghill, 1973), which they assigned to the 'complanatus' Zone (uppermost Llandovery). According to Loydell (in Clarkson et al., 1998) the formation can be refined to the spiralis Zone in age. The upper part may, however, be equivalent to the lapworthi and insectus graptolite biozones (Floyd and Williams, 2003). A significant tectonic junction with the underlying Drumyork Flags was postulated by Clarkson et al. (1998), but this was not confirmed in the recent survey. There is no tectonic evidence other than would be expected between deformed successions of differing competence.

The Blair Shale Formation is relatively soft compared to the underlying Drumyork Flags and the overlying Knockgardner formations, and therefore the Blair Shale is interpreted to occupy the hollow between Knockgardner [NS 351 034] and the Green Hill of Knockgardner [NS 353 033]. Thin-bedded, fine-grained, grey sandstones, fining upwards from sharp bases, and intercalated siltstones and mudstones exposed in the Cawin Burn [NS 373 042], are considered to belong to this formation. These beds dip about 50° to the south-east and young to the north-west. Some cross-stratification and bioturbation is reported (Clarkson et al., 1998).

Straiton Subgroup

The Straiton Subgroup (Table 1) occurs only within the Girvan main outcrop and comprises the Knockgardner Sandstone Formation overlain by the Straiton Grits Formation (Floyd and Williams, 2003). The Subgroup is unconformably overlain by the Kinnesswood Formation (Inverclyde Group).

Knockgardner Sandstone Formation (KNG)

The Knockgardner Sandstone Formation was defined in the Knockgardner area (Cocks and Toghill, 1973) as the Knockgardner Formation, and the type locality is a disused small quarry about 360 m east of Knockgardner at [NS 3550 0358] (D Palmer in Aldridge et al., 2000). In the revision of the Silurian rocks of the Girvan area (Floyd and Williams, 2003) the formation was termed the Knockgardner Sandstone Formation.

The Knockgardner beds were described by Lapworth (1882), and Peach and Horne (1899) recognised the Wenlock age of the sequence largely through their collections of shelly fossils at the quarry near Knockgardner. Cocks and Toghill (1973) confirmed the early Wenlock age of the formation, which was supported by Clarkson et al. (1977) in their study of the Silurian phacopid trilobites from Scotland. The early Wenlock age was also supported by Dorning (1982) and Wellman and Richardson (1993) on the basis of palynological studies.

The formation typically comprises thinly-bedded, grey-green or khaki weathered sandstones, usually less than 0.2 m thick, but locally thicker beds are present. Most of the sandstones are fine to medium grained, but coarse clasts are common towards the bases of beds. Interbedded siltstones and mudstone are sparse. The sandstones are sparsely shelly and in places, form a basal coarse clastic fraction. The fossils are commonly casts, which leave moulds in the sandstone. Interbedded fine- and medium-grained grey-green sandstones and mudstones are exposed along Shiel Burn at [NS 3304 0269].

The formation was estimated to be 137 m thick east of Knockgardner (Cocks and Toghill, 1973; 141 m according to Clarkson et al., 1998), where it is unconformably overlain by the Kinnesswood Formation. However, the Knockgardner Sandstone Formation is probably over 200 m thick farther north-east [NS 372 044] where it is succeeded, probably conformably, by the Straiton Grits Formation. An isolated and partly faulted inlier of Knockgardner Formation east of Glengee Wood at [NS 309 022] is overlain by Ballagan Formation.

The lithofacies of the Knockgardner Sandstone Formation was divided into three (Clarkson et al., 1998). The lowest facies, about 40–50 m thick, contains dark shales and graded sandstones with shelly accumulations and trace fossils. The middle facies is up to 100 m thick has an equal proportion of sandstone and shale. The sandstones up to 0.16 m thick, show delayed grading and cross-bedding. The uppermost facies, 30–40 m thick, contains thin-bedded shales and siltstones. The shales are dominant and locally display desiccation cracks and vertical and horizontal trace fossils.

The basal beds appear transitional with the older Blair Shale Formation (e.g. Dorning, 1982), but a tectonic contact was recorded (Clarkson et al., 1998) around Blair Farm, partly to explain the apparent time gap of two graptolite biozones if the Knockgardner Sandstone Formation is basal Wenlock centrifugus Zone (Cocks and Toghill, 1973). However, the lower zones are not necessarily missing (Floyd and Williams, 2003).

In the old quarry (now mainly in-filled) 360 m east of Knockgardner [NS 3550 0358], most beds are overturned and young to the north-west on the evidence of graded bedding. This is the fossiliferous locality where Cocks and Toghill (1973) collected shelly fossils in intercalated sandstones, siltstones and mudstones. The fauna comprised 42 per cent Protochonetes sp., 13 per cent Atrypa reticularis (Linnaeus), 13 per cent Howellella elegans (Muir-Wood) and 7 per cent Resserella cf. basalis (Dalman) as well as five other brachiopod species and 14 other taxa. Another collection by Dr I Penn from the same quarry (Cocks and Toghill, 1973) included a similar brachiopod assemblage as well as polyzoa and ostracods. Peach and Horne (1899 p. 550) give a list of fossils from this locality which, apart from brachiopods and ostracods, contains 'Ptilodictya lanceolata (Goldfuss)' 'Favosites sp.', 'Cornulites sp.', 'Orthoceras sp.' and several trilobite species.

Another locality in a small burn east of Dyke Farm [NS 363 039] exposes similar, near vertical, turbiditic strata, which yielded 'Ptilodictya', 'Orthoceras MacLareni Murchison', 'O. angulatum (Wahlenberg)' and crinoid stems (Peach and Horne, 1899 p.550).

The faunal assemblages suggested a lower Wenlock age, and although it was not possible to be more precise, they do not occur in Llandovery or upper Wenlock strata (Cocks and Toghill, 1973). The Howellella–Protochonetes brachiopod assemblage is considered to originate in a fairly shallow water shelf (Cocks and Toghill, 1973).

The trilobites, Hemiarges rolfei Lamont, Podowrinella straitonensis (Lamont) and Encrinurus sp. have been recorded from the formation (Clarkson et al., 1977; Clarkson and Howells, 1981). The trilobite fauna was detailed in a monograph by Howells (1982), and the recent reassignments of Encrinurus to Wallacia hagshawensis (Lamont) and Hemiarges rolfei, now placed within Richterarges, are referred to by Clarkson et al. (1998). The basal Wenlock biofacies around Girvan was also reviewed by the latter authors, and two faunal associations were recognised in the assemblages from the Knockgardner Sandstone Formation (Clarkson et al., 1998). The first is the characteristic Podowrinella fauna of four trilobite species (Podowrinella straitonensis, Wallacia hagshawensis, Richterarges rolfei and Warburgella (Warburgella) capetos Howells) that always occur as a resedimented assemblage within the Silurian inliers in the southern Midland Valley. This specialised and mobile assemblage, endemic to the Midland Valley, was adapted for life in shallow water high-energy conditions. The second association was the Howellella-Protochonetes assemblage of Cocks and Toghill, (1973). The four main elements of this assemblage are Protochonetes sp., Atrypa reticularis, Howellella elegans and Resserella cf. basalis. By contrast, this was a sessile benthic brachiopod assemblage that is much more cosmopolitan. The assemblage was collected mainly from three horizons within the former quarry east of Knockgardner at [NS 3550 0358]. Besides the fauna already mentioned above, Clarkson et al. (1998) described the brachiopods Craniops sp., Resserella canalis (J de C Sowerby), Dicoelosia sp., Salopina sp., Isorthis, Leptaena and Rhynchotreta. Ostracods, bryozoans, crinoids, molluscs, corals and gastropods were also present. These fossils have a wide spatial and temporal distribution within mid Silurian deposits, not just in the Midland Valley but on both sides of the closing Iapetus Ocean. Clarkson et al. (1998) suggested that the differences between the two faunal assemblages reflected contrasting reproductive strategies and larval longevities. The Knockgardner sequence is the only one in the Midland Valley inliers that indicates that fully marine conditions continued into the Wenlock.

This brachiopod assemblage was compared with a similar, but younger, brachiopod assemblage from the Anglo-Welsh area that developed at the interface between marine and nonmarine facies prior to the deposition of continental lithofacies (Clarkson et al., 1998).

Dorning (1982) examined two samples from the old quarry east of Knockgardner [NS 3550 0358] for palynomorphs. They contained similar acritarch assemblages, accompanied by rare miospores, chitinozoans and scolecodonts, which he compared to early Wenlock acritarch assemblages in the Welsh basin and Gotland. The earliest assignable age is very latest Telychian (Llandovery), and the majority of the species are recorded from the early Sheinwoodian (Wenlock). Dorning (1982) concluded that the Knockgardner Sandstone Formation was earliest Wenlock in age and that it was deposited in shallow marine conditions. The rare miospores, Ambitisporites (Dorning, 1982) and Tetraheraletes medinensis (Wellman and Richardson, 1993) suggest that there was a lack of diversity to the vegetation and/or arid conditions, possibly at some distance from the site of deposition on the coast. As a whole this assemblage appears to be different from the sphaeromorph–Moyeria–Tylotopalla assemblage ( spiralis Graptolite Zone) recorded by Molyneux et al. (2008) for the Drumyork Flags Formation.

The Knockgardner succession was described as turbiditic (Cocks and Toghill, 1973), but subsequently considered by Clarkson et al. (1998) to be a shallow water, storm influenced, deposit, grading up into shallow lagoonal facies. Their Facies 1 reflected shelf–prodelta deposition. Facies 2, exposed in the quarry east of Knockgardner, where the proportion of sandstone to shale is about equal, was interpreted as intertidal facies. The sandstones are commonly cross-stratified or lenticular and well defined, delayed grading was observed (Clarkson et al., 1998). Their Facies 3, exposed in 'a burn near Craigens Wood', comprises mainly shales with siltstones and few sandstones in a fining-upward sequence deposited in delta-front to lagoonal conditions. Thus, the Knockgardner facies is fully marine, although shallow water conditions were prevalent at the time of deposition.

In thin section (N2901), a sandstone from the disused quarry east of Knockgardner at [NS 3550 0358] contains clasts of metamorphic, igneous and sedimentary rocks. The igneous clasts include fine-grained basalts, intermediate rock types and granitic types with granophyric and graphical intergrowths of quartz and feldspar. Aggregates of chloritic minerals may be derived from serpentinite. Metamorphic polycrystalline quartzite, psammite, mica schist, mylonite and garnet are present. Monocrystalline vein quartz, carbonate and crushed shell fragments are also identifiable.

Straiton Grit Formation (STGC)

The Straiton Grit Formation (Clarkson et al., 1998; Straiton Grits of Cocks and Toghill, 1973; Straiton Grits Formation of Floyd and Williams, 2003) comprises a sequence of fine-grained to pebbly, grey turbidites, passing up overall into reddish brown, coarse-grained sandstones and conglomerates.

The lower part of the formation is exposed in a tributary to the Water of Girvan west of Straiton [NS 371 046], dipping very steeply to the south-south-east. This part of the formation consists of greenish and mid-grey, interbedded wacke sandstones, with thin siltstones and silty mudstones similar to the underlying Knockgardner Sandstone Formation; however, there are also interbeds of coarse-grained, granular and pebbly sandstones. The coarse-grained or 'gritty' sandstones are 0.4–0.8 m thick; the pebbly beds are thicker and may form lenses. Subrounded to angular clasts up to 60 mm in length were observed, including dark and greenish grey chert, pinkish felsite, vein quartz, microgranite and hard 'greywacke' sandstone. The beds young towards the north-west on the evidence of grading and sharp bases. Siltstone and silty mudstone intervals are commonly thinly bedded or interlaminated. At one locality [NS 3697 0459] in the burn west of Straiton, a restricted assemblage dominated by sphaeromorph acritarchs with rare Diexallophasis denticulata, Micrhystridium sp., scolecodonts and the cryptospore Tetrahedraletes medinensis was identified (Molyneux et al., 2008).

Within the lower part of the formation, in an old quarry north of the road from Knockgardner to Kirkbride at [NS 3445 0362], green, coarse-grained sandstones in stratigraphical continuity with red, coarse-grained sandstones and conglomerates, were found to contain a sparse fauna of beyrichiacean ostracods and bivalves (Peach and Horne, 1899 p.549; Floyd and Williams, 2003). The fauna listed by Peach and Horne (1899) was identified as Silurian rather than Devonian (Cocks and Toghill, 1973), and interpreted by them as a shallow, perhaps brackish or even fresh-water assemblage. Dorning (1982) collected grey-green silty sandstone from this quarry and recovered the acritarchs, Diexallophasis denticulata, Leiosphaeridia spp., Lophosphaeridium sp., and Veryhachium trispinosum (Eisenack) Stockmans & Willière. He interpreted this low diversity assemblage as indicating very shallow marine conditions. The whole fauna was suggestive of very shallow marine or brackish water (lagoonal?) conditions (Floyd and Williams, 2003).

At a second disused quarry [NS 3453 0368], about 60 m stratigraphically higher, near-vertical conglomerates, pebbly sandstones and coarse- to medium-grained sandstones with rare shale partings were reported (Cocks and Toghill, 1973; Clarkson et al., 1998). Some conglomerates contained mud and silt clasts, and others had clasts dominated by quartzites, quartz and jasper (Clarkson et al., 1999).

The assemblages from the lower part of the formation suggest a marginal marine environment (Molyneux et al., 2008), consistent with the record of a sparse and restricted fauna of ostracods and bivalves (Peach and Horne, 1899; Floyd and Williams, 2003), which also suggests very shallow marine or brackish water, possibly lagoonal conditions.

'Purple grits and purple-green shales' considered to be from the basal part of this formation were described by Cocks and Toghill (1973) from the Shiel Burn at [NS 3281 0274], lying unconformably below the Kinnesswood Formation.

The upper part of the formation is exposed in Craigens Wood at [NS 3727 0491] and north of the Old Bridge of Blairquhan [NS 3740 0529]. These beds include reddish brown medium- to coarse-grained, 'gritty', lithic sandstones with scattered angular to sub-angular granules and pebbles. The beds are poorly sorted and typically 0.3–0.7 m thick. The pebbles, generally 10–20 mm in length, include vein quartz, red chert or jasper, pink 'felsite' and hard purple-brown sandstone.

Cocks and Toghill (1973) and Clarkson et al. (1998) estimated the formation to be 90 m thick. In the Craigens area [NS 370 047], however, more of the formation is exposed below the unconformity with the Kinnesswood Formation, and it is estimated that it is at least 500 m thick there.

Two sedimentary facies were recognised in this formation which may interdigitate (Clarkson et al., 1998). Facies 1 consists of an upward-thickening sequence of coarse- to fine-grained sandstones, representing quiescent deposition subject to influxes of coarse-grained sediment; probably in coastal lagoons subject to unusually high tides and storm-driven swash. Fluviatile flooding may have contributed to the coarser-grained deposits. Facies 2 is characterised by conglomerates and pebbly sandstones, and it was suggested that they represented back-barriers developed on the shoreside of the lagoons.

A thin section (N2900) from a coarse-grained to pebbly sandstone in the lower part of the formation west of Straiton at [NS 3713 0461] contains mainly igneous and metamorphic rock clasts. The suite of igneous rocks includes basalt, dolerite, gabbro and chlorite/carbonate aggregates possibly after serpentinite. Intermediate, and to a lesser extent, granitic rock fragments are also present, together with metamorphic polycrystalline quartz, psammite, chloritic schist and garnet. Carbonate grains are present and the polycrystalline types were probably original clasts.

Another thin section (N2895) of medium-grained red-brown sandstone from near the Old Bridge of Blairquhan [NS 3740 0529] is better sorted, more mature with iron oxides coating the grains and less carbonate. The clast content includes more monocrystalline quartz and feldspar, but also more acid-intermediate igneous and metamorphic rock fragments than the older Silurian strata in this area.

Chapter 4 Siluro-Devonian

Lanark Group

The Lanark Basin formed during Siluro-Devonian times as a graben in the southern part of the Midland Valley Terrane, developing during sinistral strike-slip movement on basin-bounding faults within the terrane (Phillips et al., 1998). Traditionally the Siluro-Devonian strata within the Midland Valley were called the Lower Old Red Sandstone (e.g. Eyles et al., 1949), but more recently the term Lanark Group (Smith, 1999a) has been applied to the basin fill. At the base of the group the Greywacke Conglomerate Formation (Table 3) is an alluvial fan facies which thins westwards, and appears not to have extended as far as the Craighead Inlier in the south-west of the Ayr district. The overlying Swanshaw Sandstone Formation forms a substantial part of the Lanark Group in the district, and comprises a thick, predominantly fluvial, sandstone and conglomerate sequence, including the first record of aeolian facies within sedimentary rocks of this age in the Midland Valley (Smith et al., 2006). The latter formation is succeeded in Ayrshire by the Carrick Volcanic Formation (Table 3), which mainly comprises a calc-alkaline basaltic andesite suite. This formation is likely to be Early Devonian (Lochkovian–Pragian) in age by comparison with rhyolite in the Pentland Hills which gave a Rb–Sr age of 412.6 ± 5.7 Ma (Thirlwall, 1988). Early indications of this volcanic episode occur towards the top of the Swanshaw Sandstone Formation in the form of volcaniclastic beds. The top of the volcanic pile was eroded away in this district after Acadian tectonic deformation and uplift in Mid Devonian times.

Farther east, in Lanarkshire, the volcanic succession is overlain by the Auchtitench Sandstone Formation (Table 3), which includes much volcaniclastic debris. To the west, offshore, the Lanark Group extends for a short distance before apparently being overlain by Carboniferous strata or faulted against Permo-Triassic strata by the Carrick Fault, which lies only 650 m offshore from Turnberry Point.

Early sedimentological models for the Siluro-Devonian strata of the Midland Valley emphasise the importance of elongate basins filled by coarse, marginal alluvial fans and finer grained, axial braided river complexes (e.g. Bluck, 1978). In the northern Midland Valley, the development of continental fluvial basins, such as the Strathmore Basin, involved complex river systems influenced by a south-westerly axial flow (Bluck, 2000). In contrast, the sedimentology of the Lanark Basin has been less well studied; although combining observations from the Swanshaw, Duneaton and Auchtitench formations (Table 3), Bluck (2000) suggested a similar south-westerly axial flow by large river systems may have formed part of the same drainage system as that of the Strathmore Basin in the north-east. Study of the fluvial palaeoflows in this district (Jones, 2001; Smith et al., 2006) indicated that while westerly flow is dominant, there is also a component to the north-west.

Greywacke Conglomerate Formation (GRWC)

The basal coarse conglomerate lying unconformably above the Ordovician Benan Conglomerate Formation, seen around [NS 391 010], is assigned to the Greywacke Conglomerate Formation by correlation with the succession in Lanarkshire (Smith, 1999a; Browne et al., 2002). It comprises variable thicknesses of poorly sorted conglomerate mainly containing greywacke sandstone pebbles and was deposited as an alluvial fan. In this area, typical conglomerate is exposed in a small burn to the south of Genoch Farm [NS 3911 0065]. The conglomerate contains purple-grey and brown, rounded to subrounded pebbles and cobbles of mainly wacke sandstone in a sandstone matrix. Considering that the conglomerate overlies the Benan Conglomerate, it does not obviously contain pebbles derived from the older conglomerate. The formation is estimated to be about 150 m thick here, but it most probably interfingers with the overlying Swanshaw Sandstone Formation.

The Greywacke Conglomerate Formation and the conglomerates in the Swanshaw Sandstone Formation have been studied by Syba (1989), who concluded that the conglomerates were deposited in small proximal alluvial fans in north-south trending half grabens, with the controlling faults on the east side of the basins. A detailed petrological study of the wacke clasts from conglomerates in the Straiton area (Syba, 1989) indicated that they are highly quartzose, and rich in low-grade metamorphic detritus (mostly schist and quartzite). The sources are not readily identifiable in the Southern Upland Terrane and the palaeocurrent evidence from the conglomerates indicates dispersal from an easterly (and possibly southerly) source, possibly within the Midland Valley Terrane. This model is supported by the fact that the formation is not present to the west, where sandstones of the Swanshaw Sandstone Formation rest unconformably above the Craighead Inlier and the inlier of Ordovician Benan Conglomerate at [NS 330 007]. Presumably fans of Greywacke Conglomerate did not extend that far west, although there is a local lens (former channel?) of quartz-felsite conglomerate within the Swanshaw Sandstone Formation overlying the Llandovery Saugh Hill Grit Formation at [NS 286 053].

Swanshaw Sandstone Formation (SWAS)

The Swanshaw Sandstone Formation (Browne et al., 2002) is predominantly a succession of reddish brown to pinkish buff, lithic arenites with subordinate mudstones, pebbly sandstones and conglomerates. The formation is a sandstone-dominated fining-upwards siliciclastic succession, deposited in a terrestrial environment. Lithic sandstones form approximately 65 per cent of the succession, with lesser amounts of conglomerate (about 25 per cent) and mudstone (about 10 per cent). The sandstones are typically medium grained and moderately well sorted, with detrital clasts that are considered to have a recycled orogenic provenance (Phillips et al. 1998).

South of Ayr, the formation forms a substantial part of the Lanark Group (Figure 4), and comprises a thick and locally well-exposed sequence. The sandstone beds are mainly planar or cross-bedded and fluviatile in origin, but include the first record of aeolian facies within sedimentary rocks of this age in the Midland Valley (Smith et al., 2006).

A definitive biostratigraphical age for the formation is lacking but a poor spore assemblage, including Apiculiretrusispora and Synorisporites sp., recovered from the formation on the coast north of Isle Port at [NS 245 133], suggests a Siluro-Devonian age (Stephenson, 1999c). Radiometric Rb-Sr dating of the overlying volcanic rocks gives an age of around 410–415 Ma, which is just above the Silurian–Devonian Boundary (Thirlwall, 1988).

Although the Swanshaw Sandstone Formation is lithologically similar both north and south of the Straiton Fault, the sequences and thicknesses are different. For instance, south of Straiton at [NS 390 007], the Swanshaw Sandstone Formation overlies the Greywacke Conglomerate Formation (Table 3). However, north of the Straiton Fault, sandstone facies unconformably overlie the Lower Palaeozoic rocks of the Craighead Inlier.

Also to the north of the fault, several members have been recognised within the formation including the Chapelton Burn Member (Table 3), characterised by aeolian facies (Smith et al., 2006). In the upper part of the formation, the Crane Dyke Member includes mudstone facies (Smith et al., 2006) and is directly overlain by basaltic andesite lavas of the Carrick Volcanic Formation (Table 3), (Figure 4), as exposed on the coast west of Drumshang at [NS 2449 1382]. Within the Swanshaw Sandstone Formation, near the Mochrum vent (Figure 4), lenses of volcaniclastic sandstone are the products of earlier, penecontemporaneous volcanic activity. The petrographical characteristics of the Swanshaw Sandstone Formation are described by Phillips (2001).

In the coastal area, the thickness of the Swanshaw Sandstone Formation is approximately 870 m, and it appears to thicken laterally to the north-east. The formation appears to be less thick south of the Straiton Fault.

The resurvey was complemented by the later detailed sedimentological work by Jones (2001). Channel, aeolian, floodplain and mud-flat facies in the Swanshaw Sandstone Formation, distinguished mainly on the basis of lithology, bedding and sedimentary structures, are combined into genetically or environmentally related associations (Smith et al., 2006). Palaeocurrent data collected from the formation are summarised on (Figure 5), and this information together with an interpretation of the depositional facies is synthesised into a sequence of palaeogeographical phases during the accumulation of the formation (Figure 6).

The Channel facies association (facies association A) comprises conglomerates and sandstones associated with river channels, and it includes facies A1 to A8 detailed in Smith et al. (2006). The typical exposures mentioned in the text are located on (Figure 4) and examples of the facies are shown in (Plate 3), (Plate 4), (Plate 5). Channel facies can be recognised by either the preservation of a lenticular sediment body (the channel form) or the identification of erosive bounding surfaces (Miall, 1996) that separate individual channels (Plate 6). These surfaces are generally laterally continuous, flat to slightly concave-upwards, and are characterised by local cut-and-fill relief and by basal lags. Where successive channels erode down into earlier forms, individual units can be termed a storey, and the vertical stacking of these produces the multistorey complexes commonly seen in the Swanshaw Sandstone Formation.

The aeolian facies association (Association B of Smith et al., 2006; (Plate 7), (Plate 8), (Plate 9), comprises three distinct facies which have only been recognised in the Chapelton Burn Member (Table 3). Aeolian sandsheets may form up to 40 per cent of the member. This restricted occurrence is presumed to be because the preservation potential of characteristic aeolian features, such as pinstripe lamination, was low in this dominantly fluvial environment. The aeolian facies have a close relationship with the floodplain facies, as the aeolian dunes formed on the floodplain during dry periods.

Floodplain facies Association C (Smith et al., 2006; (Plate 10)) of sandstone and mudstone facies, deposited on floodplains during short-lived floods, is volumetrically less important than the channel facies. Floodplain facies interdigitate with both channel and aeolian facies, the latter having been partly reworked.

The depositional environment for the mud-flat Association D of Smith et al., 2006; (Plate 11) and (Plate 12) is similar to that of the sandflat, in that it is a low gradient plain characteristic of semi-arid and arid landscapes. However, the main difference is the sand to mud ratio, with mud deposition dominant (Reading 1996). This facies association has only been recorded in the Crane Dyke Member, near the top of the formation (Table 3) on the coast north of Isle Port.

Southern sections (Dipple to Culzean Castle) including Chapelton Burn

The coastal sections of this formation are the best in the district, and the sedimentology of the formation has been studied in detail (Jones, 2001, Smith et al., 2006). The base of the formation is marked by an unconformity over the Ordovician–Silurian sequence of the Craighead Inlier, but it is not exposed in the coastal section. The top of the formation is exposed at the contact with the Carrick Volcanic Formation, on the coast west of Barwhin Hill [NS 2190 0945]. Good exposures can be found within deeply incised burns, for example Chapelton Burn [NS 2177 0459] and Milton Burn [NS 210 051], along coastal wave-cut platforms for example north Dipple Factory [NS 1933 0306], and forming cliffs in the Culzean Castle area e.g. [NS 2385 1053].

The lower part of the Swanshaw Sandstone Formation, as seen in the Dipple area, is dominated by red to buff, coarse- to medium-grained sandstones, with trough- and cross-bedding. The beds are typically well sorted micaceous, feldspathic arenites. Clast-supported conglomerate units, up to 10 m thick, appear more common towards the base of the formation.

The sandstone sequence contains stacked, cross-bedded units, up to a few metres thick, bounded by erosion surfaces and with little lateral continuity. Basal lags of microconglomerate, a few centimetres thick, are quite common just above the erosion surfaces.

Distinct beds of clast-supported conglomerate, composed of rounded clasts, were observed in sharp, erosionally based, units approximately 0.5 m to 10 m thick. This rock type passes upwards into parallel- and trough-bedded red sandstones and examples of the conglomerate crop out west of Dowhill Farm [NS 1993 0306] and Drumuskin Bridge [NS 2115 0212].

Jones (2001) described these two rock types in more detail and interprets them as sand-filled channel and conglomerate-filled channel facies respectively, within a fluvial channel facies association. These belong to the Channel facies association A (Smith et al., 2006), and the facies are well exposed on the coastal section [NS 1990 0400–1980 0330] near Dowhill Port, the lower part of lower Chapelton Burn section e.g [NS 2177 0459] (Figure 4) and at Culzean [NS 2385 1053]. Palaeocurrents (Figure 5) derived from cross-bedding indicate flow to the west on the coastal section, and to the north-west and south-west inland. This facies is interpreted as prograding bedforms (dunes) deposited in a river bed, possibly within a major, sand-dominated channel system. These observations are in agreement with previously published work (Bluck, 2000).

Chapelton Burn section — including the Chapelton Burn Member

In the upper section of the Chapelton Burn (Figure 4), interbedded white-grey and red mottled siltstones are interbedded on a 0.1–0.2 m scale with sharp-based, laterally extensive, white-buff, fine-grained sandstones e.g. [NS 2188 0470]. Mottling and other faint traces of semi-vertical, pale grey, cemented structures were interpreted as possibly incipient calcrete development. Together, these deposits were interpreted as overbank fluvial deposits with periodic crevasse splay activity and some palaeosol development. The rare multiple sets of ripple cross-laminated fine-grained sandstone (e.g. at [NS 2216 0476] along Chapelton Burn), preserved at the tops of abandoned channels, are interpreted as low energy deposits from unidirectional turbulent flows (Channel facies A6 of Smith et al., 2006).

Chapelton Burn Member (CBUR)

Distinctive white aeolian sandstones are also mapped (Monaghan, 2001) in the upper section of Chapelton Burn (e.g. at [NS 2287 0490]). The moderate- to well-sorted sandstones contain coarse, rounded, frosted grains in a poorly sorted fine- to medium-grained sandstone matrix. Small cross-bedded units, up to 0.13 m thick, have distinct grainflow laminae (Plate 7), and some sand sheets, up to 0.1 m thick, have an internal 'pinstripe' lamination (Plate 8). The pinstripe lamination is an even, millimetre-thick alternation of fine and medium-grained sandstone in which the individual fine and medium pinstripe couplet defines a wind ripple set.

Other thinly-bedded sandstones have a wavy to convoluted nature to the internal laminae that are attributed to aeolian deposition on a damp sandsheet (Plate 9). Wave-reworked units of sheet-flood sandstone (Floodplain facies) are interbedded (Plate 10). Jones (2001) described these rock types in more detail, and distinguishes crevasse splay, sheet-flood, ephemeral lake, palaeosol and aeolian sandsheet facies. Cross-bedding in the aeolian beds indicates that palaeocurrents flowed to the east (Figures 5 and 6), distinctly different than that of the fluviatile sandstones and the crevasse splays.

The distinct characteristics of these aeolian (B1–B3) and floodplain (C1–C4) facies association rocks (Smith et al., 2006) make it possible to distinguish the Chapelton Burn Member. The type section of this member is defined in Chapelton Burn (from [NS 2200 0472] to [NS 2297 0483]), with a minimum thickness of 145 m.

Culzean Bay to Isle Port section

Sandstones in the middle part of the formation are exposed in the Swallow Craigs cliff [NS 2425 1081], Beggar's Knowe [NS 2400 1070], Dead Knowe [NS 2466 1208] and in Rancleugh Burn [NS 261 114]. This part of the formation extends up to the base of the Isle Port Conglomerate Member, near Croy, and comprises predominantly medium- and medium- to coarse-grained, quartzo-feldspathic to lithic buff-grey sandstones. The sandstones, which weather buff-brown, are commonly cross-bedded in sets between 0.4–0.8 m thick, exceptionally 0.2–1.2 m thick. Trough cross-bedding and low-angle cross-beds are common (Plate 5); planar cross-beds are rarer. These multistorey cross-bedded units belong to Channel facies A5 (Smith et al., 2006).

Pale, greenish grey mudstone clasts (10–250 mm in length) are numerous, and in places concentrated on lag surfaces and foresets. Small sandstone-filled channels have erosive bases overlain by mudstone clasts. Scattered subrounded, small (up to 10–20 mm long), lithic pebbles, including reddish chert, microgranite and vein quartz, also occur within the sandstones. Subordinate interbeds of purplish to red-brown, fine- and medium-grained micaceous sandstone are in the order of 0.1–0.2 m thick.

In better exposures on the foreshore, micaceous, fine-grained brown sandstone passes up into red-brown silty mudstone with calcareous concretions. These and local red-brown or green mottled silty mudstones, with disrupted mud-cracked layers, belong to the floodplain facies association.

Volcaniclastic sandstones occur inland, north-east of the Mochrum Hill vent rocks (see below), and they may be derived from the vent. These khaki/brown volcaniclastic rocks contain numerous small greenish and buff weathered basaltic fragments, set in a medium-grained sandstone matrix. The rocks are poorly- to moderately- sorted, with scattered small angular to subangular pebbles (up to about 10 mm in length) with darker fine-grained rims (possibly originally lapilli). In the Brockloch area, exposed in Brockloch Burn north-west of Maggie's Linn [NS 2955 1165], interbedded medium- and coarse-grained to pebbly volcaniclastic sandstones contain altered pale greenish fine-grained volcanic fragments set in a brownish, medium-grained sandstone matrix. The strata dip up to 56° to the north-west, and their steep dip may be due either to strike faulting, or to synsedimentary volcanic disturbance, similar to the disturbance around the Mochrum vent. Similar rocks were formerly exposed north of Ladycross Wood [NS 2896 1111] where dips were reported as near vertical in the previous geological survey. The area is poorly exposed and small outcrops of similar lithology continue for at least 1.5 km north-north-west of Ladycross Wood, but no clear interbedded relationships with the surrounding Swanshaw Sandstone Formation can been seen.

Isle Port Conglomerate Member (ISP)

The overlying Isle Port Conglomerate Member is well exposed on Croy foreshore [NS 245 129]. However, its base is exposed (and defined) in the tributary to Craigencroy Glen [NS 2578 1303], where a pebble conglomerate (more than 1 m thick) sharply overlies red-brown fine- and medium-grained sandstone. The top of the member is exposed in the cliffs north-east of Fairy's Well [NS 2469 1316]. The member is estimated to be about 30 m thick.

The clast-supported pebble conglomerate, in which the coarser grains are sorted and weakly bedded at its base, includes lenses of reddish brown medium grained sandstone. The pebbles and cobbles generally vary from subangular to rounded and are commonly 10–100 mm in length. The matrix is a medium- to coarse-grained, red-brown to greenish mottled sandstone, but where the matrix is sparse or lacking, the voids have been infilled with sparry calcite. The conglomerate contains a range of clasts, but fine- to medium-grained, well-sorted greywacke sandstone is the most common clast type. Other clasts include grey-green, black, red and brown chert, jasper, grey siltstone, coarse to granular sandstone, vein quartz, metaquartzite, fine-grained, red-brown sandstone, pinkish microgranite, reddish granite, purplish grey, aphyric, siliceous and amygdaloidal ?rhyolite and fine-grained, pale limestone. This conglomerate member contained the limestone clasts from which Armstrong and Owen (2000) and Dean (2000) obtained Pygodus anserinus conodonts (mid Caradoc), and they suggested that as these are similar in age to those in the Stinchar Limestone, the Girvan succession situated to the south could be the source of the clasts. Smith (1895) had previously found two fossils in limestone clasts (up to 0.3 m in length) from the member, and concluded that the clasts were 'Silurian' in age.

Near the base of the member, coarse pebble to boulder (up to 0.3 m in length) conglomerate beds, 1–2 m thick, are commonly massive. Horizontally bedded conglomerates are interbedded with greyish, locally laminated sandstones, up to 0.3 m thick (Plate 3). These and trough cross-bedded units, about 1 m thick (Plate 4), all belong to the channel facies association (Jones, 2001; Smith et al., 2006), with an example of its multistorey nature shown in (Plate 6).

Rounded greywacke sandstone pebbles dominate some of the coarse conglomerate beds, whereas some smaller pebble, gravelly conglomerates contain predominantly angular to subangular clasts of chert and siliceous rhyolitic rocks. In places, the conglomerate beds both fine- and coarsen-up, with imbrication particularly evident towards the tops of the beds. The palaeocurrents are predominantly towards the north-east. Small-scale channelling and convolution of the sandstone interbeds is present. Towards the top of the member, there is an increasing proportion of intercalated sandstone, including bimodally sorted pebbly sandstones. Low-angle foresets in the cross-bedded pebbly sandstones dip towards the north. These beds are cut by small normal, early, possibly synsedimentary faults, trending 052° and dipping 70° south-east.

The member can be mapped to the east of Knoweside Hill but no farther south due to lack of exposure. The conglomerate may pass laterally into pebbly sandstone and in a small disused quarry [NS 2654 1133] extrabasinal pebbles, including microgranite, are present within the Swanshaw Sandstone Formation. East of the Carrick Hills, a thinner conglomeratic sequence (which may lens out in places) is exposed at about the same level within the Swanshaw Sandstone Formation, for instance 100 m south of Glenbay [NS 2814 1371], and it is mapped as equivalent to the Isle Port Conglomerate Member. The member consists of clast-supported, sharp based, pebble conglomerates, up to 3 m thick, interbedded with medium- to coarse-grained and pebbly sandstones near Beoch at [NS 2935 1458]. The pebbles at this locality are generally 10–20 mm, but some are cobbles up to 80 mm in length. The rounded/subrounded to angular pebbles comprise red, grey and black chert, jasper, quartzite, vein quartz, pink microgranite and fine-grained, grey and reddish sandstone. Greywacke sandstone is a less common clast component than at Isle Port and no limestone pebbles were observed. A pebble conglomerate unit, about 2 m thick, exposed in Brockloch Glen [NS 2958 1163] is clast-supported and contains a similar suite of clast types, including rare purplish grey, amygdaloidal, basaltic andesite. However, because this section is faulted, it is not certain if this unit is equivalent to the Isle Port Conglomerate.

The conglomerates are interpreted as the distal parts of fanglomerates, built out from the east (Figure 6) in response to relative uplift in the hinterland to the south-east, as extensional/transtensional faulting within the Midland Valley took place prior to the eruption of the Carrick Volcanic Formation. Conglomerates are found at a similar horizon within the Swanshaw Sandstone Formation in the New Cumnock district (Smith, 1999a).

Crane Dyke Member (CRDY)

The Swanshaw Sandstone Formation succession above the Isle Port Conglomerate, and below the Carrick Volcanic Formation, is about 70 m thick, and has been established as the Crane Dyke Member (Smith et al., 2006). At its base, trough cross-bedded, greyish, medium-grained lithic sandstones are medium bedded, with a few interbeds of pebbly and fine-grained sandstone. The succession passes up into purple-brown and grey-green, fine- to medium-grained sandstones, which are mottled in places and contain scattered micas. Sporadic interbeds, about 0.1 m thick, of greenish grey and red-brown banded mudstone and siltstone are associated with angular to rounded mudstone clasts, which occur scattered or concentrated in lags in the adjacent sandstones. Locally the mud clasts have greenish reduced margins. One sample of brownish mudstone from [NS 2455 1334] yielded a very sparse palynomorph assemblage, including a few specimens of Retusotriletes and Apiculiretusispora, and a single specimen of Synorisporites sp. The latter is characteristic of late Silurian to Early Devonian assemblages (occurring rarely in the Late Devonian). Based on its occurrence, the sample is most likely to be late Silurian to Early Devonian in age (Stephenson, 1999c). Beaconites burrows (Jones, 2001) are recognised in the mud-flat facies of the member (Plate 12). This is the first description of this trace fossil from the Lanark Basin, although it has been recorded from the Stonehaven and Strathmore groups in the northern Midland Valley and the Siluro-Devonian of the Scottish Highlands and Islands (Allen and Williams, 1981), including the Turriff Basin (Trewin and Thirlwall, 2002). Beaconites occurs in and near active river channels, including overbank muds and calcretes; and although mud-flat facies are considered its least favourite environment (Allen and Williams, 1981), its preservation potential might be higher because the mud layers retained moisture and helped preserve the structure.

The overlying very fine- to medium-grained, chocolate-brown coloured sandstones are locally greenish mottled and contain calcrete (cornstone) concretions. Other, harder, sandstone beds have a calcareous cement. Besides the red-brown mudstone clasts, some sandstones contain rounded clasts of purplish, amygdaloidal basaltic andesite, up to 50 mm (coarse gravel) in length; some contain small feldspar phenocrysts. The overlying greenish grey sandstones are trough cross-bedded and locally convoluted. To the north of the dolerite dyke at [NS 2453 1358], pale brown, medium-grained sandstones are trough cross-bedded and include coarse grains of detrital white mica. These sandstones pass up into fine-grained, chocolate-brown sandstones and flaser-bedded, finely micaceous siltstones and silty mudstones with local reduction spots, calcareous concretions and pale grey lenses of calcrete. The silty mudstones and the calcretes (cornstones) have locally been broken up and resedimented in lenses of breccia. Mud-flat facies D1 (Plate 11) is interbedded with sheet flood sandstone (facies C2), and elsewhere in this member (Smith et al., 2006), silty claystone contains darker, silt-filled desiccation cracks.

A soft, brick-red, fine-grained sandstone including numerous small, scattered, rotten volcanic fragments, about 2 m thick, lies just below the Carrick Volcanic Formation. The volcanic fragments are associated with yellow/greenish reduction spots and may have been possible pyroclasts reworked into the background sandy sediment.

Other sections through the Crane Dyke Member are exposed in the tributaries to Craigencroy Glen [NS 256 133] and [NS 262 132]. West of Howmoor Cottage at [NS 2711 1189], just below the Carrick Volcanic Formation, about 2 m of soft, purple and green mottled, fine-grained sandstone is exposed containing poorly sorted volcanic lapilli or clasts, up to 0.12 m long, together with sporadic calcareous nodules and fine-grained brown sandstone clasts.

Farther east, around Craigskean [NS 2995 1509], red, micaceous, well-sorted arenites crop out directly beneath amygdaloidal, feldspar-phyric igneous rocks of the Carrick Volcanic Formation. The micaceous arenites may be laterally equivalent to the Craig Dyke Member. At Shank Glen [NS 2998 1566], medium-grained, well-sorted red sandstone is intercalated with altered amygdaloidal igneous rocks at the base of the Carrick Volcanic Formation.

Inland sections (Kirkmichael–Maybole–Minishant)

Around Maybole, the Swanshaw Sandstone Formation extends from the Kerse Loch Fault to the base of the Carrick Hills, which are capped by the Carrick Volcanic Formation. Because of limited and isolated exposure, the stratigraphy and depositional history of the formation is more difficult to establish inland than along the coast. Most of the beds belong to the channel or the floodplain facies associations, the latter commonly reworked in an active fluviatile system. The strata mainly dip to the north-north-west, so the oldest beds are presumed to be cut out against the Kerse Loch Fault.

Near Carsloe Glen (Figure 4), pink-brown, fine- to medium-grained, lithic sandstones crop out below a ridge striking east-south-east at [NS 3029 0650], and weather deeply to an orange colour. The ridge comprises a greyish and red-brown mottled, weakly bedded conglomeratic unit estimated to be about 20 m thick. Set in a red-brown sandy matrix, the clasts are commonly 10–60 mm in length and rounded to subrounded, although those of chert tend to be angular. The pebble content includes green, red, grey and black chert, fine-grained, reddish sandstone, purplish micaceous sandstone, grey sandstone (including some silicified clasts up to 0.14 m in length), pinkish microgranite, pink, medium-grained granite, purple-brown feldspar porphyry and andesitic rocks and fine-grained quartzite and vein quartz. Grey sandstone and chert are the most common clasts and a derivation by erosion of Lower Palaeozoic rocks with a southerly provenance is indicated by comparison with the Isle Port Conglomerate Member (cf. Smith, 2000a and Armstrong and Owen, 2000). The conglomerates near Carsloe Glen are, however, lower down in the formation than the Isle Port Conglomerate. In a good exposure at [NS 3042 0656] north-east of Craigfin, two fining upward units around 1.2 m thick with red-brown, fine- to medium-grained, cross-bedded sandstone lenses towards their tops and some imbrication in the cobbles and pebbles towards their bases, indicate that the palaeocurrents flowed to the south-west. Farther north at Spring Garden [NS 301 071], another, thinner, conglomeratic unit is exposed.

In the Capenoch Burn farther east [NS 302 072], khaki, grey-green to grey-brown, medium-grained lithic sandstones occur in beds 0.2–0.4 m thick with thin, ripple-marked sandstone interbeds. Beds vary in hardness due to variable calcareous cementation, and there are sparse silty mudstone intraclasts and scattered lithic pebbles. A typical medium-grained sandstone ((N4281) from Capenoch Burn) is, poorly to moderately sorted, with subangular, or less commonly, subrounded, heterolithic clasts in an interstitial carbonate matrix. Monocrystalline quartz, plagioclase feldspar, perthite, muscovite, chlorite, biotite and iron ores are common mineral constituents. Lithic clasts include basaltic and felsic igneous rocks, polycrystalline quartz, micrographical quartz and feldspar intergrowths and altered basic minerals. In the same burn, south of Capenoch Bridge [NS 303 077], a sequence of red-brown to yellow-brown mottled, fine- and very fine-grained sandstones is interbedded with thin red-brown siltstones and mudstones, dipping up to 14° to the north or north-north-east. Some of the sandstones are calcareous and pale grey or purple-brown, with common white mica flakes. Cracked surfaces occur locally on flaggy beds possibly reflecting original desiccation cracks. This interval of floodplain facies is at least 30 m thick, which is atypical of the Swanshaw Sandstone (or not usually exposed). To the south-east in a railway cutting [NS 308 072], about 100 m north of the Kerse Loch Fault, red sandstones contain pebble beds which dip 25° to the south-east. Since most of the formation dips north-north-east, the south-easterly dip is probably the consequence of the strata lying within the fault zone.

The prominent conglomerate unit, on the ridge [NS 304 081] from Drummullan to Kilhenzie Castle, is estimated to be 20 m thick. It is reddish brown, coarse, massive bedded and contains quartz, chert and igneous clasts up to 0.1 m in length. The clast-supported conglomerate beds are intercalated with medium- to coarse-grained, red-brown sandstone beds or lenses. A pebbly sample (N4298) from north-east of Drummullan contains clast-supported pebbles and granules, including fine-grained red mudstone, reddish radiolarian chert, vein quartz, quartzite, schist and psammite, poorly sorted, fine-grained sandstone with iron ore cement, altered, fine-grained rhyolitic and volcanic rock, quartz porphyry and altered basalt.

In Maybole, brown medium- to fine-grained sandstone belonging to the Swanshaw Sandstone Formation was encountered below Quaternary deposits at a site on Society Street (SE 3440 Borehole 5 [NS 30162 09793]), as well as in Littleton Farm Borehole (NS30NW/1) [NS 31322 08789] where over 25 m was proved.

In Dyrock Burn [NS 346 089], east of Kirkmichael, a short section exposes fine- to medium-grained trough cross-bedded, purple-brown lithic sandstone belonging to the channel facies association. Some flaggy and micaceous interbeds lie towards the tops of these units. In one section north of the burn at [NS 3462 0899], brown mudstone intraclasts up to 0.15 m in length were seen in the sandstone below a micaceous, calcareous sandstone interbed, about 0.2 m thick. The latter contained slightly rounded, pale green to dark red-brown mudstone intraclasts and grey micritic limestone fragments up to 60 mm across. These clasts appear to be reworked penecontemporaneous overbank muds and calcrete. Similar strata were formerly exposed in a quarry [NS 347 093] to the north. A sandstone sample with carbonate cement, from Dyrock Burn (N4292), contains coarse carbonate lithic clasts and partly altered hornblende, clasts of micaceous and quartzose schist, psammite and microdioritic rock.

East of Kirkmichael, north-west of the Kerse Loch Fault, the formation is exposed in Dyrock Burn [NS 353 091], where purplish brown, pink and locally greenish grey, fine- to medium-grained lithic sandstones are thin to medium bedded. Low-angle cross-bedding is present here, and one palaeocurrent reading towards 277° was recorded. Thin red-brown siltstones, up to 0.2 m thick, are interbedded with the sandstones. White mica is conspicuous on flaggy bedding surfaces and mudstone intraclasts within the sandstones. The intraclasts are commonly 10–100 mm in length and dark red-brown in colour, but pale greenish siltstone clasts up to 50 mm in length are also present. Here, the lower part of the formation appears to have been cut out by the Kerse Loch Fault.

In general, the Swanshaw Sandstone Formation is directly overlain by basalts and basaltic andesites of the Carrick Volcanic Formation. However, in the area between Kirkmichael and Dalrymple, the top of the formation is probably diachronous, as fine- to coarse-grained volcaniclastic sandstones are intercalated with basaltic rocks, with the lowest of the basalts taken as the base of the Carrick Volcanic Formation. The uppermost sandstones in the Swanshaw Sandstone Formation are greenish grey and volcaniclastic, becoming similar to volcaniclastic sandstones intercalated within the Carrick Volcanic Formation above. Greenish grey, thin, parallel bedded sandstones near the top of the formation are exposed in the disused quarry west-south-west of Cassington [NS 3518 1153]. The sandstones are fine to medium grained, with scattered rounded to sub-rounded clasts commonly between 10–100 mm in length. The beds are strongly bimodal in grain size although there are thin interbeds of medium- to coarse-grained volcaniclastic sandstone. The included pebbles are mainly of fine-grained volcanic rock, including amygdaloidal basaltic andesite. The beds are between 10–200 mm thick with some ripple and low-angle cross-bedding in the thicker beds. In exposures 100 m to the north, purple-brown sandstones are interbedded with greenish coloured types. South-south-west of Cassington [NS 3534 1123], thin-bedded, purplish, fine-grained sandstones crop out below the Carrick Volcanic Formation. These sandstones possess fine-grained white micas, a greenish mottling and irregular ripple-marks.

The conglomerate unit at Guiltreehill [NS 358 107] lies about 27 m below the top of the formation. The brownish grey pebble conglomerate is clast-supported in a purplish brown sandstone matrix. The conglomerate is associated with pebbly sandstone beds and minor medium-grained sandstone interbeds. The rounded to sub-rounded clasts, 10 to 100 mm long, include fine- to medium-grained wacke sandstone, brown sandstone, vein quartz and grey and reddish chert. Porphyritic basaltic andesite and vesicular volcanic rocks are rarer clast components. This unit can be correlated with the Isle Port Conglomerate exposed at Isle Port [NS 244 128], although the limestone pebbles that were observed there were not found and this conglomerate is thinner, only about 6 m thick. Another massive heterolithic conglomerate exposed at [NS 3505 1161], at least 1.4 m thick, is correlated with the one exposed at Guiltreehill.

North-east of Guiltreehill Farm [NS 3590 1074], a cutting exposes medium-grained, green and grey speckled lithic sandstones, containing pebble horizons and scattered chert and greywacke sandstone pebbles, as well as dark purplish brown mudstone intraclasts. Imbrication within one pebble bed indicated palaeoflow was to the south-west, toward 245°. Some sandstone beds fine up into interlaminated sandy siltstones and siltstones with minor small-scale cross-bedding.

In a burn south of Fardenwilliam [NS 368 106], fine- and medium-grained red-brown mottled micaceous sandstones contain small pebbles of purplish volcanic rock and locally greyish pedogenic carbonate nodules. Similar nodules are recorded from sandstones above High Smithston Bridge at [NS 322 118].

Around Minishant, exposures of the formation are scattered, mainly along watercourses, but it is well seen in the disused St Murray's Quarry [NS 304 113]. Trough cross-bedding in the quarry indicates dominant palaeocurrents to the north-north-west and at least six channel units can be defined on the basis of laterally persistent erosion surfaces marking the channel bases (Jones, 2001). Most of the sandstone exposures in the Minishant area are fine- to medium-grained and micaceous. Coarse-grained and conglomeratic beds are rarer; thin clast-supported conglomerate beds crop out in Pinmore Burn [NS 311 149] and a massive, partly matrix-supported conglomerate occurs west of the River Doon at [NS 339 135].

Sections south of the Kerse Loch Fault

A faulted wedge of purple-brown, fine- to medium-grained lithic sandstones around Cloncaird Castle [NS 358 075] is assigned to the Swanshaw Sandstone Formation in this resurvey (Smith, 2001). Previously, these sandstones were assigned to part of the Carboniferous Calciferous Sandstone Measures (Eyles et al., 1949). In the Water of Girvan [NS 3575 0759], the pink-brown and purple-brown lithic sandstones contain red-brown mudstone clasts, up to 0.15 m in length, and interbeds of micaceous siltstone. The sandstones vary from hard, calcite cemented beds, to softer beds with buff coloured zones at calcareous concretions. Cross-bedded and planar-bedded units occur. Planar cross-bed foresets dip towards 322° and 316° in units up to 0.8 m thick. This sequence appears to be unconformably overlain by pale, coarser-grained, quartzose sandstones belonging to the Kinnesswood Formation.

Sections south-east of the Straiton Fault

The Swanshaw Sandstone Formation, as exposed south-east of the Straiton Fault, appears to be a thinner, more rudaceous succession than that found to the north-west, particularly that along the Ayrshire coast. Overall, the thickness of the formation in the Genoch area [NS 390 011] is estimated to be 410 m. The base of the formation conformably overlies the Greywacke Conglomerate Formation, although the junction is not exposed in the Genoch area. Typical exposures in Genoch Burn [NS 3875 0059] comprise greenish grey, medium-grained lithic sandstones with scattered detrital white mica. Thin flaggy and cross-bedded units, 0.1–1 m thick, dip about 50° to the north-west and contain mudstone intraclasts and red-brown coloured zones. The sandstone succession also contains significant lenses of pebble conglomerate, that are structurally overlain to the west-north-west by basaltic andesite sheets forming the base of the Carrick Volcanic Formation. The intercalated conglomerates contain mainly pebbles, and more rarely cobbles, of wacke sandstone and some grey chert.

In Palmullen Glen [NS 380 015], grey, pink- and red-brown, fine- to medium-grained lithic sandstones dip at 12° north-eastwards below a thick purple to reddish brown pebble to cobble conglomerate. The conglomerate is roughly bedded in coarsening upwards units about 2 m thick, within which imbrication of the clasts indicates that palaeocurrents flowed to the north-west. Cross-bedded units in the underlying sandstones, up to 0.5 m thick indicate that local palaeocurrents flowed to the south-south-west and south-east (compatible with those inferred from the formation near the Ayrshire coast).

North of Cawin Hill [NS 3705 0311], just below the Carrick Volcanic Formation, a pebble conglomerate unit about 50 m thick, contains generally clast-supported, rounded to subrounded pebbles. The pebbles, generally up to 50 mm in length, comprise mainly grey and red-stained wacke, black and red chert, siltstone, small vein quartz and some igneous pebbles, including one of coarse-grained tonalitic granite with a well-developed micrographical intergrowth. Imbrication within the clast-supported conglomerate beds indicates that palaeocurrents flowed towards the north-west. The conglomerates contain lenses of medium-grained to pebbly, cross-bedded sandstone, similar to the matrix of the conglomerate. Some of the conglomerate beds coarsen upwards to contain cobbles up to 0.2 m in length. The underlying strata are mainly buff-brown to grey sandstones passing downwards into intercalated pebbly sandstones and pebble to cobble conglomerates, which include grey wacke sandstone and red and greenish grey chert at [NS 3562 0313], but here the base of the formation is cut out by the Straiton Fault.

Within the formation south of the Straiton Fault near Cairn Hill [NS 323 017], a persistent belt of conglomerate, about 60 m thick, is repeated about the syncline that extends west-south-west from Cawin Hill. The conglomerate, with a coarse-grained sandstone matrix and conglomeratic sandstone lenses, varies from grey-green to grey-pink in colour. The variegated red, grey, pink and white sandstones below the conglomerate belt are commonly coarse-grained to conglomeratic, whereas the sequence above includes thinly interbedded sandstones, siltstones and mudstones as well as conglomerate lenses.

Petrology

Volcanic rock fragments form the dominant lithic clast component within the Swanshaw Sandstone Formation, but metamorphic rock fragments are also a common component (Phillips and Smith, 2008). Sandstones from the formation show an apparent systematic variation in their polycrystalline quartz and volcanic lithic clast content with respect to monocrystalline quartz. This variation may reflect either a change in the maturity leading to an increase in the stable quartzose component of the sandstones, or, mixing of detritus from two separate source areas, one quartzose and the other volcanic in nature (Phillips and Smith, 2008). Varying degrees of compositional overlap between the Swanshaw Sandstone Formation sandstones suggest that they were derived from a similar 'well-mixed' source. In general, compositional data indicate that the sandstones from the formation in the Ayr (and New Cumnock) district have a transitional, recycled orogenic provenance (Phillips and Smith, 2008; Phillips et al., 1998).

Microconglomerates (varying from volcanic, lithic-rich to sedimentary clast rich types), quartzose, lithic-rich sandstones, and lithic-rich sandstones from the formation have been described in detail (Phillips and Smith, 2008). Lithic-rich sandstones are distinguished from quartzose litharenites by their lower detrital quartz content.

The microconglomerates and sandstones are mainly moderately to closely packed, clast- to matrix-supported, immature, heterolithic rocks. Detrital grains range in shape from angular, subangular to subrounded; occasional well-rounded grains are present in some rocks. Where the rocks are close packed, grain shapes are modified and individual grain boundaries are difficult to identify where alteration or pressure solution has occurred. Clay-grade matrix, chlorite cement, quartz overgrowths, hematitic rim cement and secondary/replacive carbonate cement are minor components of these clastic rocks.

Recognisable volcanic detrital clasts include tuffaceous clasts, quartz microporphyritic rhyolite, hematised plagioclase microporphyritic basaltic rock, very fine-grained basaltic to andesitic rock and feldspathic microcrystalline rock (felsite).

The metamorphic component includes foliated metasiltstone, fine-grained metasandstone, very fine-grained phyllitic to slaty rock, very fine-grained biotite-schist, white mica-rich phyllite or very fine-grained schistose rock, together with detrital garnet, biotite and possible staurolite. Deformed vein quartz or quartz mylonitic rock fragments and Mg-chloritic aggregates (possibly serpentinite) are rarer.

Carrick Volcanic Formation (CRKV)

The Carrick Volcanic Formation comprises lavas and associated sills, mainly of basaltic andesite with subordinate basalt and andesite. The formation name is derived from Brown Carrick Hill, south of Ayr (Browne et al., 2002) and is well exposed on the coast near Dunure, Culzean Castle and Maidens, and in the Carrick Hills. It crops out across south central Ayrshire and was formerly known as the volcanic group of the Lower Old Red Sandstone (Eyles et al., 1949). The Carrick Volcanic Formation is part of the larger calc-alkaline magmatic event, which occurred across the Midland Valley, and is penecontemporaneous (Early Devonian) with the Duneaton, Biggar and Pentland Hills volcanic formations (Browne et al., 2002). The petrogenesis of these igneous rocks has been attributed to north-west-dipping, active or recently terminated subduction (Thirlwall, 1982; 1986), or partial melting of a subduction-contaminated source (Smith, 1995: Philips et al., 1998). The formation is similar in character and composition to the Duneaton Volcanic Formation centred in Lanarkshire (Table 3).

The type section for the formation lies along the foreshore from west of Drumshang [NS 2449 1382] to Bracken Bay, south of Ayr [NS 273 182]. At the base of the formation, west of Drumshang, a basaltic andesite has a sharp, irregular junction above brick-red sandstone (Swanshaw Sandstone Formation) containing volcanic clasts and scattered pebbles of sandstone (Smith, 2000a). The top of the volcanic pile is not seen because it is unconformably overlain by rocks belonging to the Stratheden and Inverclyde groups.

The lavas in central Ayrshire are commonly subject to alteration, being partly hematised. The flows have slaggy tops that range from 3 to 15 m in thickness. Thin sandstone and conglomerate intercalations and sandstone-filled fissures in the lavas have been described several times since Geikie (1897). Smith (1910) considered the formation to be composed of terrestrial lava flows and claimed not to have seen a sill, although he admitted it would be difficult to distinguish flows from sills. In contrast, fluidisation of wet sediments during the emplacement of associated hypabyssal intrusions has been recognised more recently on the Ayrshire coast (Kokelaar, 1982), and explains some of the complex relationships between the igneous and sedimentary rocks. These are also described from the geological conservation sites, Port Schuchan to Dunure Castle, Culzean Harbour and Turnberry Lighthouse to Port Murray (Durant, 1999). Olivine-rich basalt is widespread at the base, but there is evidence of earlier volcanic eruptions from the lava pebbles present within the underlying Swanshaw Sandstone Formation. There are no clear eruptive centres or fissures feeding the formation but the Mochrum Vent, which cuts through the Swanshaw Sandstone, may be an early indication of the volcanic episode that culminated in the eruption of the Carrick Volcanic Formation.

Flow-banded andesites mapped south of Dalrymple around [NS 350 125] are considered to be extrusive, but more siliceous lavas are an known in this formation (cf. rhyolites in the Pentland Hills Volcanic Formation). It is uncertain whether there are any true ash fall tuffs in this area. A compact microporphyritic basaltic andesite flow on Highgate Hill at [NS 392 039] was distinguished from the surrounding macroporphyritic, more rubbly lava flows. In the Craig Hill area at [NS 398 021], the lavas commonly include autobrecciated or amygdaloidal zones, and between flows, conglomerates containing rounded pebbles of aphyric basaltic andesite are exposed.

The intercalated volcaniclastic conglomerates and sandstones are mainly derived locally from the volcanic rocks. In the Carrick Hills and Dunure areas, there are thin interbeds of fine-grained sandstone and sandy mudstone in which several trace fossils, including horizontal arthropod tracks, have been found (Eyles et al., 1949; Walker, 1985), and these have been used to interpret the prevailing environment as an ephemeral shallow lake.

Eyles et al. (1949) synthesised a generalised stratigraphy for the igneous succession in the Carrick Hills based mainly on specific quarry exposures which comprised: (a) olivine basalts; (b) hypersthene andesites and olivine basalts; (c) augite andesites, and (d) olivine basalts with porphyritic feldspars. During the resurvey, emphasis was placed on further recognition of an igneous succession and subdivision of the stratigraphy. However, stratigraphical subdivision did not prove possible solely from field observations because: (a) there exists considerable lateral and vertical variability within each unit, for example, in phenocryst content; (b) the composition of the majority of the units observable macroscopically is very similar, and (c) whilst some units are traceable for a few hundred metres (for example north of Brown Carrick Hill around [NS 2890 1640], as picked out by terraced (trap) features) they are not traceable for larger distances. Glacial erosion has picked out the general strike of the units but has not favoured erosion of the tops, bases or centres in particular. In addition, fault gullies in some outcrops (e.g. north of Shank Glen [NS 2990 1578]) suggest that the igneous sheets are dissected by faulting.

In fact, the above factors mean that the volcanic succession defined by Eyles et al. (1949) from specific exposures may not be representative. Aphyric units, for example, that Eyles et al. (1949) mapped at the top of Brown Carrick Hill, are the only clearly recognisable and mappable units.

Petrography

The igneous sheets consist of variably altered aphyric to macroporphyritic basalts and basaltic andesites. Most of the rocks are commonly plagioclase-phyric, and have phenocrysts of bronzite (magnesium-rich orthopyroxene) and/or augite and some olivine (Trewin and Thirlwall, 2002).

Using the BGS thin section collection, these rocks have been subdivided (Phillips, 1999), into: (a) plagioclase-olivine-clinopyroxene-phyric basalts, which are typically fine grained, inequigranular, micro-to macroporphyritic, sometimes with a pilotaxitic or hyalopilitic fabric and are variably amygdaloidal, and (b) plagioclase-clinopyroxene-orthopyroxene-phyric basaltic andesites, which are typically fine grained, inequigranular, micro- to macroporphyritic, massive to pilotaxitic and locally difficult to distinguish from the basalts (Plate 13), (Plate 14). Subsequently Phillips (2000) confirmed that the bulk of the lavas are basaltic andesites; basalts are less common. The basaltic andesites are very fine to medium grained, microporphyritic (less than or equal to 2.0 mm in size) to macroporphyritic (greater than 2.0 mm) rocks. They have a weakly to moderately developed pilotaxitic fabric, defined by either the alignment of phenocrysts or small plagioclase laths within the groundmass (Plate 13).

The phenocrysts are mainly plagioclase laths, up to 6 mm long, which in some rocks exhibit a seriate texture. The presence of cores and rims recording two main stages of plagioclase growth was noted by Phillips (2000). Small phenocrysts of clinopyroxene and in some cases orthopyroxene were also noted, and in altered rocks the pyroxene is pseudomorphed by chlorite or a mesh-textured mosaic of bowlingite and iddingsite as well as varying amounts of opaque oxide, quartz and carbonate. Orthopyroxene where present, is petrographically similar to accompanying clinopyroxene and is only distinguished by its straight extinction. The groundmass is typically a fine-grained mixture of plagioclase, pyroxene and minor opaque minerals. Small grains of biotite pseudomorphed by chlorite occur within the groundmass of some basaltic andesites, with traces of quartz and apatite. Amygdales, where present, are composed of carbonate or a mixture of carbonate and chalcedonic quartz.

In thin section, the basalts are fine-grained rocks which are commonly macroporphyritic and amydaloidal. Plagioclase is the dominant phenocryst phase forming anhedral to weakly euhedral laths up to 3 mm long. A pilotaxitic texture is crudely developed. Minor pale brown clinopyroxene phenocrysts up to 1 mm in size are also present. In the more altered basaltic rocks all the ferromagnesian minerals, which probably included olivine, are altered to an assemblage of chlorite, carbonate, quartz and opaque oxides.

The basalt groundmass is composed of fine- to very fine-grained plagioclase laths and subordinate, variably chloritised, granular pyroxene. Amygdales are composed of zoned cryptocrystalline zeolites, chlorite and quartz.

Geochemistry

Geochemically, the igneous rocks in the formation are confirmed as mostly basic andesites and andesites with a calc-alkaline character. No basalts (SiO₂ less than 52 per cent) were found in the formation and only one dacite was identified (Thirlwall, 1979; 1982; 1986). Many of the rocks contain orthopyroxene ± clinopyroxene, although the more siliceous rocks tend to lack orthopyroxene. However, there is no precise relationship between phenocryst mineralogy and host-rock silica content (Thirlwall, 1979). The single dacite (AY12 with over 65 per cent SiO₂ from near Fisherton [NS 258 170]; Thirlwall, 1986) also contains bronzite and clinopyroxene phenocrysts. Many samples are rich in Ni (about 100 ppm) and Cr (about 150 ppm) relative to normal calc-alkaline suites, and isotopic compositions are similar to lavas of a similar age in the north of the Midland Valley. Indeed, Thirlwall (1986) argued that the deep source (mantle) material for the basaltic andesites was common across the Midland Valley and the south-west Highlands.

Little or no systematic variation within the formation was found (Thirlwall, 1979), although unusually, in the Straiton area the orthopyroxene is hypersthene and the augites are more Fe-rich than usual, and contain pseudomorphs after exsolved pigeonite. The Straiton lavas tend to have lower La/Y and lower Sr than the majority of rocks in the Carrick Hills (Trewin and Thirlwall, 2002). Thirlwall (1979) concluded that the rocks are best described as calc-alkaline, although those from Straiton show some characteristics transitional into tholeiitic types.

A subsequent study of the formation (Phillips and Smith, 2008) showed that it had geochemical characteristics (Figure 7) similar to basaltic lavas erupted in either a continental volcanic-arc or within-plate setting. The lavas, mainly basaltic andesites (with SiO₂ contents between 52 and 57 weight per cent oxide) were divided into two suites: (1) a suite of high-Mg character, and, (2) a low-Mg suite more typical of calc-alkaline rocks (Thirlwall, 1981). The two analyses with SiO₂ contents typical of basalts (N2152) and (N2155) had high loss of volatiles on ignition and one (N2155) is associated with the Mochrum Hill Vent. Alteration of the volcanic rocks resulted in the variable remobilisation of the major elements Na₂O, K₂O and CaO, and trace elements Rb and Sr. The Nb/Y ratio (Figure 7) of the volcanic rocks (Nb/Y ± 0.6) clearly indicates that they are calc-alkaline in character. In order to overcome problems of remobilisation the data were plotted on classification diagrams of Winchester and Floyd (1977). These diagrams utilise trace elements (TiO₂, Zr, Nb, and Y) which are considered essentially immobile during low-grade metamorphism and alteration. On the plot of Zr/TiO₂ versus Nb/Y, the bulk of the data plot within the basalt field with a small number of analyses plotting just within the basaltic andesite field (Figure 7). A greater compositional variation is observed on the plot of Zr/TiO₂ versus SiO₂ ((Figure 7); top right) where the formation ranges from basalt to andesite in composition. It can be seen from (Figure 7) that the Carrick Volcanic Formation has the same compositional range as other Siluro-Devonian volcanic rocks. The major oxides TiO2, Fe₂O₃ total, MgO and CaO and trace elements Zn, Ni, V, Cr and Sc all show decreasing trends as SiO₂ increases, typical of calc-alkaline magmatic differentiation paths, consistent with the fractionation of olivine and clinopyroxene within the basaltic andesites, and hornblende in the more andesitic lithologies. The whole-rock data from this study indicated that there is a significant change in composition within the Carrick Volcanic Formation, marked by a break in the lithostratigraphical variations of the trace elements Nd, Y, Ce, Zr, Ni, Cr and V, and to a lesser extent, Rb and Hf. The formation can be divided into two groups in terms of their Ni and Cr contents; greater than 80 ppm high-Ni, less than 30 ppm low-Ni and greater than 100 ppm high-Cr and less than 40 ppm low-Cr (Phillips and Smith, 2008). Thirlwall (1981) concluded that the Ni- and Cr-rich lavas are primitive, and represent primary magmas which have undergone less than 25 per cent fractional crystallisation of the ferromagnesian minerals.

Dunure to Carrick Hills sections

The andesitic to basaltic igneous rocks with sedimentary intercalations are exceptionally exposed along the coast around Dunure, for example at the Port Schuhan to Dunure Site of Special Scientific Interest (SSSI) (Durant, 1999).

Earlier workers in this area, including Geikie (1897), Smith (1910) and Eyles et al. (1949), proposed that the igneous succession is composed of a series of lava flows with inter-eruption sedimentation. However, Kokelaar (1982) provided convincing evidence from Ayrshire coastal sections, including Dunure, to suggest that the succession formed as shallow level intrusions into wet sediment. Kokelaar's hypothesis was largely born out by field examination and the evidence is discussed further below.

At Dunure, over the Fisherton coastal zone [NS 256 164] to [NS 265 175] and in a few other localities, sedimentary interbeds crop out that are conformable within the volcanic succession. The Dunure exposures which prove an 80 m thickness comprise coarse volcaniclastic feldspathic arenites and volcaniclastic conglomerates. The northern end of the exposure is faulted [NS 2552 1627], but the conformable contact over the igneous succession is exposed in Dunure village [NS 2555 1594] and on the north side of the harbour wall [NS 2545 1606]. The sedimentary rocks are poorly sorted with rounded to subrounded clasts. The clasts are dominantly volcanic rocks derived from this formation together with rare baked sandstone. The clast- and matrix-supported conglomerates and coarsely laminated sandstones commonly have gradational boundaries and hints of trough- and cross-bedding. The sedimentary rocks are interpreted as locally-supplied subaqueous debris flow and fluvial deposits. The depositional basin may have formed in response to faulting similar to that observed at the northern edge of the exposure, or to local accommodation space created by a growing, but uneven, volcanic topography.

Other clearly conformable, but thinner sedimentary intercalations are observed within the formation. West of Fisherton cottage [NS 2650 1767], approximately 1.2 m of thinly-bedded, red sandstone passes up into 4–5 m of pebble–boulder, clast-supported conglomerate and thence into highly amygdaloidal igneous rock through an uneven intrusive contact. The conglomeratic unit shows a general fining upwards trend, with a poorly sorted base and moderately sorted top. The clasts are rounded to sub-rounded and are locally derived from porphyritic and aphyric igneous rocks and red and grey sandstones. The matrix at the basal part of the unit is a gravelly, coarse sand of locally derived clasts, but passes upwards into nodular and possibly reworked massive micrite with numerous sub-horizontal calcite veins. The conglomeratic facies seem indicative of a debris cone stabilised by a calcrete palaeosol. The debris cone may have formed in response to uplift/subsidence caused by local tectonism, or from local volcanic-generated topographical changes.

Red sandstones and siltstones are common conformable intercalations. Ripple traces, desiccation cracks, nodular micrite (calcrete, e.g. at [NS 2596 1715]) and red sandstones plus volcaniclastic conglomerate [NS 2665 1775] imply a shallow subaqueous to subaerial environment. Walker (1985) described trace fossils of aquatic or amphibious organisms within sedimentary rocks baked by the andesite sheets. These well-preserved arthropod trackways had been collected from three localities around Dunure (Pollard, 1995; Smith, 1909). Smith (1909) thought that the tracks had been preserved in 'quiet recesses' among the lava flows in an 'upland' (subaerial) environment, although he concluded that most of the tracks were made under water. He also noted that the tracks had been locally cut across, and were baked by what he considered to be lava flows passing over the fossiliferous strata. The subsequent observation that many of the volcanic sheets are intrusive (Kokelaar, 1982; see below) explains many of the relationships more adequately. From the more than 300 specimens Smith (1909) collected, he described 23 ichnogenera and 51 ichnospecies; however, these needed later revision in the light of modern knowledge of trace fossils. Pollard (1995) and Walker (1985) concluded that approximately ten of Smith's ichnogenera and twenty of his ichnospecies are valid. The modern ichnotaxa names which have been given (Pollard, 1995) to this collection of trails, trackways, swimming and resting traces include Siskemia, Stiaria, Danstairia, Kiercalia, ?Diplichnites, Stiallia, Isopodichnus (Rusophycus), ?Scolicia, and Mermia. It is considered that these trace fossils were made by crustaceans, eurypterids/scorpionids and possible 'myriapods' (Walker, 1985). There was no definite evidence of purely terrestrial animals, but several of the groups may have been amphibious at that time. The single body fossil in John Smith's collection from the shore close to Dunure Castle (Eyles et al., 1949), was earlier identified as the myriapod Kampecaris tuberculata Brade-Birks (Brade-Birks, 1923). A recent study of K. tuberculata (Wilson and Anderson, 2004) has shown that it is not a kampecarid myriapod but instead the archidesmid millipede Palaeodesmus tuberculata (Brade-Birks) and it is placed in the Order Incertae Sedis within Archipolypoda.

This fossil community comprised diverse arthropods, living in shallow water or marginal mud-flats of freshwater pools depositing soft muddy to silty sediments. These were later intruded by andesite sheets with consequent baking, fluidisation and production of gas flow and bubble structures. The depositional environment of the sediments, interpreted as a shallow ephemeral lake, is in agreement with the sedimentary structures observed during this resurvey.

Sedimentary-igneous relationships

As well as discrete sedimentary interbeds, the lower Devonian igneous succession contains laterally discontinuous sedimentary rocks, usually green-grey or reddened, fine sandstones, which have complex contact relationships. Kokelaar (1982) proposed that the complex sediment-igneous contact relationships (e.g. pillowed contacts, lobes and fingers of andesite in sediment, sediment-filled fractures), peperites and hyaloclastites, plus the replacement of sediment by igneous material, seen at Dunure and other Ayrshire coastal sections, are attributable to wet sediment fluidisation during the emplacement of shallow-level sills. Key lines of evidence which distinguish fluidised sediments from those incorporated at the base or top of lava flows include: (a) peperites are supported by sediments; (b) the sediments are commonly vesicular; (c) sediments do not contain andesite epiclasts; (d) planar lamination is preserved but sediment is reconstituted close to contacts; (e) the sediments are commonly enclosed, usually in pillowed andesites; voids are absent, and, (f) sediment structures are not consistent with those formed by being washed into voids (Kokelaar, 1982). Kokelaar's interpretation that the volcanic sheets are intrusive and post-date sediment deposition is confirmed by field examination. However, the maintenance of bedding and features such as the local folding of sediments near the top of irregular igneous contacts (e.g. west of Old Drumbane [NS 2596 1715]), imply that the sediments may have been semi-lithified when the intrusion took place. Fluidisation of wet sediment is a viable explanation for many of the features seen, but others could be explained by forceable intrusion into wet, but semilithified sediments, at shallow levels. In places, such as the Croy area (see below), amygdaloidal basaltic andesite occurs as loaded, pillowed structures, and may have been lava ploughing on to soft sediment.

Inland and down section of the coastal zone, irregular and baked sediment fragments are commonly found at the tops and bases of units. Where specific geometries can be observed (e.g. near Dunduff [NS 2657 1598]) the mode of genesis proven at the coastal exposures is appropriate. The individual igneous sheets are commonly a few to 10 m thick.

Culzean–Dipple area

The exposures of the Carrick Volcanic Formation on the coast at Maidens [NS 200 076] and Culzean [NS 230 102] are part of the Maidens to Doonfoot SSSI [NS 21 08]–[NS 31 19], as are the Culzean Harbour [NS 231 102] and Turnberry Lighthouse to Port Murray [NS 196 072]–[NS 207 081] Geological Conservation Review (GCR) sites (Durant, 1999). They have been described in detail by Kokelaar (1982) and Durant (1999), as well as in older publications by Tyrrell (1914), Eyles et al (1949) and references therein. These outliers of the formation show similar characteristics to the main Carrick Hills outcrop to the north-east (Sowerbutts, 1999).

The sequence is composed of amygdaloidal basalts and basaltic andesites, with many exposures of interbedded red sandstone and siltstone (Plate 15). The complex textures at the boundaries of the basalt and basaltic andesite sheets, such as peperites, hyaloclastites and pillows, and their interrelationship with the sedimentary rocks, led Kokelaar (1982) to propose that the igneous sheets were intruded at shallow levels into semi-lithified sediments. This view was discussed and supported by Durant (1999) and Sowerbutts (1999). The detail of the rocks and their mode of genesis are therefore not discussed again here. However, some additional features are described.

The base of the formation is patchily exposed on the coast west of Barwhin Hill [NS 2190 0945] and appears conformable over the Swanshaw Sandstone Formation. The top of the latter formation at this locality is marked by a grey mudstone. This is overlain by 5 m of lapilli-tuff and laminated tuff. Above these are a sharp-based basal breccia about 1 m thick, created from locally derived basaltic and sedimentary rock clasts, and an igneous sheet about 8 m thick, with a pillowed base and including sedimentary rafts and pieces. The lapilli-tuff contains numerous porphyritic vesicular and non-vesicular pieces of Carrick-type volcanic rock, up to 50 mm in diameter, and is interlaminated with better-sorted tuffs. The clasts of the basal breccia are considered to be more siliceous than any others seen in the coastal sequences of the formation, and may be representative of an earlier, siliceous and explosive volcanic phase (Durant, 1999). The lapilli-tuff, tuff and breccia represent a different phase, of either extrusive volcanism or epiclastic volcanic deposits, than the overlying intrusive sheets. This interpretation is in agreement with the suggestion by Kokelaar (1982) that some extrusive andesites existed close by, or that the shallow intrusive rocks were subject to penecontemporaneous erosion.

In the Culzean and Maidens sections, the igneous sheet boundaries are more easily mapped out than in the Dunure area to the north-east (Sowerbutts, 1999). The boundaries are commonly marked by layers of intact or disrupted sedimentary rocks, and can generally be traced for a few tens of metres. The mapped boundaries define igneous sheets, commonly 8–25 m thick, although Culzean Castle is built on a 35 m-thick andesite sheet (Durant, 1999).

The interbedded sedimentary rocks range from laminated siltstones and fine sandstones, to cross-bedded and rippled sandstones (e.g. north Turnberry Lighthouse [NS 199 075]) and a matrix-supported conglomerate (e.g. Culzean slipway [NS 2312 1027], interpreted as a debris-flow deposit in Durant, 1999). These intact sedimentary rock interbeds range from less than a metre in thickness, to about 6 m north of Turnberry Lighthouse, and probably a few tens of metres at Port Carrick [NS 221 906] and Port Murray [NS 206 080]. Desiccation cracks and raindrop prints were observed in the sedimentary rocks near Maidens [NS 2091 0815]. Durant (1999) recorded arthropod trackways in the sedimentary intercalation in Broad Sands Bay [NS 197 073].

Croy area

On the coast near Croy, the basal part of the Carrick Volcanic Formation is well exposed from 100 m south of Katie Gray's Rocks [NS 245 138] northwards. Much of the high ground to the east is free of superficial deposits, but good exposures are few because the formation weathers readily. Locally, to the west of Beoch [NS 286 145], 'trap topography' (prominent terracing) has developed due to differential weathering of the basaltic sheets, but this is not extensive probably due to faulting, jointing and the irregular but pervasive alteration within the formation. The formation here comprises a suite of aphyric to porphyritic basalts to basaltic andesites, commonly amydaloidal, in sheets or flows with sparse intercalations of volcaniclastic sandstones and breccias. The igneous rocks contain rafts, lenses and screens, usually of fine-grained and laminated siliciclastic sandstone. The relationship between the igneous and the sedimentary rocks is complex, and many features have been described by Kokelaar (1982).

The base of the formation at [NS 2449 1382] shows a sharp, irregular and complex junction of basaltic andesite above brick-red, fine-grained sandstone with clasts of volcanic rock and scattered pebbles of sedimentary rock. The basal zone of the basaltic andesite is very amydaloidal, and appears to be a flow which has ploughed and loaded into the underlying sediment. The amygdales are variably stretched into lobate, almost pillowed, forms. The amygdales are commonly infilled with calcite or green chloritic minerals; farther north, small agates occur. Lumps of basaltic andesite (up to 0.3 m in length) within the sandstone appear to be disconnected from the main igneous mass, and may have foundered in wet sediment. Some of the basaltic andesite is brecciated and former voids have been infilled with calcite. Hematitic alteration is variably developed in the basaltic andesite, which passes up into types in which plagioclase phenocrysts (up to 5 mm in length) are conspicuous. The more massive centres to the sheets or flows are characterised by blocky jointing and generally lack amygdales. The amygdaloidal zones, which are locally flow orientated, are usually associated with rafts or screens of fine sandstone, which locally include mudstone clasts and pebbly beds. The intercalated rafts or xenoliths of fine sandstone are commonly laminated, although where strongly baked the lamination is destroyed and locally vesicles develop in the sandstone. This is probably due to the original water content of the sandstone, which was coherent enough to retain its lamination. The incorporated fine sandstone fragments are locally wavy bedded and reorientated. At a small headland [NS 2456 1435], the basaltic andesite contains thermally metamorphosed, pale pink pedogenic carbonate (cornstone) nodules as well as mudstone fragments. Where the burn at Dunure Mill reaches the coast [NS 246 145], the purplish grey porphyritic andesites contain zones with large calcite-filled amygdales up to 80 mm long, and pinkish agates up to 100 mm long. Smaller amygdales contain chloritic minerals and quartz.

In the southern part of the outcrop around Howmoor Quarry [NS 278 119], a grey to purplish basaltic andesite, generally lacking in phenocrysts, is at least 5 m thick. This unit can be traced to the west [NS 266 122] but does not appear to be represented at the coast. A zone of finely amygdaloidal basaltic andesite, including reddish, fine-grained sandstone blocks, is exposed at the top of the quarry face. Large-scale spheroidal weathering joints are present within the quarry. A local 'zone' of olivine-rich basalt has been recognised above this level (Eyles et al., 1929, 1949).

There are few intercalated volcaniclastic beds, one of which is exposed in the old railway cutting to the north of the road to Dunure [NS 2504 1499]. Farther south [NS 2537 1425], flaggy, fine-grained sandstone, about 2 m thick and dipping 16° to the west, is interbedded within the basaltic andesites.

Dalrymple area

In the Dalrymple area, the Carrick Volcanic Formation comprises basalts and basaltic andesites with a high proportion of volcaniclastic sandstone near its base. The formation is exposed near Cassington [NS 355 116], and south of Montgomerieston [NS 357 121]. In the steep bank south of Montgomerieston, purple-grey and khaki weathered basaltic andesite is considered to lie above the Swanshaw Sandstone Formation, although the latter is unexposed. Some of the basaltic andesites are very vesicular; others are amgydaloidal with green chloritic infillings. The porphyritic basaltic andesite contains scattered green phenocrysts (possible pyroxene) 2–3 mm in length. There is a rough jointing in the volcanic rock, which is subparallel to the bedding observed in the intercalated volcaniclastic sandstones. An intercalation of thinly parallel bedded, greenish sandstones, containing reworked volcanic clasts and dipping between 3° and 8° to the south-south-east, is exposed in a disused quarry at [NS 3703 1218], south of Montgomerieston. The base of this intercalation has a complex relationship with the underlying igneous rocks, with fragments of slightly baked greenish sandstone and siltstone included within the basaltic andesite, and the latter may therefore be a sill. In the burn north-east of Cassington [NS 3564 1178], spheroidally weathered, more massive, grey basaltic andesites are intercalated with rubbly amygdaloidal zones. Some of the amygdales are infilled with calcite and/or chloritic minerals. Irregular, slightly baked, green and brown mottled sandstone inclusions may be xenoliths or rafts picked up by lava flows. Near the burn south of Cassington [NS 3536 1124], purplish grey, vesicular basaltic andesites have rotted to a khaki colour and lie above fine-grained sandstone belonging to the Swanshaw Sandstone Formation. Some of the volcanic rocks contain curved, aligned vesicular patterns, suggesting the igneous material balled up during flow. The rubbly, possibly autobrecciated volcanic rocks also contain irregular brown and pale green sandstone fragments/xenoliths. The lumps of vesicular basaltic andesite within intercalated sandstones could be interpreted either as volcanic bombs or epiclasts.

South-east of Fardenwilliam [NS 3693 1120], a thick unit of basaltic andesite is exposed below slightly baked, greenish grey, fine- to coarse-grained, volcaniclastic sandstone. The igneous unit is therefore possibly a high-level sill within the Carrick Volcanic Formation, although the base of the unit is not exposed. This basaltic andesite varies from a massive, purplish grey rock to one that contains small green phenocrysts and amygdales of agate and chlorite minerals.

Closely jointed, grey andesitic or rhyolitic rock with pale pink or yellow weathered feldspar phenocrysts, is exposed in a disused quarry [NS 3624 1158] south of Ballycoach. The rock has a set of low dipping joints dipping 5°–10° east-south-east, whereas in the pit at [NS 3612 1206] the joints dip 10° north-eastwards. North of Guiltree Hill [NS 359 113] this rock has a rubbly fracture. It is in part separated from the basaltic andesite at Cassington by a wedge of sandstone [NS 3591 1195], and the andesitic rock may have been extruded above the latter to form the upper part of the Carrick Volcanic Formation in this area. Previously, the andesitic rock was thought to be a 'felsite' sill (IGS, 1978a). The hill south of Burnbank Farm [NS 351 124] is largely formed by pinkish grey andesitic rock. Here the jointing has a spacing of 50–100 mm, which is observed to be parallel to a weak flow foliation within the rock dipping generally east to east-north-east. However, the flow foliation is locally irregular, and dips up to 70° towards the north-east at [NS 3536 1220]. This is interpreted as indicating that the body was either an andesitic flow or dome, rather than a felsitic laccolith (Eyles et al., 1949).

Straiton area

The Carrick Volcanic Formation, as exposed in the hills south of Straiton, mainly comprises porphyritic basalts and basaltic andesites (Phillips, 1999). They are typically macroporphyritic to microporphyritic and variably amygdaloidal. The rocks are commonly altered so that ferromagnesian minerals such as olivine and pyroxene are usually pseudomorphed. Pilotaxitic and hyalopilitic textures involving the alignment of plagioclase laths are common. Geochemically the lavas in the Straiton area are slightly different to those to the north (Thirlwall, 1979, 1981). Thirlwall's studies of the geochemistry of the Lower Devonian volcanic rocks in Ayrshire, indicated increases in the La/Y ratio and Sr content of the rocks from the Straiton area, through the Carrick Hills, to Distinkhorn in the Kilmarnock district to the north. This may be a systematic variation or suggest later strike-slip movement on the Straiton Fault.

The base of the formation south of Straiton [NS 372 034] is intercalated with conglomerates belonging to the Swanshaw Sandstone Formation. The top of the volcanic formation has been eroded away but it is estimated that the formation is about 300 m thick in this area.

The formation crops out on the hills either side of the Water of Girvan, particularly on Bennan Hill [NS 379 034] and Craigengower [NS 390 037]. On the crags north of Bennan Hill, towards the base of the formation, purplish grey basaltic andesites with dark green phenocrysts (possibly pyroxene) and quartz- and calcite-filled amygdales are exposed. Some basaltic andesites contain spherical vesicles up to 10 mm in diameter and also irregular greenish grey siltstone xenoliths or rafts, which are baked and intimately mixed with the igneous rock at [NS 3789 0359]. Just above this horizon, the basaltic andesite is grey and massive, with few amygdales or vesicles and this is probably the central portion of a flow or sheet. Locally the basaltic andesites are weathered to a pale purple or khaki colour.

Near Bennan Hill at [NS 3746 0327], the base of the formation lies directly above a pebble conglomerate at the top of the Swanshaw Sandstone Formation. The basaltic andesite at the junction is amydaloidal, with aligned vesicles in a plane striking 064° and dipping 25° to the south-east. There is no sign of hydrothermal interaction so the pebble bed is presumed to have been dry when it was overlain or intruded.

On Highgate Hill [NS 3915 0394], farther up the volcanic succession, the basalts and basaltic andesites are commonly macroporphyritic, with plagioclase laths 2–3 mm in length. Slightly baked, silty mudstone inclusions are grey-green to purple in colour and generally aligned, dipping about 45° to the south-east, although some have variable orientations. Adjacent igneous rock contains vesicles 10–50 mm in length aligned in a plane striking 042° and dipping up to 75° to the south-east. A distinctive flow or sheet of grey, compact, very fine-grained microporphyritic basaltic andesite (N2897) has been mapped south of Highgate Hill at [NS 3937 0396].

East-south-east of Craig Farm, at [NS 3898 0259], grey basaltic andesite is autobrecciated and cemented with hematitic volcanic fragments; baked, brown, fine-grained sandstone also occurs as inclusions. Around Craig Hill [NS 393 019] most of the basalts and basaltic andesites are macroporphyritic, but a 'trap topography' has developed on the hill slopes where intercalated, water-lain, red-brown siltstone and basaltic andesite pebble conglomerates are present, for example, on Craig Hill at [NS 3977 0210]. These sedimentary rocks indicate that the igneous rocks were being eroded for brief periods between eruptions.

Chapter 5 Upper Devonian

Stratheden Group

The Stratheden Group (Browne et al., 2002) comprises mainly red-brown sandstones, conglomerates and subordinate mudstones. Its main outcrop is in the northern and eastern Midland Valley. The group was previously described as part of the 'Upper Old Red Sandstone' (Eyles et al. 1949), which also included the overlying cornstone-bearing Kinnesswood Formation (Inverclyde Group). The Stratheden Group outcrop in this district exposes a relatively thin succession, which has not been subdivided but may be a condensed correlative of the Kelly Burn Sandstone Formation.

The group is faulted out to the north by an east–west fault south of Heads of Ayr, so that its crop does not extend offshore.

Bracken Bay exposures

The Stratheden Group exposed at Bracken Bay, south-west of the Heads of Ayr, comprises white, pink and green sandstones, pebbly sandstones and conglomerates that lie unconformably on an uneven palaeotopography of the Carrick Volcanic Formation. The group here reaches a maximum thickness of 110 m and is apparently conformably overlain by the Kinnesswood Formation 800 m inland of Bracken Bay (Sowerbutts, 1999). The well-exposed coastal section in Bracken Bay [NS 2763 1818] to [NS 2810 1823] commences with approximately 1 m of a basal angular conglomerate composed of micritic and calcareous, probably palaeosol-derived clasts. The conglomerate passes upwards into a 10 m wide zone of trough cross-bedded, buff sandstones, cherty sandstones and argillaceous green-grey sandstones. Above this, the general succession is one of coarse, moderately clean sandstones and conglomerates, that are commonly cross- and trough-bedded, with some local erosional scours. White, buff and pink mottled colours are common. Cross-bedding commonly exhibits a sigmoidal geometry and is locally oversteepened, overturned and slumped. The palaeocurrent direction derived from cross-bedding indicates that an overall north-easterly to easterly flow predominated. Conglomerate clasts are commonly poorly sorted and some are angular. The well-exposed sedimentary structures in this section are indicative of a braided fluvial environment dominated by superposed barforms (Sowerbutts, 1999).

Inland, such as at Carwinshoch Burn [NS 2903 1750] to [NS 2871 1803], exposures are dominated by pebbly sandstones and conglomerates, with rare red and green sandstones. During the 1920s survey, plant and fish remains were recorded on the field maps from a thin lens of red marl within this burn but were too poorly preserved for identification (Eyles et al., 1949).

Age and possible correlation

The succession as a whole is probably Late Devonian in age (Browne et al., 2002) on the evidence of fish faunas found at Bracken Bay [NS 274 181]. These include Bothriolepis leptocheira Traquair and Holoptychius sp., the former allowing tentative correlation with the upper Alves Beds of Morayshire and the suggestion of an early Fammenian age (Miles, 1968; Mykura in Craig, 1991). Bracken Bay is the type locality for B. leptocheira where specimens, originally described by Geikie (1869) as Pterichthys major Agassiz, included detached head-shield and trunk-armour plates and incomplete pectoral appendages. Traquair (1902) originally associated the bothriolepids at Bracken Bay with B. major, but later regarded them as B. leptocheira, characterised by the length and slenderness of the pectoral appendages (Eyles et al., 1949 and references therein). The exact bed from which the specimens came is not known, but during the 1920s revision, the geological survey collector W Manson found a fish horizon near the top of the sea cliff. Genus Bothriolepis, which is confined to the Upper Devonian in Scotland, is found in an equivalent stratigraphical position elsewhere in Scotland (Miles, 1968). B. leptocheira is closely similar and possibly related to B. jarviki Stensiö found within the lower part (which lacks Phyllolepis) of the Phyllolepis Series of East Greenland (Miles, 1968).

The field determination of Asterolepis sp. by Bassett (1958), in the roof of one of the sea caves at Bracken Bay, was regarded as suspect by Miles (1968) who cited no firm records of the genus Asterolepis, which is Middle Devonian (Givetian) in age in Scotland. The fossil fragments are set in variegated, fine-grained sandstone of fluvial origin. The fossils appear to have been transported and only the more resistant fragments are likely to be preserved in this fairly active environment (Dineley and Metcalf, 1999). Subsequent studies of the placoderm antiarch genus Bothriolepis (Dineley and Metcalf, 1999) indicated that they were freshwater forms, bottom dwellers and 'mudgrubbers'. The genus had paired sacs that may have functioned as 'lungs', which suggests that they inhabited streams or pools that occasionally dried up or became stagnant. The pectoral assemblages may have helped propel the fish along the bottom, or served as props or braces. The subspecies B. leptocheira curonica Gross is found in the north-western part of the East European Platform (Latvia, Lithuania and western Russia), where it belongs to the B. leptocheira Zone of the Eleja Baltic Regional Stage (lower Famennian; Lukševicˇs, 2001).

The fossil fish fauna and lithostratigraphy allow assignment of the beds at Bracken Bay to the Stratheden Group (Browne et al., 2002). The strata may equate to the Kelly Burn Sandstone Formation, which is exposed to the north on Sheet 22W (Irvine), and has similar palaeocurrent flows to the north-east (Bluck, 1978; 1980). However, this sequence is condensed compared to the Kelly Burn Sandstone near Hunterston, where the formation is considered to be over 400 m thick (Monro, 1999), and it thickens further across Great Cumbrae to around 1500 m in the Greenock district (Paterson et al., 1990). The strata could well extend offshore in the Firth of Clyde, but could not be distinguished from the Carboniferous rocks shown on the Ayr Bedrock map. It may be difficult to prove the connection below the Carboniferous rocks in the Ayrshire basin, but the fossil fish genera Bothriolepis and Holoptychius are widespread (Dineley and Metcalf, 1999) and probably did not live in an isolated sub-basin. There could therefore be an extension of the Stratheden Group basin to the south of that shown in (Figure 3) of the Irvine Memoir (Monro, 1999 p.18), along the present Firth of Clyde.

Petrology

The sedimentary rocks are generally poorly sorted, with rounded to angular clasts. Quartzite is the dominant conglomerate clast type, with varying amounts of chert, jasper, greywacke, phyllite, porphyry and reworked intraclasts. Volcanic rocks of the underlying Carrick Volcanic Formation are absent or very scarce, so it is likely that the clasts were remnants of those eroded from Lanark Group conglomerates (Browne et al 2002). In the type area of the Kelly Burn Sandstone Formation (Paterson et al., 1990), the clast content includes low grade metamorphic Dalradian rock types as well as reworked Siluro-Devonian clasts. The range of detrital clasts found in the Stratheden Group south of Ayr (see below) is compatible with that of the Kelly Burn Sandstone Formation.

In one thin section (N2057), (Plate 16) of cross-laminated, medium-grained, calcareous sandstone (Phillips, 2000), the rock is moderately to poorly sorted, clast to cement supported, with rounded to slightly irregular mudstone and calcareous siltstone rip-up clasts. The sand grains vary from subangular to rounded, with a moderate to low sphericity. The clast assemblage is dominated by monocrystalline quartz with subordinate to minor rock fragments. The minor to accessory detrital components include polycrystalline quartz, microcline, cherty rock, perthite, plagioclase, garnet, opaque minerals, altered biotite, apatite and very fine-grained sandstone.

Traces of a quartz rim cement and an earlier fine hematitic rim cement, were observed to precede a replacive fine, sparry carbonate cement, which is the main mode of cementation in the rock. Localised etching of clast grain boundaries during carbonate cementation has in places modified the detrital grain shape. The development of carbonate within the mudstone clasts before their incorporation into the sediment suggests a semi-arid climate, and the few rounded grains suggest an aeolian component.

Chapter 6 Carboniferous

Carboniferous strata form an important part of the outcrop in the Ayr district, and have been divided into four lithostratigraphical groups: Inverclyde, Strathclyde, Clackmannan and Scottish Coal Measures.

Offshore, in the Firth of Clyde, the Carboniferous groups are undifferentiated but are likely to be largely Coal Measures to the north of the subsurface limit of igneous intrusion indicated by magnetic anomaly, as shown on the Bedrock edition of the Ayr Sheet. This anomaly appears to be due to a sill or sills within the Coal Measures. However, in the offshore area west of Longhill Point up to the Carrick Fault, it is likely that older Carboniferous groups crop out on the sea floor. As the Carrick Fault downthrows to the west, Coal Measures, being younger, are likely to be offset southwards on the downthrow side, but the red and grey sandstone and marl encountered in a borehole (Lat. 55° 25.06' Long. 4° 49.50' W, about [NS 21 17]) has been interpreted as early Carboniferous (Chesher et al., 1972; British Geological Survey, 1985a). However, no spores were recovered from rocks in the borehole, so it is possible that they are from the Upper Coal Measures. West of the district, 9 km west of Maidenhead Bay around [NS 013 085], an offshore borehole (Borehole 72/1; McLean and Deegan, 1978) encountered Middle Coal Measures coal and disrupted dark grey mudstone containing carbonate nodules. The mudstone contained miospore assemblages indicative of Westphalian B age (Owens in McLean and Deegan, 1978). The offshore extent of the palaeohigh in the vicinity of Ayr during early Carboniferous times remains speculative.

Inverclyde Group

The Inverclyde Group includes most of the strata previously assigned to the 'Upper Old Red Sandstone' and the lower part of the 'Calciferous Sandstone Measures' (Eyles et al., 1949; compare Geikie et al., 1869). It comprises the Kinnesswood, Ballagan and Clyde Sandstone formations (Browne et al., 1999), which are Mississippian in age. In the Ayr district, the group crops out over a considerable area south-east of Ayr and between the Kerse Loch and Straiton faults. Except to the east of Dailly, relief over the crop of the group tends to be low due to the relative softness of the strata, particularly the Ballagan Formation. Good exposures are limited to watercourses or the coast south of Ayr.

Kinnesswood Formation

Within the Ayr district, the Kinnesswood Formation lies at the base of the Inverclyde Group conformably overlying the Stratheden Group south of Heads of Ayr, but stepping down on to the Lanark Group farther to the east, and down again farther south on to Silurian rocks in the Main Girvan Inlier. In several places in the district the formation is seen to be overlain conformably by the Ballagan Formation. The maximum thickness of the formation is estimated to be between 200 and 250 m in this district, but the sections are commonly faulted so its true thickness is difficult to ascertain.

The formation is characterised by a succession of mainly red-bed type fluviatile sandstones, some of which are calcareous and locally contain nodules of pedogenic carbonate (cornstone) as a precursor to calcrete (caliche) development in palaeosols. The sandstones are commonly mottled white, pale buff, yellow, purple-red and grey-purple. The subordinate siltstones and mudstones tend to be red-brown and locally, mottled grey. Beds adjacent to calcrete developments are usually mottled purple and yellow. Most of the calcrete is limestone but dolostone is reported from the formation in the north of the Midland Valley (Paterson et al., 1990).

The sandstones are mainly medium grained and medium to thick bedded. The units are commonly cross-bedded and locally have sharp, erosive bases. Some bases are lined with scattered white quartz and quartzite pebbles, and some beds include scattered pebbles together with reworked subangular clasts of mudstone and pedogenic carbonate from the floodplain deposits. The upward fining fluvial cycles usually have a lower, coarse-grained part followed by an upper, finer grained part laid down as the channels filled and were covered by floodplain sediment. Where the channels were stacked, multistorey sandstones are the result.

Although the formation is considered to have been deposited on a broad alluvial plain by an eastward-flowing river system (Hall et al., 1998), the current indicators from exposures in this district are too few and variable to confirm such a model. Deposition by small meandering streams with a reduced rate of deposition, indicated by the development of calcrete, is a likely scenario (Bluck, 1978).

The silty mudstone intervals range in thickness from about 0.02 to 1 m, and in places they contain interbedded siltstones and fine-grained sandstones. There are local sand-filled desiccation cracks, dewatering structures and disruption as the calcrete nodules developed. In immature calcretes, small carbonate concretions or nodules may have grown by evaporation of carbonate-bearing groundwaters in either the sandstones or the mudstones as the water table fluctuated in a semi-arid climate. The thicker carbonate lenses tend to be concentrated in the finer grained floodplain deposits, where they had a longer time to develop.

The mudstone intervals are generally barren of spores but occasionally sparse palynomorph assemblages have been identified (see below) as pre-CM Biozone, which is consistent with the evidence that the overlying Ballagan Formation contains CM Biozone (Tournaisian) assemblages in this district.

Between Ayr and Dalrymple, small exposures west of Pleasantfield [NS 3530 1636] and in the burn south of Abbothill [NS 3712 1869] include sandstones that are red to orange-brown, white and buff coloured. They are medium grained and parallel laminated, with nodules or lenses of pale grey pedogenic micrite. In the small burn south of Abbothill, red-brown and grey calcareous silty mudstones and siltstones including some micritic nodules, lie below a bed of coarse to granule grade sandstone with numerous micritic clasts. The greyish silty mudstones at this locality contain a very sparse organic residue with very rare, mainly fragmentary palynomorphs (Stephenson, 1999a). While no evidence of the age of the sample could be given, the palynomorph taxa and assemblage were similar to those found in the Kinnesswood Formation on the New Cumnock 1:50 000 Sheet (Smith, 1999a; Turner, 1994; McNestry, 1993), which was considered Tournaisian in age (LN-PC biozones). The succeeding red-brown silty mudstones and grey siltstones lie below purple-grey, coarse- and medium-grained sandstones near the top of the formation.

Around Dalrymple, the base of the formation is faulted against the Lanark Group, and the extent of its crop at rockhead is tentatively inferred due to lack of exposure. Poor exposures of pale brown and buff, medium-grained, fairly well-sorted sandstone occur to the north and south-east of Burnbank Farm [NS 353 129]. South-east of the farm [NS 3575 1259], there are also blocks of sandstone containing pebbles of vein quartz and calcrete typical of the Kinnesswood Formation. Larger exposures occur near Montgomerieston Quarry [NS 3640 1205], where over 2 m of yellow, brown to pale buff, medium-grained, quartzofeldspathic sandstone is exposed. The sandstone is well sorted in beds 0.4–0.8 m thick, some of which are massive although others are parallel and cross-bedded in sets up to 0.7 m thick. In general, these beds have low dips to the north-east. The few foresets measured indicate palaeocurrent flow was to the north-west. Some horizons contain scattered grey mudstone intraclasts, and sandstone blocks within the quarry contain scattered subangular quartz pebbles up to 10 mm in length. In the burn to the west, soft, weathered, grey and greenish grey, silty mudstone and siltstones are interbedded with the sandstones. These dip up to 20° to the north-east due to the fault lying to the west which downthrows these beds against Lower Devonian andesites. In a disused quarry farther south at [NS 3644 1176], fine- and medium-grained, quartzose, pale buff sandstones dip about 20° north-west. The sandstones occur in beds up to 0.8 m thick, interbedded with thin, red-brown and pale green siltstones and silty mudstones.

A faulted block containing Kinnesswood Formation strata is mapped on the north-eastern side of Barnshean Loch [NS 379 115], where well-sorted quartzitic, fine- to medium-grained sandstones vary in colour from buff to pink to brown where weathered. This sandstone sequence dips up to 15° to the south.

In the Kirkmichael area (Smith, 2003), the formation comprises mainly medium-grained quartzose sandstones with subordinate red and green argillaceous beds, and calcrete lenses and nodular horizons. The formation is conformably overlain by the argillaceous Ballagan Formation, which is known to be CM Biozone in age in the Straiton area to the east (Smith, 2001). While the Kinnesswood Formation is considered to be older, the boundary between the two formations could be slightly diachronous across the district. The junction between the two is not exposed in the Kirkmichael area, but it is inferred to cross the Water of Girvan at [NS 347 081], and is likely to be transitional. In the Straiton area to the east (Smith, 2001), the upper boundary of the Kinnesswood Formation is taken at the top of a predominantly sandstone succession, with subordinate reddish mudstones containing calcrete nodules. The formation crops out in the tributary south of Rig Burn [NS 3485 0505], where fine- to medium-grained, well-sorted, quartzose sandstones range in dip from 15° west up to 20° to the north-west. The fresh sandstones are generally pink-brown to purplish grey in colour and include horizons of calcrete concretions. Weathered sandstones are rusty brown and slightly carious. Cross-bedding is common in units 0.2–0.5 m thick and foresets dip north-west- and north-eastwards. Rarely exposed, purplish grey, soft, sandy siltstone and muddy siltstone interbeds occur up to 0.1 m thick, and at one locality lie above horizontal burrow marks on finely rippled sandstone.

An old limekiln at [NS 3472 0504] is situated just to the north of a quarry which exploited a thick, but impersistent, calcrete (cornstone) horizon within the formation.

The Kinnesswood Formation west of Straiton is unconformable on the Silurian inlier near Knockgardner [NS 351 035]. The low dip of the formation at Knockgardner creates a somewhat sinuous basal junction with the Silurian Knockgardner Formation, and the junction is interpreted as an unconformity (compare figure 1 of Clarkson et al., 1998). Here, pink-brown, cross-bedded sandstone beds are commonly coarse-grained and contain scattered small quartz pebbles. Elsewhere on the south side of the Midland Valley, for example north of New Cumnock, a well-developed calcrete bed occurs at the base of the formation (Smith, 1999a). This is not found above the unconformity at Knockgardner, perhaps due to a more active fluvial depositional environment, which would have prevented development of mature calcrete. However, a few tens of metres above the base of the formation, two calcrete beds, up to 1.5 m thick, crop out and were formerly quarried near Three Thorns [NS 3566 0435, 3586 0459] and east of Balgreggan [NS 3566 0492]. These beds dip 10–12° to the north-west and also south-west, below mainly fine- and medium-grained, pale buff and purple-brown sandstones containing calcareous calcrete nodules. The purple-brown sandstones tend to be softer as they lack calcite cement.

The best exposures in the Straiton area are found in the banks of the Water of Girvan from [NS 359 060] to [NS 358 072]. The section lies to the north of the Blairquhan Fault, which separates Kinnesswood successions dipping generally to the north-west (to the south) from that dipping to the south-east (north of the fault). Along the Water of Girvan, medium-grained, pink-brown mottled sandstones are interbedded with softer, silty sandstones and siltstones. Some of the sandstones are coarse grained and many contain red-brown to grey-green mudstone intraclasts. Some beds contain so many pink and pale green mottled mudstone intraclasts that they are conglomerates. Units, 0.4–0.5 m thick, of low angle cross-bedded sandstone, exhibit palaeocurrent indicators with a flow to the south-east (n=5).

At Tranew Linn [NS 3597 0708], medium-bedded, medium-grained, buff to brown, calcareous, quartzose sandstones contain micrite (calcrete) nodules up to 0.1 m in diameter. Interbedded, soft, brown to grey, silty mudstones with local calcareous concretions were sampled, but found to be barren of palynomorphs. The sandstones are commonly trough cross-bedded in units between 0.3–1 m thick, which are in places stacked together. Fine- to medium-grained quartzose sandstones tend to be better sorted, and occur interbedded with micaceous siltstones and silty mudstones in local coarsening upwards sequences, about 1 m thick.

Farther south, in the north-west bank of the Water of Girvan at [NS 3595 0606], buff to grey, medium- to coarse-grained sandstones contain small pebbles of micrite (calcrete), mudstone and vein quartz. Locally the coarse-grained beds coarsen upwards and are cross-bedded; planar foresets here dip to the north-east and north as well as south-east.

South of the Blairquhan Fault at [NS 3618 0538], harder, buff to pink quartzose sandstones are interbedded with soft, pink to red-brown sandstones and finely micaceous siltstones dipping up to 20° to the north-west. Locally the red-brown sandstones are green mottled adjacent to pale grey micrite nodules.

In Littleton Glen [NS 379 062], soft and hard, buff to red-brown quartzose sandstones are intercalated with red-brown and locally green mottled siltstones. The sandstone beds are generally between 0.1–0.3 m thick, locally cross-bedded with mudstone intraclasts. This succession dips gently to the north-east and north-north-east, below the predominantly grey siltstones and mudstones of the Ballagan Formation. A similar succession is exposed in March Burn [NS 3759 0699], where the top of the formation is marked by a 0.1 m thick calcrete bed lying above a red-brown to buff mottled, soft, fine-grained sandstone. The latter overlies a predominantly sandstone sequence, with local pale to mid grey mudstone interbeds which contained sparse palynomorphs of indeterminate age (Stephenson, 2001a). In contrast, the grey mudstones within the overlying Ballagan Formation contain palynomorph assemblages characteristic of the 'lower' Ballagan Formation (Stephenson, 2001a).

The succession of medium-grained, pink, quartzose sandstones to the east of the north-north-westerly trending fault south of Loch Spallander is thought to be Kinnesswood Formation. It contains sparse quartz pebbles and occasional thin, pale grey mudstone interbeds, one of which (MPA 49689) contained a sparse palynomorph assemblage (with Auroraspora macra Sullivan but lacking Schopfites claviger Sullivan) which may be pre-CM Biozone (Stephenson, 2001a).

A small half-graben of Kinnesswood Formation has been mapped around Pinmerry Farm [NS 347 112], where pinkish mottled sandstones with mud clasts are exposed in the burn south of the farm. The mud clasts contained a sparse palynomorph assemblage (MPA 49990) including A. macra and other taxa common in the CM Biozone, however, S. claviger is absent, so it may be a pre-CM Biozone assemblage (Stephenson, 2001b). Stephenson concluded that the composition of the assemblage and the sparse nature of the residue suggested similarities to the samples north of Straiton.

South of Heads of Ayr (Sowerbutts, 1999), the Kinnesswood Formation is considered to lie conformably between the Stratheden Group (Kelly Burn Sandstone Formation) and Ballagan Formation, though the contacts are nowhere exposed. The Kinnesswood Formation is distinguished from the underlying Stratheden Group by the lack of pebble conglomerates and the presence of nodular calcrete horizons (cornstone).

The exposures downstream of Low Glenayes, in a small gorge [NS 2893 1824] to [NS 2876 1808], comprise 18 m of coarse to medium-grained, well-sorted and rounded, micaceous white-buff sandstones. Some clean, trough cross-bedded sandstones contain well-rounded quartz grains that are probably reworked aeolian clasts. Minor beds of matrix-supported chert-quartz-intraclast conglomerates contain clasts less than 10 mm in size and a nodular micritic palaeosol horizon (calcrete). In the previous survey, a lenticular interbed of red marl yielded plant and fish remains from this section (Sowerbutts, 1999).

Petrology

The majority of the formation is made up of slightly feldspathic quartz-rich sandstones, ranging from fine to coarse grained. Sample (N2589) from near Doonfoot is coarse grained and poorly sorted (see Phillips and Smith, 2008 for more detailed descriptions). The detrital grains are angular to subrounded in shape and supported in medium- to coarse-grained, sparry carbonate cement; however, the grain shapes have been modified by etching during the development of the cement. The detrital assemblage is mainly composed of monocrystalline quartz with minor K-feldspar (including microcline). Other minor to accessory clasts include polycrystalline quartz, plagioclase, opaque minerals, garnet, biotite, felsite, tuffaceous rock, mesoperthite and muscovite.

Nearby, calcareous sandstone (N2590) is clast to cement supported with well developed carbonate cement. The cement apparently replaced the original matrix and unstable detrital components within this sandstone because outlines of highly altered/replaced rock fragments can still be recognised. The detrital assemblage was originally composed of monocrystalline quartz, subordinate rock fragments and minor feldspar clasts. The rock fragments that can be recognised include a very fine-grained tuffaceous rock. Other minor detrital clasts include plagioclase, muscovite, K-feldspar, microcline, chloritised biotite, opaque minerals, polycrystalline quartz and chlorite pseudomorphs.

Examination of a thin section (N4290) of pinkish sandstone from near Blairquhan showed a well-sorted, fine- to medium-grained arenite, mainly composed of quartz, plagioclase and K-feldspar. The compact texture contains mainly subangular to subrounded grains with little or no interstitial clay or carbonate. Minor opaques or iron ores are present together with scattered white mica, chlorite and possibly zircon grains. There are also a few lithic clasts such as chert and quartzite.

A thin section of a calcareous microconglomerate (N2586) shows a very poorly sorted, open packed rock with granule- to pebble-sized clasts of micrite, the shapes of which have been modified due to pressure solution during compaction. Other similar sized clasts are composed of very fine-grained, possibly tuffaceous rock, strained polycrystalline quartz and hematised cherty rock. The sandstone matrix consists of etched clasts of monocrystalline quartz set in well-developed carbonate cement which has replaced the original matrix and/or unstable detrital grains. The minor detrital grains recognised include polycrystalline quartz, plagioclase, sheared vein quartz or quartz mylonite, K-feldspar, muscovite and tourmaline.

Two thin sections (N2585), (N2591) of pedogenic limestone from the Kinnesswood Formation have been examined. These brecciated, very fine-grained calcareous rocks comprise irregular, angular to subangular micritic carbonate fragments, together with a slightly coarser-grained sparry carbonate. In places, these rocks contain angular to subrounded monocrystalline quartz grains which may represent relict detrital clasts and/or diagenetic quartz crystals. Rounded to subrounded micritic carbonate clasts were also recognised within these calcareous rocks. In thin section (N2585), the sparry carbonate cement is texturally zoned and forms fine-grained, granular-looking rims upon the rock fragments, with the remainder of the vugs/fractures being composed of coarser grained, anhedral granular carbonate. By way of contrast, the fractures in thin section N2591 are filled by texturally zoned, cryptocrystalline to very fine-grained quartz.

Ballagan Formation (BGN)

Previously termed the Cementstone Group of the Calciferous Sandstone Measures, the Ballagan Formation conformably overlies the Kinnesswood Formation. Within the Ayr district, the characteristic finely laminated mudstone, siltstone and dolomitic limestone sedimentary rocks forming the Ballagan Formation (Browne et al., 1999) reach up to 500 m in thickness. Generally the carbonate beds (cementstones) within the formation are argillaceous, ferroan dolostones (Read et al., 2002), but no analyses of the dolomitic limestones (which mostly react with dilute HCl) in the Ayr district are available to confirm their composition. There are also considerable facies variations within the district, indicated by sandstone-dominated strata prevalent in the south, and grey mudstone facies which are predominant farther north. In the Dailly area, the three lithological subdivisions of the Ballagan Formation recognised by Eyles et al. (1949) have been further distinguished into the basal Drumwhirn Member, the Lindsayston Burn Member and an upper unit with typical Ballagan Formation facies (Monaghan, 2004).

As a result of palynological studies during the latest resurvey, it appears that the presence of Auroraspora macra and Schopfites claviger throughout the Ballagan Formation, together with the absence of Lycospora pusilla (Ibrahim) Somers (index taxon for the Visean Pu Bizone), indicate a CM Biozone age for the entire formation in Ayrshire. The presence of ostracods (e.g. Sucella cf. affiliata (Jones & Kirkby)) that elsewhere are considered typically Visean, in the upper part of the formation at the Heads of Ayr, may suggest that this section extends into the Visean. Since L. pusilla is rare in the lower part of its range, it is possible that it remains unrecorded in the section at the coast. However, this seems highly unlikely considering the number of samples examined and the presence of other lycopsid spores (Stephenson et al., 2002). Another possibility is that the ostracod ranges require recalibrating (Stephenson et al., 2002).

A comparative study of algal palynomorphs and ostracods within the Ballagan Formation at the Heads of Ayr showed that the algal palynomorphs Botryococcus spp. Chromotriletes spp., Leiosphaeridia spp. and Leiosphaeridia sp. A. were associated with the low salinity to freshwater ostracod Sulcella cf. affiliata. Modern Botryococcus are confined to freshwater or brackish bodies of water, which supports the interpretation that there was a variation in facies from freshwater (containing Sulcella sp and Botryococcus) to tidal (containing ostracods Cavellina coela (Rome) and C. incurvescens (Jones & Kirkby)) within the formation in this section (Stephenson, 2002).

The restriction of the formation to the CM Biozone is consistent with the palynological record across in the Cultra area of Northern Ireland (Ballycultra Formation of similar facies and age; Mitchell, 2004). The formation in the Ayr district may be partly diachronously younger than the formation in the New Cumnock area (Smith, 1999a) where PC Biozone spores were identified (Turner, 1995). It may also suggest younger diachronous deposition of the formation in the eastern Midland Valley; as at Cockburnspath, the formation appears to span the CM–Pu Biozone boundary (Clayton, 1971; Neves et al., 1973). This was also found to be the case near the Firth of Tay, where most of the formation was assigned to the CM Biozone but the uppermost samples ranged into the Pu Biozone (Browne, 1980). This is also the case in Berwickshire (Scott et al., 1984). Other evidence from Oxroad Bay, near Dunbar, also indicates volcanic activity during CM Biozone times (Scott et al., 1984).

Heads of Ayr section

The best exposures of the Ballagan Formation are in the coastal sections between Bracken Bay [NS 279 183] and Greenan Castle [NS 310 192] ((Figure 8); the Heads of Ayr section). The base of this section in Bracken Bay is faulted against the Stratheden Group. Regional field relations and palynomorph biostratigraphy indicate that a few tens of metres of the lower part of the formation is missing from this section, particularly the pedogenic limestone–dolostone (calcrete) horizons known from the lowest Ballagan Formation inland (Stephenson et al., 2002). In the Heads of Ayr section, the formation is overlain by the Greenan Castle Pyroclastic Member, which is related to the Heads of Ayr Vent (Figure 8), and the evidence that the Ballagan Formation was lithified before its eruption indicates that there is probably a disconformity between the two (Stephenson et al., 2002). Within the section, the Ballagan Formation comprises interlaminated or interbedded green-grey-brown silty mudstone, thin, nonmarine, dolomitic limestone ('cementstone', in places nodular) and yellow, brown and grey sandstone (Plate 17). Sandstones are of variable composition, ranging from clean quartz arenites to dirty, green-grey, possibly volcaniclastic, micaceous arenites. Sedimentary structures observed include desiccation cracks, halite pseudomorphs, parallel lamination in siltstones and carbonate-bearing mudstones, cross, trough and flaser type bedding in sandstones and interbedded units, and straight-crested wave ripples.

Fossils such as carbonaceous petrified wood fragments, Spirorbis sp., Lingula mytilloides J Sowerby, Modiolus cf. latus (Portlock), ostracods, estheriids and fish fragments (including rhizodont scales) have been found previously (Eyles et al., 1949). Planolites-type trace fossils were observed at several localities (Sowerbutts, 1999) and L. mytilloides and fish scales were identified from a locality [NS 2964 1875] near Long Rue in the lower part of the formation (British Geological Survey, 2003). The fossil assemblage that typically includes worm tubes, M. latus and ostracods is termed the 'Modiola-phase', and is best interpreted as quasi-marine (see Dean, 2002 and references therein). Palynological studies on the formation in the Heads of Ayr area (Stephenson, 1999b; 2000a; 2001b) are summarised in Stephenson et al. (2002). Ostracods collected from the middle of the formation, just to the east of the Heads of Ayr vent, yield abundant cavellinids (Cavellina coela and C. incurvescens; which are definitely Tournaisian in age), Paraparchites and rare Bairdia (Williams and Taylor, 2001). These assemblages suggest marginal marine facies, perhaps supratidal — transitional to marine at some horizons, and more marine at others (Williams and Taylor, 2001). Higher in the middle part of the formation, possible brackish water facies are suggested by Acutiangulata-rich faunas. A stressed (marginal marine or brackish?) environment is suggested by the low-diversity, Sulcella-dominated faunas of the upper part of the formation (Williams and Taylor, 2001).

The overall depositional environment is interpreted as a semi-arid to arid coastal floodplain with periodic emergence. This is supported by ostracod assemblages and stable isotope studies on the ostracods, that are indicative of marginal marine, brackish or hypersaline environments rather than marine conditions (Stephenson et al 2002; Williams et al., 2005; Williams et al., 2006). The rocks were deposited in closer proximity to a clastic and volcaniclastic source area than is typical of the Ballagan Formation. Some sediment may have been supplied from the erosion of Lower Devonian volcanic rocks or from contemporaneous volcanism, possibly associated with early activity on a proto-Heads of Ayr vent. An upwards reduction in fluvial activity and increase in aridity is implied, with fewer and thinner sandstones, in contrast to palynomorph evidence which suggests fresh or brackish lagoons and high diversity assemblages, with lycopsids suggesting wetter conditions (Stephenson et al., 2002).

Detailed palaeontological sampling from mudstones and silty mudstones from the Heads of Ayr section, as well as inland sections, gave a CM palynomorph biozonation and a Beyrichiopsis glyptopleuroides–Eriella ostracod biozonation, both indicative of Tournaisian age (Stephenson et al., 2002).

The study also established a local, lower palynostratigraphical subdivision which lies to the west of the Deil's Dyke, characterised by a low-diversity, terrestrially derived Colatisporites-dominated flora, including Auroraspora macra, Retrusotrilites incohatus Sullivan and Schopfites claviger. An upper part is characterised by east of the Deil's Dyke in the Heads of Ayr section (Stephenson et al., 2002; Williams et al., 2005) high-diversity assemblages alternating with Leiosphaeridia-dominated assemblages, aquatic algal palynomorphs and the ostracod Sulcella cf. affiliate. It is notable that the change in flora in the upper part of the Heads of Ayr section appears to be independent of any major sedimentary facies change within the formation (Stephenson et al., 2002).

Dailly sections

Well-exposed sections in the Ballagan Formation around Dailly, where there is a strong fluviatile influence in its deposition, have allowed the subdivision of the formation outlined below.

Drumwhirn Member (DRWN)

The Drumwhirn Member is the local, basal member (Monaghan, 2004) of the Ballagan Formation. It is characterised by white, buff or yellow, thinly planar-bedded sandstone intercalated with finely laminated, grey calcareous siltstone and mudstone, and rare thin, nonmarine, dolomitic limestone ('cementstone'). It is poorly exposed east of Whitehill Farm (e.g. from [NS 275 005] to [NS 300 016]). Palynological samples (MPA51242) and (MPA51243) gave a CM miospore zonation, indicative of a Tournaisian age (M Stephenson, unpublished report). The unit appears to be at least 275 m thick and is overlain by the Lindsayston Burn Member.

Lindsayston Burn Member (LSBU)

The Lindsayston Burn Member (Monaghan, 2004) is characterised by red, green, grey and purple sandstone, conglomerate, pedogenic carbonate (cornstone) beds or nodules, and red-green siltstone and mudstone (Figure 9). It is almost continuously exposed in the Lindsayston Burn between [NS 2821 0082] and [NS 2755 0093] reaching a thickness of 134 m, though in other areas it may reach up to 285 m thick.

In the lower, south-easterly part of the section between [NS 2821 0082] and [NS 2272 0089], the strata are characterised by buff-white-grey, hard, flaggy sandstone, interbedded with soft, red-green sandstone containing pedogenic carbonate nodules and pipes, and rarer beds of red-green siltstone and mudstone with pedogenic carbonate nodules. Thick channelised sandstone and conglomerate, a few metres thick and tens of metres wide, are observed at several localities. These units are typically erosively based with local scoured surfaces and fine upwards, for example from coarse- to medium-grained sandstone to more laterally extensive fine-grained sandstone beds. Conglomerates are typically clast supported and poorly sorted, with subangular clasts of chert, quartz, greywacke, altered igneous rocks and some sandstone intraclasts, up to 0.2 m in diameter; rarely, they are cross-bedded. Lenses and lags of conglomerate and rip-up clasts are also found within the massive, cross- and trough-bedded flaggy sandstone that forms a major component of this lithofacies. In the most extensive exposures, lateral facies variability was observed, for example in the thickness of channelised sandstone beds.

The sedimentary facies observed are interpreted as deposition in a fluvial environment. Thick-bedded, erosively-based sandstone and conglomerate represent channel deposits; laterally extensive sandstone beds are interpreted as sheet-flood and crevasse splay deposits, and siltstone and mudstone with pedogenic carbonate (calcrete) represent overbank deposits. The observed facies variation illustrates how major channel facies pass laterally to channel margin, crevasse splay and overbank deposits. The environment must have been semi-arid, with relatively well-drained soils for the pedogenic carbonate nodules and pipes to develop, but with a sizeable amount of water flow from time-to-time reflected in the width of channel sandstones and cobble-size of conglomerate clasts. Six palaeocurrent indicators show a dominant north-westwards flow direction for the fluvial system. Three palynology samples (MPA 51238, 51239 and 51240) from Lindsayston Burn Member gave a CM miospore zonation, indicative of a Tournaisian age (M Stephenson, unpublished report). The member is overlain by the upper part of the Ballagan Formation (see below).

In the north-western part of the Lindsayston Burn section from [NS 2272 0089] to [NS 2749 0097], the upper part of the Lindsayston Burn Member records a gradational change to the upper part of the Ballagan Formation. Medium- to coarse-grained sandstone interbedded with red-grey siltstone with pedogenic carbonate are still common, but intercalations of grey, calcareous siltstone and mudstone, and thinly-bedded, fine-grained sandstone, in packages up to about 0.7 m thick, are also observed (e.g. at [NS 2772 0089]). Sedimentary structures include mudcracks, raindrop prints and tool-marked bases of thicker sandstone beds. Indeterminate bioturbation was also observed in grey siltstone. One sample (MPA 51241) contained the palynomorphs, Leiosphaeridia spp., including sp. A and B.

A similar fluvially dominated depositional environment to that described for the lower parts of the Lindsayston Burn Member is envisaged for these strata. The intercalation of grey, calcareous siltstone and mudstone with mudcracks and bioturbation, is interpreted as a gradual transition and change in environment to the lacustrine — marginal marine lagoonal — coastal plain environment envisaged for the upper part of the Ballagan Formation.

Inland exposures south-east of Ayr

South-east of Ayr, the Ballagan Formation underlies a broad belt from Friarland to Knockjarder (Smith, 1999b), but is poorly exposed. In the small burn south-west of Friarland [NS 3675 1916], near horizontal, grey, carbonate-bearing, silty mudstones are interbedded with earthy, impure limestones and fine-grained, pale buff sandstones up to 70 mm thick. In an old pit west-south-west of Friarland [NS 3673 1942], impure limestones and laminated dolomitic micrites are thinly interbedded with grey siltstones and mudstones; the latter contain a rich organic residue including well preserved palynomorphs. The palynomorph assemblage, including Colatisporites sp., Schopfites claviger and Auroraspora macra, is typical of the CM Biozone (Stephenson, 1999a) from the lower part of the formation (Stephenson et al., 2002). The Springwater Fishery development, near Knockjarder [NS 364 154], created several exposures within the formation, including some mudstones and siltstones which besides being mid-grey and greenish grey, were locally red-brown mottled. Associated with these beds are thin, laminated micrites and pinkish gypsum/anhydrite veinlets parallel to bedding. Subordinate buff-grey, fine- to medium-grained sandstones up to 1 m thick are micaceous. A grey mudstone sample (MPA 48206 from [NS 3638 1538]) was again found to contain palynomorphs, typical of the CM Biozone (Stephenson, 1999a) and the lower part of the Ballagan Formation.

Inland exposures on the River Doon near Dalrymple

North of the River Doon [NS 3745 1367], exposures of pale to mid-grey mudstone and silty mudstone pass up into a fairly massive bed, 0.27 m thick, of calcareous, silty sandstone with sparry calcareous nodules towards the top (Smith, 2000b). Above the silty sandstone is 0.3 m of mid grey mudstone, with scattered, fine-grained mica and carbonaceous flakes. This bed was sampled for palynomorphs and yielded an assemblage (including A. macra and S. claviger) assigned to the CM Biozone (Stephenson, 2000c). The assemblage was similar to the low diversity 'lower' Ballagan assemblages described from the Heads of Ayr (Stephenson, 2000a, b). The succeeding mid grey sandy siltstone, 0.32 m thick, is less fissile and overlain by 0.45 m of darker grey, faintly laminated mudstone and silty mudstone. South of the River Doon, in a section in Burnton Burn [NS 3789 1321] (MPA 49028), fine-grained, buff to grey sandstones and thin-bedded to laminated, dark to mid grey mudstones also yielded a 'lower' Ballagan assemblage (Stephenson, 2000c).

On the south-west bank of the River Doon [NS 3759 1339] a sequence of pale grey, medium-grained sandstones, over 2 m thick, includes coarse, white micas and plant remains (including stem imprints) on bedding surfaces. The beds are up to 0.66 m thick, with softer micaceous laminated tops. Small scale ripple cross-bedding occurs in sets up to 0.15 m thick. Also on the south-west bank of the River Doon [NS 3727 1376], outcrops of pale buff, medium-grained quartzose sandstone occur in beds 0.3–0.4 m thick.

A fault trending north-westerly is inferred within the formation, crossing the Brunton Burn south of the Doon. Grey mudstones farther south in the burn at [NS 3794 1295] contain a diverse 'upper' Ballagan assemblage (MPA 49030), including Radiizonates mirabilis Phillips & Clayton as well as A. macra and S. claviger (Stephenson, 2000c; Stephenson et al., 2002). In the absence of this fault, north-east dipping strata would otherwise apparently lie below 'lower' Ballagan assemblages (MPA 49027 and MPA 49028).

Mid grey mudstones and dolomitic micrites near Netherton [NS 3714 1360] contain an 'upper' Ballagan assemblage (MPA 49029), supporting the presence of the fault. The dolomitic micrites near Netherton are 0.1 m thick and weathered to a pale cream colour with a pelleted to laminated texture. Interbedded mudstone, siltstone and fine-grained, ripple-marked sandstone include some desiccation cracks.

Doonfoot section

In the valley of the River Doon, near Doonfoot [NS 324 187], exposures of Ballagan Formation comprise fine-grained sandstones and grey mudstones with thin, dolomitic micrites. Medium- and coarse-grained sandstones lie towards the base of the formation. Three palynomorph bearing samples were collected and examined from the Doonfoot area. Samples MPA 49703 [NS 3236 1857] and MPA 49705 [NS 3238 1870] are from the mid to upper parts of the formation and yield sparse, poorly preserved assemblages including Leiosphaeridia spp. These may correlate with the Leiosphaeridia-type assemblage of the upper Ballagan Formation in the Heads of Ayr section, although the poor preservation and sparseness of the assemblages precludes precise correlation (Stephenson et al., 2002). Sample MPA 49704 [NS 3231 1847] yields an assemblage rich in A. macra, Colatisporites decorus (Bharadwaj and Venkatachala) Williams in Neves, et al. (1973), and Colatisporites denticulatus Neville. Also present are S. claviger, Apiculiretrusispora spp., Discernisporites crenulatus (Playford) Clayton, Knoxisporites spp., Auroraspora cf. macra, Cyclogranisporites spp., Convolutispora spp., Retrusotrilites spp., Perotrilites cf. perinatus Hughes and Playford and Verrucosisporites nitidus (Naumova) Playford (Stephenson, 2001c). The absence of taxa indicative of palynomorph assemblages of the Upper Ballagan Formation suggests that this sample is most likely equivalent to the lower Ballagan Formation as exposed on the coast, although the proximity to the samples MPA 49703 and MPA 49705 suggests a stratigraphical position high in that interval.

March Burn section

At March Burn [NS 376 070], the lowest part of the formation is conformable with the Kinnesswood Formation. The base of the Ballagan Formation (Figure 10a) is established in the March Burn section at [NS 3759 0699] where, above a calcretised red and buff mottled sandstone belonging to the Kinnesswood formation, a 0.1 m thick dark grey, earthy, calcareous siltstone is overlain by 0.15 m of mid grey mudstone. The mudstone contains a high abundance but low diversity palynomorph assemblage (MPA 49697) from the CM Biozone, typical of the lower part of the Ballagan Formation, but also contains Grandispora cornuta Higgs (previously recorded from the LL to PC biozones; Stephenson et al., 2002). This species is significant, as it has not been recorded from the lower Ballagan Formation at Bracken Bay, where its absence supports the field evidence that a few tens of metres of the Lower Ballagan Formation are faulted out at the coast.

In March Burn, above the basal argillaceous beds, medium-bedded, fine- to medium-grained sandstones are greyish buff in colour and flaser rippled. Interbedded sandstones are softer, fine grained and thinly bedded. Some sandstone beds fine up into rippled and bioturbated tops, overlain by pale grey siltstones and mid grey silty mudstones. Thick interbeds of blocky siltstone and silty mudstone, up to 2.2 m thick, are purple brown towards their tops, with scattered small calcareous concretions. Some siltstones and mudstones are red-brown with a local, pale green mottling. They include lenses of fine-grained sandstone up to 0.3 m thick. Within the lower part of the formation, a prominent bed of granular to pebbly 'calcrete' conglomerate is exposed [NS 3773 0712] which has a sharp, erosive base and varies in thickness up to 0.5 m. Similar conglomeratic beds up to 0.14 m thick are interbedded with the overlying grey siltstones, sandstones and mudstones. Some of these siltstones are bioturbated and contain small mica and cabonaceous flakes. Within the overlying sandstones low angle cross-beds have foreset bedding dipping northwards. These sandstones vary from pinkish brown to grey-green, and locally contain mudstone rip-up clasts. Interbedded siltstones are locally red-brown with calcareous concretions.

Nearby in Littleton Glen [NS 3823 0654], grey mudstones contain a CM Biozone palynomorph assemblage which is probably 'lower Ballagan', as the sample contains no taxa indicative of the 'upper Ballagan' and has a lower diversity than is typical of the 'upper Ballagan'. In Littleton Glen, the succession is predominantly one of fine- to medium-grained sandstones with interbedded mid grey, grey-green and pink mudstones, with sparse dolomitic micrite interbeds. The top of the Ballagan Formation is cut out to the east of Littleton Glen by the Lambdoughty Fault running north-north-west through Loch Spallander.

Troquhain sections

Sections in the upper part of the formation are exposed in tributaries east of Troquhain [NS 378 093], and the composite graphical log of the exposures (Figure 10b) also indicates the palynological sample horizons. Sample MPA 49684 is the youngest from this vicinity and contains a CM Biozone assemblage with a high diversity, including Lophozonotriletes spp., cf. Ancistrospora spinosa Menéndez & Azcuy and Densisporites spp., which suggests similarities with the upper part of the Ballagan Formation on the Heads of Ayr foreshore (Stephenson, 2001a). Samples MPA 49683 and 49682 are slightly older but also contain upper Ballagan assemblages.

The sample MPA 49685 [NS 3808 0919] is considered to come from the lowest part of the section, and contains a sparse palynomorph assemblage with the distinctive Leiosphaeridia sp. A, which was found associated with the upper part of the formation on the Heads of Ayr foreshore. Also at the horizon of MPA 49685 east of Troquhain, Williams and Taylor (2001) recognised a very sparse ostracod fauna comprising Sulcella cf. affiliata? and a robust ?cavellinid. They remarked that the fauna is similar to horizons on the Ayr coast, which suggests that the ostracods come from the upper part of the formation. Assuming that the assemblage once lived at that horizon, the fauna is consistent with a stressed (probably not fully marine) environment.

The uppermost part of the Ballagan Formation north-east of Troquhain is not well exposed. However, red-brown, fine- to medium-grained sandstones, cropping out in Rodgerston Burn [NS 3800 0958], contain calcareous nodules which may indicate proximity to the overlying Clyde Sandstone Formation.

The calcareous, quartz-rich sandstones of the Ballagan Formation are in general medium- to coarse-grained, moderately to well-sorted, mature quartz-arenites (Phillips and Smith, 2008). The clast assemblage is dominated by monocrystalline quartz, which is unstrained to weakly strained. The shape of the quartz grains has been modified by local pressure solution (producing quartz overgrowths) and grain boundary etching, where there is carbonate cement. Some samples contain small, pebble-sized clasts of carbonate replaced mudstone or micritic limestone. Other minor to accessory detrital components include plagioclase, opaque minerals and tourmaline. The sandstones have a well-developed carbonate cement which appears to be replacing the original matrix or earlier cement. The carbonate cement contains some subhedral rhomb-shaped crystals, which suggest that the cement is or was dolomitic.

A laminated mudstone sample (N2584) was observed to be completely replaced by massive, fine-grained micritic carbonate. The silty layers and bases to normally graded laminae contain fine-grained clasts of monocrystalline quartz. Apart from white mica, the other detrital clasts in the rock are similar to those found in the sandstones.

A microconglomerate (N2578) sampled within the formation contains rounded to irregular clasts of limestone, carbonate replaced mudstone, sandstone and some siltstone, set in a slightly feldspathic, quartz-arenite matrix with a similar range of clasts to the sandstones.

Clyde Sandstone Formation (CYD)

The Clyde Sandstone Formation lies at the top of the Inverclyde Group, but because there is an unconformity at the top of the group, the formation is locally missing in this district (e.g. around Doonfoot). In the past this formation formed the lower part of the Calciferous Sandstone Measures as mapped in this district (Institute of Geological Sciences, 1978a).

The Clyde Sandstone Formation typically comprises red-bed arenaceous facies, and indicates that a major change in palaeoenvironment occurred at the top of the Ballagan Formation. The formation appears to represent a return to a fluvial system with numerous channels and bars, with intervening areas of overbank deposits where pedogenic carbonates developed. The development of pedogenic carbonates within fluvial deposits indicates a semi-arid, terrestrial environment.

The formation comprises brown, reddish brown and brick-red, medium-grained sandstones interbedded with reddish and, locally greyish siltstones and mudstones. The sandstones are locally coarse-grained to granular or pebbly. Some beds are calcareous with sporadic, pale grey calcrete nodules. However, no distinct calcrete horizons can be mapped, and the palaeoenvironment was probably more active than that of the Kinnesswood Formation which is finer grained in general, and contains mappable pedogenic limestone beds. The thickness of the formation is variable but reaches 240 m in the Dailly area.

In the Dalrymple area, the outcrop of the Clyde Sandstone Formation, is inferred from field evidence and palynological determinations (Smith, 2000b). East of Yonderton, in the stream section south of the River Doon at [NS 3969 1361], a grey, silty mudstone intercalated within reddish siltstones and sandstones contains a palynomorph assemblage (Stephenson, 2000c) including Lycospora pusilla, indicating the presence of the Pu Biozone of early Visean (Chadian) age. The assemblage is therefore younger than would be expected for a red-bed succession which formed part of the Kinnesswood Formation (or 'Upper Old Red Sandstone'), with which it was formerly correlated (Eyles et al., 1949). Consequently, this red-bed succession is correlated with the Clyde Sandstone Formation, and is considered to lie conformably above the Ballagan Formation cropping out to the west, but is faulted against older rocks to the north-east.

The stream section referred to above comprises fine- and medium-grained, red-brown sandstones, commonly about 0.2–0.3 m thick and including coarse, detrital micas. The beds fine upwards into rippled sandstones, locally containing mudstone intraclasts. In places the sandstones are calcareous and have a greenish-grey mottle. Some of the fine-grained sandstone interbeds, about 0.1 m thick, are parallel bedded. Prominent brown, medium- to coarse-grained sandstones, up to 0.5 m thick, contain scattered small pebbles but become more fine grained and parallel laminated towards their tops. Some of the granular to pebbly beds fine upwards rapidly over 60–70 mm. The pebbles, which include quartz, grey sandstone and microgranite clasts, are commonly angular to subrounded. In the upper part of the section, the sandstones are cross-bedded and up to 1 m thick. Local, small scale current ripples indicate that the palaeocurrents flowed toward the south-west. Reddish brown siltstone interbeds, up to 0.6 m thick, contain fine-grained, copper-coloured micas. One pale, greenish-grey, calcareous silty mudstone interbed contains hard, pale grey calcrete nodules up to 0.1 m in length near its base.

Fine- and medium-grained pink and red-brown quartzose sandstones, interbedded with siltstones and coarse-grained to pebbly sandstones, exposed for example in Dyrock Burn [NS 382 107], are also considered to be part of the Clyde Sandstone Formation. The formation is affected by faulting in the area south-east of the Kerse Loch Fault zone, but the beds generally dip eastwards below the Lawmuir Formation. Some of the harder sandstone beds, up to 0.4 m thick, are calcareous and some are locally reduced to produce a greenish mottle. However, no evidence of calcrete nodules was seen in these limited exposures. The clasts in the rudaceous beds include vein quartz, pink microgranite, and dark grey, pale grey and red-brown chert.

The formation is also exposed in the eastern part of Hannah Glen [NS 384 098], where a variable sequence of red-brown sandstones is correlated with similar strata to the north near Dalvennan [NS 382 107]. In Hannah Glen, the formation comprises purple to red-brown coarse- and medium-grained sandstones, with variable dips due to faulting and open folding about roughly north-south axes. The sandstones are commonly mottled green and are calcite-cemented with poorly developed micritic (calcrete) concretions. Bedding is generally massive, up to 1 m thick or, less commonly, cross bedded. Some sandstones are fine- and medium-grained and others contain granules and rounded pebbles up to 30 mm in length. The pebbles comprise vein quartz, wacke sandstone and micrite. One breccio-conglomerate bed consisted of micritic (calcrete) clasts set in a red sandstone matrix.

The Clyde Sandstone Formation crops out on both limbs of the Dailly Syncline. It is well exposed on the north-west limb in Glenton Glen, west and north of Glenton Farm (e.g. at [NS 2887 0422]; (Figure 11)), but poorly exposed to the south-east. In the former exposures the formation contains fine- to very coarse-grained, red-buff, locally green or grey sandstone with parallel lamination and scattered quartz clasts. Cross- and trough-bedding and scoured, erosive surfaces are common in the sandstone.

Clast-supported conglomerate beds are common, typically with erosive channelled and scoured bases (Figure 11), and containing poorly sorted, ungraded, angular quartz, chert, jasper and intraformational siltstone and sandstone clasts up to 0.15 m in diameter. Interbedded with the sandstone and conglomerate are strongly coloured, red-purple mudstone and siltstone with pedogenic carbonate nodules and pipes (probable rhizoconcretions). A sample collected for palynological analysis was barren.

North of Roan of Craigoch, the formation is moderately exposed in a stream section [NS 2924 0475] characterised by yellow-orange, medium- to coarse-grained flaggy sandstone, with rare cross- and trough-bedding and microconglomerate beds. The range of clast types in the sandstone appears to be quartz, chert, lithic and detrital mica. At Quarrelhill Burn [NS 2570 0235] the formation characteristically comprises deep red-purple mottled siltstone and sandstone. In Baldrennan Burn, no exposures exist between the Ballagan and Clyde Sandstone formations, but the transition must be abrupt. The basal unit of the Clyde Sandstone Formation west of Glenton at [NS 2849 0405] consists of fractured, carbonate-cemented, red-green mottled sandstone that is interpreted as a well-drained palaeosol, and passes up to red mudstone with pedogenic carbonate.

The thickness of the formation in the Dailly area is interpreted to vary from 160–240 m. The formation boundaries are moderately well constrained in the Glenton Glen area giving a thickness of approximately 240 m.

No outcrops of Clyde Sandstone Formation are known at the north-east end of the Dailly Syncline, but the formation is inferred to be at rockhead in the area between exposures of Lawmuir Formation and Ballagan Formation. Blocks and pebbles of medium-grained, red-brown sandstone and coarse-grained to granular, carbonate-cemented sandstone were found in the southern bank of the Water of Girvan at [NS 310 056], north-west of Knockroon. These are typical of the Clyde Sandstone Formation seen to the north-east in the Ayr district. It is therefore considered that, along with the evidence of its exposure in the Dailly area, the formation continues round the hinge zone of the Dailly Syncline in a belt between the Lawmuir and Ballagan formations.

A thin section (N4283) of a coarse-grained, brown mottled, carbonate-cemented sandstone block, found north-west of Knockroon, contains mineral and lithic clasts generally in the order of 1–2 mm, but some are up to 7 mm in length. The loosely packed clasts range from subangular to rounded in shape and are cemented by fine- to coarse-grained sparry calcite matrix. The lithic fragments include vein quartz, spherulitic chalcedony, coarse-twinned K-feldspar with a fine-grained inclusion of biotite, perthite, radiolarian chert, fine-grained sandstone, micrite, fine-grained volcanic rock and igneous quartz-feldspar-epidote rock.

Strathclyde Group

The Strathclyde Group (late Chadian to mid Brigantian) is a varied succession of sedimentary and volcanic rocks. The Clyde Plateau Volcanic Formation lies at the base in the west of the Midland Valley, and in the north-west it is overlain by reworked volcaniclastic sedimentary rocks belonging to the Kirkwood Formation. In the south-western part of the Midland Valley, including this district, the volcanic rocks (of which the only representative south of Ayr is the Greenan Castle Pyroclastic Member) are separated by an unconformity from the overlying Lawmuir Formation. The Lawmuir Formation is variable in facies but characterised by its carbonaceous content.

Clyde Plateau Volcanic Formation (CPV)

The Clyde Plateau Volcanic Formation is the result of a brief, but intense eruption of subaerial basalts from linear vent systems and later central volcanoes (Read et al., 2002). Although most of the volcanic formation crops out to the north of the Inchgotrick Fault, the Greenan Castle Pyroclastic Member and the Heads of Ayr Vent have been related to the formation.

Greenan Castle Pyroclastic Member (GCP)

The Greenan Castle Pyroclastic Member consists of well-bedded and massive, grey-green or purplish red-stained, lithic lapilli-tuff and coarse volcaniclastic sandstone, about 20–25 m thick, forming the promontory on which Greenan Castle stands. On either side of the promontory the softer sedimentary rocks were cut back by the sea at the time of the formation of the lowest (Holocene) raised beach. Minor faulting also affects the shore outcrops. The coarse volcaniclastic sandstone is well exposed in cliffs just to east of, and below, Greenan Castle [NS 3120 1933]. The beds dip east-north-eastwards (Plate 18) and are locally trough cross-bedded.

The reworked beds contain accretionary lapilli and subordinate sedimentary rock fragments. The material forming the reworked tuff is remarkably similar to that at the Heads of Ayr Vent, which is therefore considered to be the source of the pyroclasts. No fossils have been obtained from the member, but since the age of the underlying Ballagan Formation is Tournaisian and the overlying Lawmuir Formation is assigned to the mid Asbian–early Brigantian (Paterson et al., 1998, and this district), both the vent and the member are limited to a time within this range (i.e. probably early Visean, Chadian–Asbian). This age assignment correlates with the timing of active Clyde Plateau volcanism in the northern part of the Midland Valley.

Previous work by Tyrrell (1920) termed the rock 'a mugearitic agglomerate' and implied that the unit is conformable, with material sourced from a vent which may be concealed locally or lie out to sea (i.e. it was younger than and not directly related to the Heads of Ayr Vent). Eyles et al. (1929, 1949) described the member as bedded agglomerate lying within the 'Cementstone Group'. They also described its content as including fragments of sedimentary rock and lava from the 'Lower Old Red Sandstone', together with fragments of Ballagan Formation, pieces of fossil wood replaced by calcite, nodules of carbonated peridotite and an isolated large crystal of alkali feldspar. They concluded that the Heads of Ayr Vent was the source of the Greenan Castle Pyroclastic Member. This hypothesis is supported by the study of Stephenson et al. (2002).

A thin section (N2583) from this outcrop is of a poorly sorted, moderately to open packed, immature, lithic-rich, weakly laminated, coarse volcaniclastic sandstone (litharenite). The packing of the sandstone varies from very close to open, with the higher degree of packing occurring in the finer grained laminae. These finer grained laminae also exhibit a higher intensity of alteration/degradation of the unstable lithic clasts. Detrital grains are angular, subangular to slightly irregular in shape, with a low sphericity and a weakly developed preferred shape alignment of elongate clasts can be seen. The clasts are mainly composed of chloritised basaltic rock fragments which exhibit minor alteration to secondary carbonate. These volcanic lithic clasts range from medium-sand to granule-grade, and are enclosed within or coated by a thin oxide rim. Other minor to accessory detrital components include plagioclase, monocrystalline quartz, hematised rock fragments and chlorite-carbonate-quartz vein material or highly altered rock. Traces of a chloritic cement and pore-filling sparry carbonate cement occur. Compaction has resulted in localised pressure solution and embayment of unstable lithic clasts (Phillips and Smith, 2008).

Lawmuir Formation (LWM)

The Lawmuir Formation is the uppermost formation in the Strathclyde Group and, south of the Kerse Loch Fault, passes up conformably into the Lower Limestone Formation at the base of the Clackmannan Group. The formation typically varies from coarse-grained and pebbly sandstone to grey or reddish siltstone and mudstone, with impersistent seatrocks and coaly wisps. No marine limestone beds or coal seams have been found in this district, in contrast to the Hamilton area (Paterson et al., 1998). The limited exposures observed make it difficult to detail a depositional environment for the Lawmuir Formation, though it appears to be nonmarine, probably fluvial or fluviodeltaic, with some overbank areas rich in plants. The thicker, coarser-grained succession, forming the Lawmuir Formation at the eastern end of the Dailly Syncline, could result from a relatively long period (mid Asbian–early Brigantian) of fairly high energy, fluviatile to fluviodeltaic deposition prior to the marine incursions of the Lower Limestone Formation. Alternatively, a greater accommodation space may have developed in the hanging wall of the Kerse Loch Fault.

The base of the formation is unconformable on the Inverclyde Group in this district, except on the coast south of Ayr, where it overlies (probably unconformably) the Greenan Castle Pyroclastic Member. The formation was previously included in the upper part of the Upper Group in the Calciferous Sandstone Series (Eyles et al., 1949), and included in the Calciferous Sandstone Measures on the earlier 1:50 000 Solid edition of the geological map (Institute of Geological Sciences, 1978a). In the Belston area around [NS 385 285], south-east of Ayr, the formation appears to be absent due to non-deposition or erosion over the palaeohigh around Ayr, however, it has been interpreted to lie unconformably on the Ballagan Formation in boreholes penetrating below the Coal Measures Group east and north of Ayr e.g. Monktonhill Borehole, (NS32NW/4), [NS 3457 2849] . The Lawmuir Formation is also interpreted to occur at outcrop on the coast south of Ayr, where it forms the sandstone-dominated sections unconformably overstepped by the Passage Formation. Interbedded grey mudstones (e.g. MPA 49711 and MPA 49712) from the coastal section yielded spore assemblages suggestive of the NM Biozone (Stephenson, 2001c; Stephenson et al., 2002). As the Lawmuir Formation in this district mainly overlies the Clyde Sandstone Formation, considered to be Pu Biozone in age, the TS and TC biozones appear to be missing indicating that a significant unconformity lies at the base of the Lawmuir Formation.

Sections around the Dailly Syncline

In the Dailly area, in exposures just south of the Ayr sheet, Floyd (1999) described an abrupt change from the red sandstone and mudstone with pedogenic carbonates of the Inverclyde Group, to the grey sandstone and silty mudstone with carbonaceous debris and rootlets that are representative of the Lawmuir Formation. This lithological change between the Inverclyde and Strathclyde groups can be recognised in the wider Dailly area.

In the hinge zone of the Dailly Syncline, the bulk of the exposures around Little Craigfin are fine- and medium-grained, pale buff to brown sandstones. Some exposures near Little Craigfin [NS 302 055] include gritty to pebbly sandstones, which are greenish, brown or white, and thinly bedded to fairly massive. In one exposure close to the road [NS 3023 0544], a channel filled with pebbly sandstone over 1.2 m thick, cuts down into fine-grained sandstone with plant remains. Black and pale grey cherty pebbles, up to 30 mm long, are present in the rudaceous beds as well as vein quartz, 'felsite' and pale greenish mudstone. Locally the medium-grained sandstones include plant remains, coaly streaks, rootlets and carbonaceous flakes (Smith, 2003). These sandstones are usually buff-weathered with ochreous brown spots, and tend to be non-calcareous containing kaolin, probably from weathered feldspars. Near the top of the formation, in Black Glen [NS 3015 0528], fine-grained grey sandstones contain hard, iron-stained, probably sideritic concretions. Micaceous, fine-grained sandstone and siltstone interbeds are commonly rippled and cross-bedded in units 0.1–0.15 m thick. Subordinate reddish to greenish, sandy, silty mudstones and grey fireclays are interbedded. In the Water of Girvan beside the tributary junction at [NS 3038 0510], fine-grained, greyish sandstones containing some carbonaceous flakes occur in trough cross-bedded units about 0.5 m thick. From the disposition of the troughs it is inferred that the palaeocurrents flowed south-west (to 229°). In the river bank above this sandstone, a soft, reddish brown, silty mudstone bed, about 1 m thick, becomes buff-grey coloured towards the top and is sharply overlain by medium-grained brown sandstone.

In the southern bank of the Water of Girvan, north of Knockroon [NS 307 055], sections in the Lawmuir Formation dip up to 20° to the south-east, apparently lying on the north-west limb of a subsidiary syncline to the main Dailly Syncline. At the base of the section in the river is a soft, pale grey fireclay overlain by channelled sandstones with subordinate siltstones and mudstones. The pebbles in the pebbly bases include quartz, feldspar, black and greenish chert as well as grey mudstone intraclasts up to 10 mm in length.

Farther south, in the east bank of the Water of Girvan [NS 3040 0516], medium-grained, buff sandstone includes coaly wisps and a carbonaceous impression of a plant trunk up to 80 mm wide (Smith, 2003). In the south bank of the Water of Girvan [NS 3047 0507], pale grey and khaki coloured sandstones are intercalated with grey siltstones and seatearths, some of which are rubbly textured with plant stems or roots (Smith, 2003). The sandstones contain scattered white feldspars and white micas. On the northern bank these beds lie in open folds plunging gently to 228° south-westwards.

Due to lack of exposure, the base of the Lawmuir Formation can only be inferred as being folded around the Dailly Syncline (see section on Structure) and to lie unconformably above the Clyde Sandstone Formation. Its top is taken at the base of the Hurlet Limestone, which can be mapped around the syncline but is cut out by the Kerse Loch Fault to the west. The full thickness of the Lawmuir Formation in this area is difficult to determine because of the folding and faulting near the Kerse Loch Fault. The formation is about 30 m thick in the Glenton Glen area around [NS 291 044], on the northern limb of the Dailly Syncline, but is interpreted to be substantially thicker on the southern side of the syncline where exposure constraints are particularly poor. It is estimated to be about 100 m thick around the hinge near Little Craigfin around [NS 303 056], which is considerably more that the 50 m estimated to form the southern limb of the Dailly Syncline on the 1:50 000 Carrick Sheet (Floyd, 1999). The increase in thickness may be due to tectonic deformation, or the result of deepening of the basin adjacent to the syndepositional Kerse Loch Fault (Figure 12).

The relatively sparse spore assemblage (MPA 51248) from grey siltstones considered to come from low down within the Lawmuir Formation west of Knockroon [NS 3068 0551] indicates the possible presence of the NM Biozone (mid to late Asbian). Another spore assemblage (MPA 51244) from higher in the formation (Smith, 2003) near Black Glen [NS 3027 0526] suggests correlation with the younger VF Biozone (late Asbian to early Brigantian).

North-east of Dailly, the formation is moderately exposed in the stream north of Roan of Craigoch from [NS 2926 0463] to [NS 2930 0456]. The contact with the underlying Clyde Sandstone Formation is not exposed, but stratigraphical constraints considering the NM–VF late Visean miospore zonations in the hinge zone of the Dailly Syncline (Smith, 2003) mean the surface must be an unconformity. The Lawmuir Formation strata are dominantly green-grey, quartz-rich sandstones and siltstones with rootlets, coaly streaks and plant pieces up to 200 mm in length. Near Glenshalloch Wood at [NS 2929 0460], a layer of concretionary ironstone nodules was observed just above a rootlet bed, possibly indicative of an iron hardpan palaeosol. A slight disconformity between the Lawmuir Formation and the overlying Hurlet Limestone (base Lower Limestone Formation), exposed in this section at [NS 2930 0456], is not considered to represent a significant time gap. Beneath the Hurlet Limestone, erosively based, carbonate-cemented sandstone cross-sets cut down into finely laminated, rippled and bioturbated siltstone.

Grey-buff-green, fine- and medium-grained, parallel-bedded sandstone with rootlets and plant fragments, minor conglomerate and red mottled siltstone belong to the formation, as seen below the Hurlet Limestone south of Captain's Bridge [NS 2855 0357].

Thin sections of Lawmuir Formation sandstones from the eastern end of the Dailly Syncline show that the lower sandstones (N4280, N4284) contain mainly subangular to subrounded monocrystalline and polycrystalline quartz, in a compact texture with little matrix. Lithic grains include radiating chalcedony from vugs or amygdales, quartzite including some strongly deformed clasts, psammite, chert, iron ores, felsitic aggregates, fine-grained, poorly sorted sandstone and mudstone. Detrital feldspar and mica clasts are relatively rare. A sample (N4285) from higher up in the formation is grey and fine grained, with siltstone laminae and small-scale sedimentary structures. The finer grained bands commonly contain poorly sorted, altered volcanic clasts, biotite, chlorite and iron ores. They therefore seem to have a volcanic or tuffaceous component, which could mean they are partly equivalent to the Kirkwood Formation elsewhere (Paterson et al., 1998), or possibly partly derived from the slightly older Clyde Plateau Volcanic Formation.

Exposures and borehole data east of Dalrymple

A narrow strip of the Lawmuir Formation is inferred to lie east of Keirsmill around [NS 400 110], south of the Kerse Loch–Littlemill Fault zone. A small exposure of sandstone belonging to the formation has been recorded from this area at [NS 3995 1074], where it is intruded by a dolerite dyke. The formation is also exposed farther east at [NS 4003 1073] where it comprises reddish-brown and brown sandstones which include plant remains.

Near Martnaham Loch, the crop of Lawmuir Formation near [NS 390 168] is inferred from borehole data (Martnaham No. 1 (NS31NE/9) at [NS 39168 16452] and Martnaham No. 2 (NS31NE/8) at [NS 39583 16754] in which reddish and grey argillites, including an impersistent coaly mudstone, are interbedded with pale sandstones. The formation in these boreholes is about 16 m thick, and thought to have been deposited close to the palaeo-high near Ayr, which persisted throughout most of the Carboniferous Period.

Exposures and borehole data around Ayr

In the exposures near the coast south of Ayr, about 20 m of thinly laminated mudstones rich in plant debris pass upwards into medium-grained sandstones also containing plant debris (Stephenson et al., 2002). These beds lie unconformably above the Greenan Castle Pyroclastic Member and have been assigned to the mid to late Asbian NM Biozone, in agreement with Neves et al. (1973). The strata have therefore been correlated with the Lawmuir Formation.

Boreholes which penetrate the Clackmannan Group (e.g. Monktonhill, (NS32NW/4) [NS 3457 2849]) cut through about 5 m of pale grey, muddy sandstones with pyrite grains and thin brown and green siltstones above the Ballagan Formation. Included at the top of this section are 0.2 m of coal and carbonaceous, pyritic mudstone, which are probably equivalent to the Hurlet Coal elsewhere in the Midland Valley (Forsyth et al., 1996). This sequence is therefore correlated with the Lawmuir Formation. In Holm Borehole No. 6 (NS32SE/15) [NS 3751 2223] the equivalent beds are coarse-grained, grey, greenish and white sandstones, some of which were interpreted to contain volcanic detritus.

Just south of the River Doon at [NS 3248 1915], a fine-grained, massive, quartz-rich sandstone from the formation was sampled. The corresponding thin section (N2588) is of a fine-grained, closely packed, relatively mature, quartz-arenite containing small pockets or patches of replacive carbonate cement. Detrital grains are subangular to subrounded in shape with a low sphericity. The shape of the clasts is locally modified by the pressure solution of quartz and localised development of early quartz overgrowths. The clast assemblage is mainly composed of monocrystalline quartz. Other minor to accessory detrital components include polycrystalline quartz, degraded/altered rock fragments (protolith uncertain), K-feldspar, plagioclase, opaque minerals, white mica, epidote, variably chloritised biotite, chlorite and apatite.

Minor kinking of detrital micas was noted. Pressure solution appears to be responsible for the main form of cementation within this sandstone. However, minor amounts of a pore filling, intergranular, cryptocrystalline to very fine-grained, radial fibrous feldspathic or clay matrix or cement are also present. This matrix component is variably replaced by red-brown, hematite-stained carbonate. Carbonate includes rhomb-shaped crystals which may be either dolomite or pseudomorphs after dolomite (de-dolomitisation).

Clackmannan Group (CKN)

The Clackmannan Group comprises the Lower Limestone, Limestone Coal, Upper Limestone and Passage formations (Browne et al., 1999). These formations are characterised by cyclical sequences of sandstone, siltstone, mudstone, limestone, coal and seatearth. Marine limestones, which are persistent in the Lower and Upper Limestone formations in other parts of the Midland Valley, are less well developed in the Ayr district as a result of a topographical high persisting from the early Carboniferous into Coal Measures times. Exposures previously seen in Annfield Burn [NS 3644 2017] south of Ayr, showed the Passage Formation directly overlying the Inverclyde Group. It was therefore deduced that the Lower Limestone, Limestone Coal and Upper Limestone formations had pinched out over the palaeohigh around Ayr. South-east of the Kerse Loch Fault, in the Dailly Syncline, the formations are better developed, including a thicker coal-bearing sequence within the Limestone Coal Formation (Figure 12). Eyles et al. (1930; 1949) gave detailed accounts of these strata around the Ayrshire Coalfield, and Simpson and MacGregor (1932) for the Dailly Coalfield. Mykura (1967) produced a detailed revision of the strata above the Lower Limestone Formation for south-west Ayrshire. The formations were further described from boreholes and mine plans.

Lower Limestone Formation (LLGS): upper Visean–lower Pendleian

The Lower Limestone Formation was previously known as the Lower Limestone Group and is poorly developed south of Ayr, mainly as a result of the condensed sequence which pinches out over the palaeohigh in the area. Farther east, around Martnaham Loch, the formation is inferred to crop out from borehole evidence. A condensed limestone sequence was encountered at depth in boreholes Jelliston No. 1 (NS31NE/2) [NS 39825 17669] and Jelliston No. 2 (NS31NE/3) [NS 39551 17502], as well as Martnaham No. 1 (NS31NE/9) [NS 39168 16452] and No. 2 (NS31NE/8) [NS 39583 16754]. The formation is up to 8.25 m thick in these boreholes and is characterised by pale to dark grey marine limestone beds up to 1 m thick, interbedded with calcareous mudstones and siltstones locally containing calcareous and ironstone nodules. Some of the limestones are fossiliferous and interbedded with hard (probably calcite-cemented) sandstones. Commonly the condensed limestone units lie within calcareous mudstones and it is difficult to correlate them with the key marker-bed components, such as the Hurlet Limestone, established elsewhere in the Midland Valley. However, in the Monktonhill Borehole (NS32NW/4) [NS 3457 2849], north-west of Monkton, a total of 7.3 m of Lower Limestone Formation was penetrated, ranging from the basal Hurlet (Broadstone), through the Blackhall (Dockra), to the Top Hosie limestones.

South of the Kerse Loch Fault, around the Dailly Syncline, the Lower Limestone Formation is represented by a less condensed sequence of bioclastic marine limestone beds, calcareous mudstones and sandstones. It is estimated that the formation is between 10 and 15 m thick in the Dailly area, which is about twice the thicknesses found to the north of the Kerse Loch Fault, but still well below the typical 50 m+ for the Midland Valley (Read et al., 2002).

The exposures in Black Glen [NS 3008 0529] consist of dark grey, fine-grained argillaceous limestone with shell fragments. A spore assemblage (MPA 51245) from grey mudstone within the formation in Black Glen [NS 3008 0526] indicates that it lies in the range NM to NC biozones (Asbian to Pendleian; Stephenson, written communication, 2003). In better exposed sections in Captain's Glen, farther west [NS 2850 0355], the basal dark grey limestone bed (Captain's Bridge Limestone of Eyles et al., 1949) contains crinoidal debris and is correlated with the Hurlet Limestone (formerly Patna Limestone in the Patna area to the east). It varies from about 2 to 8 m thick. The overlying sandstones, siltstones and fireclays are 5 m thick and succeeded by two thin (about 0.15 m thick) marine limestone beds with interbedded calcareous mudstones, correlated with the Second and Top Hosie limestones (Eyles et al., 1949). These were recorded from the burn west-north-west of Dalzellowie [NS 2936 0456], and are considered to equate with the Doon Limestone (the two Upper Hosie Limestones) in the Patna area [NS 403 120] on the eastern edge of the Ayr sheet (Eyles et al., 1949). More details of the formation as it relates to the Dailly Coalfield are given by Simpson and MacGregor (1932) and Eyles et al., (1949). Full macrofaunal lists with biostratigraphical comments for each of the limestones in the formation in the Ayr and Maucline districts were given by Dean (2002).

The depositional environment was fluviodeltaic with brief but widespread cyclical marine incursions. The cycles are considered to reflect glacioeustatic rises and falls in sea level. The marine limestones mark the interglacial maximum flooding periods. Thick channel sandstones and subaerially exposed interfluves mark the major glacial lowstands. Other cycles are superimposed on these, producing complex cyclical successions that are commonly difficult to interpret (Read et al., 2002).

Limestone Coal Formation (LSC): lower Pendleian

The Limestone Coal Formation is defined as the succession between the the top of the Top Hosie Limestone and the base of the Index Limestone. It is dominated by cyclical fluviodeltaic deposits including significant coals which were formerly mined in the Dailly area.

South of the Kerse Loch Fault, the formation lies on the flanks of the Dailly Syncline (Figure 12) and typically contains interbedded sandstones, fireclays and coal seams. The coal is commonly cannel-like and interbeds of siltstones and mudstones are rare. The formation is between 49 and 92 m thick in this area, thinning south-westwards along the syncline into the Cairnhill area on the Carrick Sheet to the south. The succession is thickest in the central part of the syncline around Dalquharran [NS 270 023], where up to nine named coal seams were once worked. There is a marked thinning of the seams and a reduction in the number of coal-leaves in the south-west of the syncline, such that only three seams were workable in the thinner succession around Cairnhill [NS 234 001]. The reduction in thickness of the coals is less marked in the north-east of the syncline around Glenshalloch Wood [NS 298 050], but the three lowest seams are split into a large number of leaves (Mykura, 1967).

The lowest worked coal known as the Hartley Coal (Mykura, 1967; Institute of Geological Sciences, 1978a; Floyd, 1999), Rotten Coal (Simpson and MacGregor, 1932) or the Harvey Coal (Eyles et al., 1949) was between 1.17–2.9 m thick, forming one to four leaves. It was one of the most persistently worked seams. The Hartley Coal was recorded as around 1 m thick in the eastern (Kilkerran) end of the Dailly Coalfield.

The succeeding (in ascending order) Lower Splint, Craigie, Corral, Parrot, Ell (or Little), Ground, Main and Wee coals are detailed in Mykura (1967), Eyles et al., (1949) and Simpson and MacGregor (1932).

Since 1962, the natural exposures in the Dailly Syncline have been much reduced by waste from opencast mining and by forestry. This resurvey therefore focussed on the Limestone Coal Formation strata exposed by the disused Glenshalloch opencast site [NS 2975 0505]. The succession in the pit comprises the Hartley and Corral coals overlain by a thick quartzose sandstone. Southwards, an east-north-east trending fault drops down a succession of thin coals, seatrock and mudstones. Neither of the pit faces could be examined in detail because of rock instability. Towards the top of the formation a thick, massive and cross-bedded sandstone was once quarried at Brunston [NS 257 017]. This persistent sandstone, below the Index Limestone, occurs at the same horizon as the Patna Sandstone near Patna (Eyles et al., 1949). In the area 540 to 900 m north of Kilkerran Station near [NS 297 049], a coarse-grained pebbly sandstone, 12.8 m thick, rests on the Parrot Coal; this suggests that the overlying coal seams were eroded in this area prior to the deposition of the sandstone. The mean dip of the foresets in the sandstones is to the south-east and slump bedding indicates downslope movement to the south or south-south-east, away from the Kerse Loch Fault. The environment is interpreted to be an active fluviodeltaic system in which little argillaceous material was deposited.

A partial section in the lower part of the formation is exposed along Captain's Glen [NS 284 035]. In Black Glen [NS 3004 0530], near the base of the formation, fine-grained medium-bedded, pale grey to buff-weathered, root-bearing sandstones are interbedded with dark grey, subfissile mudstones, siltstones and thin ironstones. A spore assemblage (MPA 51246) from near the base of the formation in Black Glen [NS 3004 0531] indicates a possible NC Biozone age (Pendleian) (Stephenson, written communication, 2003). The Johnstone Shell Bed was tentatively correlated with a marine band below the Hartley Coal, and the Black Metals Marine Band with a Lingula band above the Corral Coal (Mykura, 1967). The macrofaunas of these marine bands in the Ayr and Mauchline districts are poor and mainly consist of brachiopods (especially inarticulate forms).) Faunal lists and reference details of Lingula-bands and a nonmarine fauna with Naiadites sp. were given by Dean (2002).

In the footwall block north-west of the Kerse Loch Fault, the formation is poorly developed. In the Martnaham Loch area, it was partially penetrated in boreholes Jelliston No. 1 (NS31NE/2) [NS 39825 17669] and Jelliston No. 2 (NS31NE/3) [NS 39551 17502], as well as Martnaham No. 1 (NS31NE/9) [NS 39168 16452] and No. 2 (NS31NE/8) [NS 39583 16754]. The succession mainly comprises sandstones with interbedded pale to dark grey mudstones, siltstones, coaly mudstones, coals and ironstones. Thin coals were intersected in the lower part of the formation. Towards its top, three thicker coal seams, up to 0.79 m thick, represent the Patna coals observed farther south, and are overlain by a thick sandstone unit (Eyles et al., 1949). The general thickness of the formation in this area is estimated to be 48 m but farther west it pinches out and it was not deposited over the palaeohigh near Ayr (Mykura, 1967).

North of Prestwick, in the deep Monktonhill Borehole (NS32NW/4) [NS 3457 2849], approximately 32 m of Limestone Coal Formation was proved (including 14.6 m of igneous sill), comprising sandstone, ganister and coaly traces.

Upper Limestone Formation (ULGS): Namurian, Pendleian to Arnsbergian

The Upper Limestone Formation includes the cyclic strata between the base of the Index Limestone and the unconformity at the base of the Passage Formation (E1 to E2 Goniatite zones, TK to SO Miospore zones). It is usually characterised by repeated upward-coarsening fluviodeltaic cycles including marine limestones and mudstones, followed by siltstones and sandstones, capped by seatrocks and thin coals (Browne et al., 1999). Descriptive details for the district are given in Eyles et al. (1930, 1949), Simpson and MacGregor (1932) and Mykura (1967). In Ayrshire, the north-easterly trending Kerse Loch Fault exerted a control on the thickness of the preserved sedimentary rocks in the same manner as occurred in the deposition of the Limestone Coal Formation, but to a greater extent (Eyles et al., 1949). The number and thickness of the limestones increases to the south-east of the fault as does the thickness of the interbedded coals.

In the area around Ayr, this formation is poorly developed and poorly exposed as a result of thinning of the variable sediment supply onto a palaeohigh, with non-deposition occurring over an area just south of Ayr, indicated by the formation's absence in the coastal section at Longhill Point [NS 316 194].

North of Ayr in the Monktonhill Borehole (NS32NW/4) [NS 3457 2849], a 10 m sequence of fossiliferous limestones (in part concretionary) and mudstones is referred to the formation. Exposures in a tributary to the Water of Coyle [NS 3958 2089] comprise patchily exposed, fine-grained calcareous sandstones, shelly mudstones and grey mudstones that appear to form a succession 15–20 m in thickness.

To the east of Ayr up to three limestones are recorded within the formation but they are not easily correlated between boreholes. A hard blue limestone about 1.5 m thick was recorded 3.28 m below the top of the formation in Whitehill Tileworks No. 4 Borehole (NS31NE/7) [NS 39983 17066], below brownish sandstones belonging to the Passage Formation. The thickness of the formation in this area is estimated to be 17–24 m and represents a condensed sequence compared to those in the Dailly Syncline and farther east in the Midland Valley, where it is commonly 100 m or more thick (Read et al., 2002).

In the Dailly Syncline, the formation lies partly faulted in the fold core. The top of the formation has been eroded but a thickness of at least 106–151.8 m is preserved. Exposures are few and the stratigraphical information is known from boreholes. According to Mykura (1967), although eight major marine horizons have been recognised south of the Kerse Loch Fault only the two lowest comprising the Index and the Keirs limestones, can be recognised throughout. The Index Limestone is 1.83 m thick and is a hard, grey, shelly bed above a brown-grey compact limestone. The fauna is characterised, as elsewhere, by an abundance of Latiproductus latissimus (J Sowerby) among other brachiopods. Dean (2002) listed the brachiopod- and/or mollusc-dominated macrofaunas of all the main limestones within the formation in the Ayr and Mauchline districts, and provided further commentary on correlation. In more complete borehole logs, the Index Limestone is overlain by about 14.6 m of mudstone which is commonly calcareous and shelly with a thin earthy limestone bed up to 0.23 m thick containing Fenestella sp., productoids, Bellerophon sp. and Orthoceras sp. Exposures of the Index Limestone are recorded from near Burning Hill Wood at [NS 2781 0331] and at Maxwell Colliery at [NS 2742 0296] and [NS 2747 0299] as well as Baldrennan Burn at [NS 228 040], above Cosyglen Bridge.

The overlying strata in the Baldrennan Burn section include pebbly sandstone, fissile mudstone with ironstone ribs and hard root-bearing sandstone, all underlain by a dark sandy limestone with brachiopods tentatively correlated with the Keirs Limestone (Eyles et al., 1949). The latter limestone has also been tentatively correlated with the Lyoncross Limestone elsewhere in south Ayrshire (Wilson, 1983). The overlying limestones may correlate with the main beds in south Ayrshire, i.e. the Orchard (possibly equivalent to the Dunaskin Limestone), Calmy (Waterside) and 'Pleans' (Drumgrange, Polquhairn and Dalharco limestones), but as in the Muikirk area which has an analogous position just south of the Kerse Loch Fault/Bankend Fault, there are other candidates for correlation such as the Huntershill Cement Limestone (Birchlaw Limestone of Davies, 1972). There are at least three coals, up to 0.91 m thick, recorded in the boreholes penetrating the formation. These usually occur below the limestone successions which are relatively thick, suggesting that marine incursions were prolonged, with little penecontemporary erosion (or little siliciclastic input) as the basin subsided.

Passage Formation (PGP): Arnsbergian–lower Langsettian

Across most of the Midland Valley the Passage Formation comprises mainly sandstones, with interbedded clayrocks and a few coals, mudstones, ironstones and limestones. Cyclical deposition becomes less well developed upwards and the formation was deposited in a more proximal, fluvial environment than the older formations in the Clackmannan Group. It includes the major mid Namurian break (Chokierian–Alportian), equivalent to the Mississippian–Pennsylvanian hiatus in North America, together with more minor breaks (Read et al., 2002).

The fossil evidence summarised by Dean (2002) points to a Namurian to Westphalian age (upper Arnsbergian to lower Langsettian, E2 to G2 Goniatite zones, SO to SS Miospore zones) for the formation. The lateral variation in thickness of the formation, and its lithofacies in Ayrshire were summarised by Mykura (1967), revising the former stratigraphy in which the formation made up most of the 'Millstone Grit' (see Simpson and MacGregor, 1932; Eyles et al., 1949). The Passage Formation is absent from the area between the Kerse Loch and Littlemill faults. In this area, strata above the Upper Limestone Formation limestones are assigned to the Lenisulcata Chronozone, within the Scottish Lower Coal Measures Formation (Mykura, 1967).

In Ayrshire, the subaerial volcanic episode which produced the Troon Volcanic Member interrupted the cyclic deposition prevalent elsewhere in the Midland Valley. In this district north of the Kerse Loch Fault, the Passage Formation is characterised by a basal arenaceous succession followed by a thicker member, the Troon Volcanic Member, mainly composed of olivine basalts, impersistantly overlain by volcanclastic sedimentary rocks. The base of the formation is erosional on the underlying Upper Limestone Formation, overstepping on to the Inverclyde Group over the palaeohigh around Ayr. The top of the Troon Volcanic Member may also be an erosion surface in this vicinity, since the Ayrshire Bauxitic Clay Member which formed by weathering of the top of the volcanic pile, is poorly represented around Ayr.

Borehole records in areas where the formation was completely penetrated, indicate that it has a variable thickness which is mainly due to the irregular thickness of the Troon Volcanic Member. For example, Holm Borehole No. 6 (NS32SE/15) [NS 3751 2223] near Thornyflat Farm, encountered 43.6 m of variegated sandy clay and ashy 'whin', whereas north of Ayr, the Monktonhill Borehole (NS32NW/4) [NS 3457 2849] penetrated 166 m of variegated ash, ashy clay and lava.

The arenaceous lower part of the formation is rarely exposed, but 28.6 m of mainly white to brownish sandstone with minor pale buff mudstone and fireclay were recorded in the Whitehill Tileworks Borehole No. 4 (NS31NE/7) [NS 39983 17066]. In the faulted succession exposed in the burn north-east of Milncraig at [NS 4014 2102], pale greenish and buff, fine and medium sandstones, with some quartzose beds up to 0.7 m thick, are interbedded with root-bearing siltstones below dark grey mudstones. At the top of the section, below the Troon Volcanic Member, silicified, baked, green and white sandstone rests on seatclay. West of Oldhall Wood [NS 403 217], a small faulted inlier consists of fine, quartzitic sandstone overlain by the Troon Volcanic Member. The base of the sandstone succession is not seen but the beds are baked against the overlying volcanic rocks.

Troon Volcanic Member (oBTVL)

This member extends south from the Dalry area in North Ayrshire to just south-east of Coylton (on Sheet 14E; Monro, 1999; Eyles et al., 1949). Mapping of the extent and thickness of the volcanic member around Ayrshire proved it is thickest (greater than 160 m) close to the Inchgotrick Fault (Monro, 1999), and the isopachytes suggest that there was contemporaneous movement on the fault. This volcanic field may have extended westwards, as similar volcanic rocks are known locally within the Passage Formation on Arran (north of Brodick at [NS 018 390]), and basalts with basaltic tuffs over 140 m thick are known from the Passage Formation in the Macrihanish Coalfield around [NR 660 213], interbedded with sandstones, coaly mudstones and a marine horizon (Stephenson and Gould, 1995). However, the original extent of the member is less than previously supposed (Monro, 1999), as no volcanic rocks of similar age are known from Northern Ireland (Mitchell, 2004), and the single basalt sheet in the Stranraer area at [NX 015 677] is considered to lie within Westphalian strata (Stone, 1995). Since the volcanic rocks in Ayrshire are centred in the Irvine district, more details of the member are given in Monro (1999). There, interbedded sedimentary rocks yielded a miospore assemblage from the KV Biozone, indicating an age within the Kinderscoutian to early Marsdenian stages of the Namurian. The eruptions therefore appear to have occurred after the marked fall in sea level that corresponds to the Mississippian–Pennsylvanian hiatus (Read, 1988). A minimum K-Ar age for the member was determined as 305 ± 6 Ma (De Souza, 1979; 1982), although recent U-Pb dating of Carboniferous–Permian lavas in the Midland Valley indicates that this needs revision and is probably in the range 317–315 Ma (Monaghan and Parrish, 2006). The lava pile has a maximum thickness just north of Troon (Monro, 1999), and appears to have poured out in near horizontal subaerial flows as the basin was sinking, with some coaly fireclays occurring between flows (Richey et al., 1930). The source of the lavas has not been identified; it may be centred below the lavas themselves or else lie offshore to the west.

The Troon lavas are almost all olivine basalts, with microphenocrysts of olivine pseudomorphed by dark red 'iddingsite'; some contain black clinopyroxene and plagioclase microphenocrysts (Monro, 1999). Their textures vary from coarse grained to weakly pilotaxitic microporphyritic basalts (Phillips, 1999). Whole-rock and trace element analyses of fresh lavas (none from the Ayr district) have shown that they form a 'transitional' series straddling the critical plane of silica undersaturation, almost equal numbers of these analyses were silica saturated and silica undersaturated (Monro, 1999; Macdonald et al., 1977; Wallis 1989). A plot of normative plagioclase composition against differentiation index shows that most of the lavas can be classified as basalts, although they range into the basaltic hawaiite field (Monro, 1999). Geochemical studies (Macdonald et al., 1977) show that these lavas start a second magmatic cycle of increasing silica undersaturation from mid Namurian into Permian times. They are classed as within-plate alkali basalts (Wallis, 1989). Olivine fractionation dominated the evolution of these lavas almost to the exclusion of clinopyroxene (Wallis, 1989). Wallis (1989) concluded that the basaltic lavas represented partial melts from a heterogeneous asthenospheric source with little crustal contamination.

In the Ayr district, the outcrop of the member extends from Longhill Point [NS 317 194] on the coast to near Milncraig [NS 403 206] inland. It comprises weathered basaltic and tuffaceous rocks and is therefore commonly rotten or poorly exposed. Much of the alteration is thought to be due to penecontemporaneous subaerial weathering in a wet tropical climate. From borehole evidence the member is known to thin from 166 m in the Monkton area to less than 10 m thick near Milncraig.

Two small inliers of rotten basalt occur to the north of Milncraig, both affected by faulting, one beside the River Ayr [NS 403 217] and the other south of the Water of Coyle around [NS 401 210]. The former, beside the River Ayr west of Oldhall Wood, exposes rotten, brown-weathering basalt of the Troon Volcanic Member overlying quartzitic sandstones. The basalts are up to 10 m thick, and overlain by seatearth belonging to the Scottish Lower Coal Measures Formation. The faulted section in the Water of Coyle [NS 4004 2104] exposes rotten olivine basalt, which is underlain by silicified green and white sandstone and is overlain by weathered sandstone, rich in volcanic clasts, at the base of the Scottish Lower Coal Measures Formation.

A third small inlier of olivine basalt and volcanic detritus, farther west at [NS 376 218], was formerly exposed along the River Ayr. These outcrops are considered to be directly overlain by the Scottish Lower Coal Measures Formation, but farther south-west, overlying reworked volcanic detritus and claystone at Holmston Quarry [NS 3640 2085] are assigned to the Passage Formation (see below).

Weathered amygdaloidal basalt is reported from Annfield Burn [NS 3558 1975] and site investigation boreholes in Kincaidston [NS 352 193]. A larger, tongue-shaped crop of volcanic rocks around Lochfergus near [NS 391 184] includes rubbly and spheroidally weathered olivine basalts with common small amygdales. Some of these lavas are coarse-grained and some have a weakly pilotaxitic groundmass (Phillips, 1999). Baked, grey, silty mudstone inclusions within the lavas are recorded at several localities but with little evidence of individual flows. Near Martnaham Loch, the Bowmanston Borehole (NS31NE/1) [NS 39742 18398] proved at least 57 m of red and green weathered basaltic rock unconformably overlying the Inverclyde Group. The outcrop pattern and the increasing thickness of the volcanic rocks provide evidence for the local erosion of the underlying rocks during the Namurian, prior to the volcanic outpouring. In the area to the north of Martnaham Loch, the broad tongue of lavas is considered to have infilled a palaeovalley eroded earlier in Passage Formation times (Eyles et al., 1949, p.76).

Farther north within the Ayr Coalfield, boreholes encountered the Troon Volcanic Member at depth, for example at Holmston [NS 3580 2112] and [NS 3568 2126].

Sedimentary strata above the Troon Volcanic Member are characterised by volcaniclastic rocks. Some may be pyroclastic, but most beds are too altered to be confidently interpreted.

Previous exposures at Holmston Quarry [NS 3640 2085] have been described as fine- to medium-grained, ashy sediments with some compact, grey, possibly bauxitic clay, red mudstone and thin carbonaceous bands (Eyles et al., 1949), or, well-stratified, brownish to purplish, porcellaneous claystone composed of altered lava detritus or tuff, alternating with fine unbedded claystone (Mykura, 1967). The section included some carbonaceous layers and beds with sphaerosiderite. These limited lithofacies appear to lie above the Troon Volcanic Member and may be laterally equivalent to the Ayrshire Bauxitic Clay Member established in the Irvine district (Monro, 1999), but are not extensive enough here to be shown separately on the 1:50 000 Bedrock map. None of the clays which were analysed from this horizon in the district has proved to be bauxitic, but one sample contained a high ferrous iron content and the clay was considered to be chamositic (Eyles et al., 1930). The Passage Formation exposed north of Gateside Farm [NS 3680 2119] to [NS 3694 2115] comprises a variety of bedded lithofacies including pale pink crystal tuff, purple tuff, basaltic lava, red lapilli-tuff with large angular fragments and grey-purple clay.

On the foreshore near Longhill Point [NS 3175 1949], brown and blue-green siltstone with clasts of volcanic rock form part of an impersistant lens of sedimentary rock lying above the Troon Volcanic Member near the coast.

Scottish Coal Measures Group (CMSC)

In the Ayr district, the Scottish Coal Measures Group extends from the coast at rockhead, below Ayr eastwards to Annbank, where it lies on the south-west flank of the Mauchline Basin. The strata consist of coal-bearing cyclic, fluviodeltaic deposits. Stratigraphically they range from the unconformity with the underlying Passage Formation to the unconformity with the overlying Mauchline Volcanic Formation. The marine band containing Gastrioceras subcrenatum (Schmidt), which is taken as the base of the Coal Measures in England and Wales, has not been found in Scotland, so the base of the Scottish Coal Measures is presently taken at the Lowstone Marine Band or its equivalent (Browne et al., 1999). The base of the Coal Measures in the Ayr district was formerly taken at the base of the Dalmellington Blackband Ironstone (Eyles et al., 1949), but a Lingula band above the 1.3 m coal in the Dalcairnie Burn Bore (NS40SE/7) [NS 4505 0364] referred to by Mykura (1967) is now taken to mark the local base, and as such it may correlate with the Lowstone Marine Band of central Scotland (see Dean, 2002). The stratigraphical palaeontology of the Westphalian of the Ayrshire Coalfield was summarised by Brand (1983) with modifications by Dean (2002).

The Scottish Coal Measures Group comprises the Scottish Lower, Middle and Upper Coal Measures formations and these are all represented in the district.

Scottish Lower Coal Measures Formation (LCMS): Langsettian

The Scottish Lower Coal Measures Formation comprises up to 75 m (but commonly 60 m; Mykura, 1967) of cyclical mudstones, siltstones, sandstones, seatrocks, thin coals and rare ironstones, deposited in a nonmarine environment with the exception of the Lingula bands. The top of the formation is taken at the top of the Vanderbeckei (Queenslie) Marine Band. This band is poorly developed in the vicinity of the Inchgotrick Fault, and this district may be marginal to the original basin of deposition of the marine band (Brand, 1977).

The base of the formation in this district is taken at the incoming of black carbonaceous mudstones above grey bauxitic clay assigned to the Passage Formation, as logged in the section through the formation penetrated by Enterkine No. 6 Borehole (NS42SW/12) [NS 4037 2323]. In this borehole the position of the Beoch Musselband horizon was identified about 31 m below the Lingula bearing mudstone representing the Vanderbeckei (Queeenslie) Marine Band. The Vanderbeckei Marine Band was also tentatively identified in the Pisgah Borehole (NS32NE/15) [NS 3960 2883]) where a Lingula bearing mudstone band also contained fish scales.

Although numerous boreholes penetrate the formation within the district, the coals, commonly between 0.2 and 0.7 m thick, can rarely be correlated; partly due to the many minor faults. Few fossil horizons have been identified within the formation (Mykura, 1967), although two mussel bands were recognised in the lower part of the formation in the Monktonhill Borehole (NS32NW/4) [NS 3457 2849]). A 0.1 m coal just above the Troon Volcanic Member (Passage Formation) in the Monktonhill Borehole may be the equivalent of the Raise Coal which lies just above the Ayshire Bauxitic Clay in the Irvine district (Monro, 1999). In the north of the district, it is apparent from borehole evidence (Mykura, 1967, (Plate 1)) that the Lower Coal Measures thin drastically over the volcanic pile that constitutes the Passage Formation in this area. North of the Kerse Loch Fault generally, the Lower Coal Measures are thin and none of the fossil horizons within the Lenisulcata Chronozone have been recognised in the condensed succession, only the Communis Chronozone and higher (Mykura, 1967).

In outcrops north of the Kerse Loch Fault, the base of the formation is marked by a thin black mudstone and thick clean, white, medium-grained sandstone lying above the Passage Formation north of Gateside Farm [NS 3680 2119], Belston Quarry [NS 3765 2100] and Water of Coyle [NS 3953 2099] (Sowerbutts and Smith, 1999). An erosively based channel sandstone is exposed in Witchbrae Wood [NS 3754 2148], underlain by over 40 m of interbedded siltstones, seatrocks, thin coals and fine-grained sandstones. North of Sundrum Castle [NS 409 214], cross-bedded, buff-weathered quartzose sandstones are predominately exposed in units 0.3–0.4 m thick. The evidence from foreset beds indicates that the palaeocurrents flowed to the north-west and south-east and occasionally south-west.

A typical section north of Maxwelton Wood [NS 4052 2125] comprises:-

Sandstone, fine-grained with interbedded siltstones	5.0 m
Siltstone, grey	0.1
Coal	0.12
Fireclay, root-bearing, weathered	1.0
Sandstone, iron-rich with interbedded pale grey siltstone	>0.25 m

Other thin coals are underlain by up to 0.5 m of coaly to shaly mudstone. Within the exposures at the Water of Coyle there are also beds of soft, pale grey mudstone with harder ironstone lenses. South of the fault near Privick Mill, the section along the River Ayr is poorly exposed but two thin interbedded coals are recorded as being formerly worked from the surface, and a non-marine bivalve, Carbonicola sp., was collected from black mudstone near the base of the Coal Measures [NS 4015 2178].

Near Ayr the Lower Coal Measures are exposed in Annfield Burn [NS 3501 1959] where thin bedded sandstones dip westwards and contain mica and carbonaceous flakes. Near the base of the formation, in a small burn east of Roodland [NS 3921 1941], fine- to medium-grained, pale quartzose sandstones contain some root-bearing beds, carbonaceous flakes, fine-grained pyritic concretions and grey siltstone interbeds. Farther up the succession in Annfield Burn, two carbonaceous mudstone samples were examined for palynomorphs (MPA 49707) [NS 3492 1952] and (MPA49708) [NS 3487 1954]. Both samples have diverse palynomorph assemblages in a very rich organic residue. The presence of Radiizonates aligerens (Knox) Staplin and Jansonius and absence of Florinites junior Potonie & Kremp and Microreticulatisporites nobilis (Wicher) Knox suggests that they can be assigned to the RA Biozone, which is late Langsettian in age (Stephenson, 2001c), confirming that the strata belong to the Lower Coal Measures Formation. Since the base of the RA Biozone correlates closely with the base of the Communis Chronozone (Nonmarine Bivalve Zone; Higgs and O'Connor, 2005), the identification of the RA Biozone at this level supports Mykura's correlation (1967) of Lower Coal Measures strata in south-west Ayrshire.

Farther up the succession, the Dalmellington Blackband Ironstone was encountered in Whitehill Tileworks Borehole No. 4 (NS31NE/7) [NS 39983 17066]), and a 0.25 m-thick coal was penetrated near the base of the formation.

Near Ayr, the top of the formation is marked by a marginal marine band, representing the Vanderbeckei (Queenslie) Marine Band, within fissile black mudstones exposed near Laigland Farm [NS 3835 2243] and Potterhill Glen [NS 3944 2137].

South of the Kerse Loch Fault in this district, the Lower Coal Measures Formation is only present (but not exposed) in a sliver north-east of Shankston Loch [NS 400 122], where it is downthrown between the Kerse Loch and Littlemill faults.

Scottish Middle Coal Measures Formation (MCMS): Duckmantian

The Scottish Middle Coal Measures Formation comprises similar cyclic lithofacies to the Lower Coal Measures, however, the coals are thicker and can be correlated regionally (Brand, 1983). The formation ranges from the Modiolaris to the Lower Similis-Pulchra chronozones. In the past, the Ayr Hard (oldest), Ayr Jewel, Ayr Diamond, Ayr Major, Crawfordston and Ayr Ell coals were mined within the formation (Figure 13), which is generally 75 m thick, although there is some variability apparently unrelated to faulting (Mykura, 1967; Brand, 1983). The abundant borehole data in the district indicate that coals are thin or absent in the area between Monkton [NS 360 276] and Underwood [NS 390 293], a zone with a high proportion of sandstone between the Diamond and Major coals, which may have been dominated by a channel system during the middle part of Middle Coal Measures times (Mykura, 1967). In the upper part of the formation, the Blackband and upper Ell marine bands are recognised in this area (Mykura, 1967), implying that a slightly lower palaeotopography developed. In the Heathfield area [NS 354 235], the interval between the Jewel and Major coals is commonly very sandy with the intervening Diamond Coal locally missing, suggesting that a channel system also ran through this area at that time.

The best exposures of the basal section occur in the banks of the River Ayr, near Laigland Farm [NS 3835 2243] by Auchincruive, where there are characteristic thick, cross-bedded and channelised sandstones. At the base of the formation [NS 3840 2243] the Vanderbeckei (Queenslie) Marine Band with Lingula mytilloides was recognised (Mykura, 1967). In complete borehole sections in the Enterkine area (Mykura, 1967), the lower part of the marine band was also observed to contain the ostracods Geisina arcuata (Bean) and Hollinella cf. bassleri (Knight); the latter is considered to be a marine form. A foraminifera facies (Brand, 1977) indicates a shallower, inshore aspect to the deposition of the unit, supporting the idea that there was less subsidence in this part of the Ayrshire basin. In the Wallacetown boreholes Nos.5 (NS32SW/56) [NS 34186 2181]) and 6 (NS32SW/57) [NS 34225 21738]), the Vanderbeckei Marine Band is represented by about 0.38 m of carbonaceous, silty mudstone containing Lingula, Spirorbis, some coprolites, ostracods and some poorly preserved mussels. Composite logged sections from north-east of Tarholm [NS 3954 2235] and Brockle Wood cliffs east of Auchincruive College [NS 3976 2362] illustrate typical successions (Figure 14).

The shore section at Bentfield [NS 3399 2482] exhibits coarsening upwards sandstone cycles and interbedded, rippled and bioturbated siltstones and sandstones dipping up to 20° north-east.

A section on the west bank of the River Ayr [NS 4003 2344] exposes 0.2 m of coal, partly replaced by secondary limestone, dipping 8° northwards in upward-coarsening fluviodeltaic cycles:-

Sandstone, fine to medium grained, rippled, micaeous	0.8 m+
Sandstone, fine grained, thin parallel bedded, iron-rich nodules	0.1
Seatearth, sandy, silty, pinkish grey with sharp top	1.0
Sandstone, fine grained, micaceous, interlaminated with siltstone	0.6
Mudstone, rubbly, pinkish purple-grey	0.5
Sandstone, fine grained, micaceous interbedded with siltstone	0.8
Sandstone, fine to medium grained, rippled, pale grey	0.15
Siltstone, grey, thin bedded, fining down into dark grey silty mudstone	1.5
Coal, black, with secondary limestone	0.2
Seatearth, muddy, rooty, dark grey	0.5+

The medium-bedded, medium-grained, pale buff sandstones are commonly cross-bedded, and some of the fine-grained sandstones have asymmetrical ripples indicating that palaeocurrents flowed westwards. A few sandstones in this section contain grey silty mudstone clasts, whereas others are purplish brown in colour due to high iron oxide content.

Near Privick Mill [NS 408 223], a relatively thick sandstone sequence is exposed which is reddish brown in colour due to its proximity to a substantial north-east-trending fault. In a faulted section along the River Ayr [NS 4008 2172], thick bedded and planar-laminated sandstones are exposed with foresets indicating that palaeocurrents flowed to the north-east. Interbedded root-bearing, silty sandstones contain ironstone nodules and interbedded mudstones contain thin ironstone lenses, plant fragments and flattened mussels. Large limestone/ironstone concretions (up to 2 m diameter) crop out in a few localities along the River Ayr (e.g. north of Tarholm, at [NS 3938 2263] enclosed within grey-black siltstones and mudstones. The large concretions consist of several superimposed spheres with concentric and radial cracks.

A description of the depositional environment can be built up (Table 4) from the various but limited recorded exposures. A scheme based on Guion et al. (1995) for Westphalian A (Langsettian) and B (Duckmantian) coal-bearing rocks in Northern England can be successfully applied. The main lithofacies observed on an upper delta plain are shown in (Table 4).

Some depositional cyclicity can be observed, for example in the coarsening upwards prograding lacustrine deltas capped by seatrock and coal. The overall depositional environment of the Middle Coal Measures in the Ayr district appears to be one of low relief on an upper delta plain, which at times was crossed by minor channels and where extensive mires built up from time to time to be swamped on occasion by marine incursions.

Scottish Upper Coal Measures Formation (UCMS): Bolsovian

The Scottish Upper Coal Measures Formation is characterised by a succession of red-purple, micaceous sandstones with subordinate siltstones, mudstones, thin limestones and rare marine bands. The reddish colouration is thought to be due to oxidation of the strata by groundwaters percolating down from the overlying Permian rocks, because in some cases the reddening is associated with pre- and post-Permian faulting. The formation also contains secondary gypsum veinlets as reported in the log of the Pisgah Borehole (NS32NE/15) [NS 3960 2883]).

However, some of the oxidation may have taken place during or soon after deposition. With the change of climate to one with seasonal aridity, coals disappeared and fluvially deposited red beds with pedogenic carbonates became dominant (Read et al., 2002).

The base of the Scottish Upper Coal Measures Formation in this district is taken at an arbitrary horizon believed to be equivalent to the base of the Aegiranum (Skipsey's) Marine Band, which is present farther east towards Cumnock. An arbitrary horizon is taken as even in continuous boreholes such as Pisgah (NS32NE/15) [NS 3960 2883]), in which limestone beds up to 0.3 m thick occur, no equivalent of the Aegirinum Marine Band has been identified. The top of the formation has been taken where the volcanic strata belonging to the Mauchline Volcanic Formation (Stewartry Group) appear.

The boreholes in the area prove up to 430 m of strata with strong lateral variations (Eyles et al., 1949; Mykura, 1967). There are no persistent coals and those present are apparently replaced by carbonate. The limestone beds within the Pisgah Borehole are variously described as reddish, sandy without fossils, pale purple tinged, and crystalline to fine-grained, pale grey limestone 'ribs' with ostracods and ?Spirorbis. There is no evidence that any of the limestones are marine, but which is pedogenic, biogenic or replacive is difficult to assess. The sole plant record of fine-ribbed Calamites sp. (horsetail) in the Pisgah Borehole may indicate unstable fluvial environments such as river point bars and lakesides (Read et al., 2002).

Small exposures of purplish and white sandstones occur in the north of the district, for example Pow Burn [NS 3758 2857]. The depositional environment is believed to have become more arid and fluvially dominated (Mykura, 1967) than during that of the underlying Middle and Lower Coal Measures.

Chapter 7 Permian

Stewartry Group

The Stewartry Group includes early Permian red-bed (mainly sandy) clastic strata with sedimentary breccias and basic volcanic rocks near its base. In this district the group crops out on the south-west margin of the Mauchline Basin and the lower boundary is the unconformity with the Scottish Coal Measures Group. Only the lower constituent of the basin (the Mauchline Volcanic Formation) is represented in the Ayr district as the upper unit (the Mauchline Sandstone Formation) is exposed only in the core of the basin which lies to the east on Sheet 14E.

The group is likely to infill the faulted eastern margin of the Permo-Triassic Carrick Basin (Figure 15), which is interpreted to lie offshore (McLean and Deegan, 1978). It is less likely that Triassic strata are present, as they are not preserved in other mainland basins in the Midland Valley or Southern Uplands, although they are in Arran.

A faulted subsidiary basin crops out onshore at Ballantrae [NS 086 835] in the Girvan district (Cameron et al., 1986; Stone, 1988), where supposed Permian breccio-conglomerate and red sandstones of the Park End Breccia and Corseclays Sandstone formations are estimated to be about 1000 m thick. These sedimentary rocks, which include aeolian sand grains, are considered to be continuous with the larger offshore outcrop in the Firth of Clyde (Brookfield, 1978). An early Permian age for these strata is more likely if the small lensoid dolerite sill or lava within the breccio-conglomerate (Cameron et al., 1986) is equivalent to the Mauchline Volcanic Formation. It is evident that away from a zone up to 20 km wide trending north-west to south-east through Mauchline, early Permian volcanic activity drops off rapidly and is not likely to be a big component in the offshore basin in the Firth of Clyde (compared to the Larne basin, see below). An offshore borehole (Chesher et al., 1972) drilled 13 km east of Ailsa Craig (at Lat. 55° 14.54' N Long. 4° 55.17' W) which is just to the south of the district, penetrated fine-grained, quartzose, brick red sandstone and siltstone at rockhead which can be correlated with the Corseclays Sandstone Formation at Ballantrae (Stone, 1988).

Mauchline Volcanic Formation (MVL): early Permian

The Mauchline Volcanic Formation comprises mainly basaltic lavas (B_MVL) with subordinate tuffs (Z^B_MVL) which lie mainly towards the base of the formation. They crop out on the south-western margin of the Mauchline Basin which is controlled by north-west to south-east faulting developed in an extensional setting. It is possible that these volcanic rocks were part of a more extensive field (Williamson, 2003a) as farther south-east, across into the Southern Uplands Terrane, olivine-basalts (Carron Basalt Formation) occur in a similar position within the Sanquhar and Thornhill basins. To the south-west, the north-west-trending Larne Basin includes the early Permian Inver Volcanic Formation (Mitchell, 2004) which is over 600 m thick and includes basaltic to trachyandesitic and trachytic rocks. Similar rift systems associated with alkaline magmatism developed across north-west Europe during the early Permian. Other rift-related magmatism is known from the North Sea, northern Germany and the Oslo Graben (Neumann et al., 2004). The rifting is possibly part of the southward propagation of the Norwegian–Greenland Sea rift system (Read et al. 2002).

The formation unconformably overlies the Scottish Upper Coal Measures, although with no marked discordance, and is overlain by the Mauchline Sandstone Formation farther east towards the centre of the Mauchline Basin. A lower unit comprising intercalated red tuffs, basic volcaniclastic conglomerates, red sandstones and mudstones underlies the lavas in this district where the formation reaches 36 m thick, but it is up to 235 m (770 ft) in the Redcraig Borehole (NS42NE/15) [NS 4747 2773]) towards the northern margin of the basin. This lower unit was also encountered in the Pisgah Borehole (NS32NE/15) [NS 3960 2883]) where the lowest 22.5 m was described as 'red and green conglomerate with red marl beds'.

The lateral variations in thickness and facies of the formation suggest that the ejecta erupted during localised events from a number of separate centres (Williamson, 2003b). This is supported by the 60 or so volcanic necks, including intrusions of olivine analcimite and monchiquite, which occur within a radius of 30 km of the Mauchline Basin.

The lava flows are generally of microporphyritic olivine basalt, although some basanites containing analcime or nepheline are recorded (Eyles et al., 1949). In addition to olivine, there are in places augite phenocrysts which commonly show the purplish brown tint characteristic of titanaugite (Eyles et al., 1949). Detailed petrographical descriptions have been published previously (Tyrrell, 1912, 1928a). Macdonald et al. (1977) and Wallis (1989) analysed several of the lavas from the formation; no analyses from the Ayr district are known but those relating to the Irvine district are reproduced by Monro (1999). Volcanic rocks associated with the intrusive Permian vents have been related to the Mauchline lavas (Wallis, 1989) and they are described below. The lavas have strongly alkaline characteristics and are mainly silica-undersaturated (nepheline normative), although Macdonald et al. (1977) and Wallis (1989) also analysed some mildly alkaline transitional (hypersthene-normative) types. The few basaltic hawaiites and hawaiites represent more differentiated rocks (Wallis, 1989).

The lavas are fairly well exposed around Barnweill Hill, at [NS 407 296] just north of the district on Sheet 22W (Monro, 1999). Here local 'trap topography' (prominent terracing) indicates several flows of olivine basalt; the bases of the flows are massive but the tops, which are extremely amygdaloidal, weather more readily. Pseudomorphed olivines are a common feature, giving the rocks a speckled appearance. Exposures of the formation on Sheet 14W are scarce but north-east of Pisgah [NS 405 288] and north of Bourtreebush [NS 409 283], fragments of altered purple-grey basalt have been ploughed up. Red-brown, tuffaceous, coarse and medium sandstones exposed near Bourtreebush at [NS 4006 2815] are considered to be part of the volcaniclastic sedimentary unit at the base of the formation.

Fossil evidence for the age of these strata is scant, but a probable early Permian (early Autunian; Assellian) age has been suggested on the basis of plant material from sedimentary rocks intercalated with the lava flows at the base of the formation in the River Ayr near Stairhill [NS 4521 2423] (Wagner, 1983; Cleal and Thomas, 1995; Dean, 2002). K/Ar whole rock dates of 281 to 278 Ma have been given for these lavas (De Souza, 1979; recalculated with more recent constants). A recent ⁴⁰Ar/³⁹Ar date on the uppermost Mauchline Volcanic Formation of 289 ± 5 Ma (Monaghan and Browne, 2010) is in agreement with the intercalated early Permian plant material, and within error of the younger dates on alkaline intrusions in Ayrshire (about 298–292 Ma; Monaghan and Pringle, 2004).

The thin interbeds of red sandstone with rounded grains indicate that the lavas erupted subaerially and intermittently in a climate which appears to have become increasingly arid. The lack of discordance with the underlying upper Carboniferous strata suggests that any post-Bolsovian to early Permian earth movements were not strong in this district and that the topography was subdued at that time. The development of a north-west to south-east trending half graben during the early Permian could partly account for the present outcrop pattern of the Mauchline Basin.

Chapter 8 Intrusive igneous rocks

Siluro-Devonian intrusions

The minor intrusions of Siluro-Devonian age in this district are relatively widespread compared to other parts of the Midland Valley. They range from mafic dolerites, basalts and andesites, to dacites and felsites. The intrusions are generally related to the Carrick Volcanic Formation in which subvolcanic intrusion/volcanic eruption appears to have peaked in Early Devonian times (Thirlwall, 1988), although volcaniclastic debris occurs towards the top of the Siluro-Devonian Swanshaw Sandstone Formation. Most of the intrusions are sills or laccoliths intruded into levels of the Swanshaw Sandstone Formation below the Carrick Volcanic Formation. Basaltic andesite sills which intruded wet sediment are included in the Carrick Volcanic Formation, and were probably intruded between 415 and 410 Ma. Controls on the Siluro-Devonian magmatism were thought to be due to partial melting at a destructive plate margin (Thirlwall, 1981). A development of this model proposed Andean-type subduction with roll-back (Oliver, 2002). Alternatively, since this magmatism continues well after collision and across the suture, a model involving slab break-off causing a rising of the underplate accompanied by melting (Atherton and Ghani, 2002; Oliver, 2002) has been supported more recently (Macdonald and Fettes, 2007).

Midland Valley Siluro-Devonian Mafic Intrusion Suite

The Midland Valley Siluro-Devonian Mafic Intrusion Suite comprises doleritic (microgabbroic) to andesitic and lamprophyric rocks, which are commonly associated with the main phase of volcanism of Early Devonian age prevalent across the Midland Valley. In the Maybole–Straiton–Dalmellington area there is a distinct set of mafic intrusions belonging to this suite.

Dolerite and basalt, quartz-dolerite and quartz-basalt (QD_D and D_D)

The Craigdow Hill and Lochspouts topography is underlain by a quartz-dolerite sill (at for example [NS 2835 0584]). A similar smaller sill at [NS 300 063] lies to the north of Craigfin. The two quartz-dolerite sills were classed as of uncertain age (Institute of Geological Sciences, 1978a), perhaps because their petrological affinity was most closely allied to Paleocene intrusive rocks (Sowerbutts, 2000). However, Eyles et al. (1949) considered them to be of Early Devonian age and that is considered the most likely conclusion here, although the northern sill must be unconformably overlain by Ballagan Formation at [NS 302 064] if the palynological dating is correct. These sills are generally fine-grained and weakly plagioclase microphyric (Phillips, 1999) and a fine-grained sample (S25231) from the thinner northern sill on the sheet is described more specifically as altered quartz-bearing, plagioclase microphyric basalt. Accessory cubic crystals of pyrite occur within the dolerite.

The Early Devonian age for the quartz-dolerite sills is most likely since there is a close spatial relationship with the Early Devonian andesitic sills which also intrude the Swanshaw Sandstone Formation. Both sets of sills are cut by the major north-east-trending Kerse Loch Fault. North-easterly trending faults are not believed to have had major displacements in the Cenozoic, and elsewhere intrusions of Paleocene age continue unbroken/undeflected across faults such as the Southern Upland Fault. The quartz-dolerite sills are also cut by other smaller faults with various orientations (west-north-west to north-east); it is unlikely that all of these faults are Paleocene or later. These quartz-dolerite intrusions lie well to the south of the late Carboniferous suite of similar composition, and do not appear to be part of it. Close to the Kerse Loch Fault at [NS 3016 0616], major sinistral shear joints striking 030° and dipping 78° north-west within dolerite are consistent with the model of sinistral shear movement on the Kerse Loch Fault in Devonian to early Carboniferous times (see Structural Section). The sills are similar to those near Knockgardner south of the Straiton Fault (Smith, 2001).

An altered quartz-basalt sill (Phillips, 1999) has been mapped at Hollow Dale [NS 243 070]. The alteration of the rock, and affinity with other mafic intrusions in the area, suggest that it is probably part of the Midland Valley Siluro-Devonian Mafic Intrusion suite. A basaltic sill within the Swanshaw Sandstone Formation around [NS 223 083], extending from south of Glenside to the shore at Culzean Bay, is also considered part of this suite.

Andesitic sills (A_D)

The high-level fine- to very fine-grained, feldspar-phyric andesite intrusions west of the Kerse Loch Fault are probably related to the high-level microdioritic (or basaltic andesite) intrusions and the extrusive Lower Devonian Carrick Volcanic Formation.

Kildoon Hill [NS 295 072], on the ridge north of Lower Burncrooks, and that including Knockmill Quarry [NS 2935 0680], south of Lower Burncrooks, are both formed of andesitic sills. The andesitic sill forming Kildoon Hill extends eastwards to [NS 307 070] and west-north-west to [NS 282 079]. A ground magnetometer survey indicates that this andesite intrusion extends westwards at depth and is probably related to the sills north of Hollowshean (see below), which the magnetometer survey showed extended eastwards at depth.

In earlier surveys the sill at Kildoon Hill was described in the field as a pink 'porphyrite' or felsite, and on the earlier 1:50 000 map (Institute of Geological Sciences, 1978a) as a 'plagiophyre' containing highly altered, sparse phenocysts of albitic feldspar. More recently, the intrusion (S25228) has been described as a plagioclase-microphyric andesite with secondary quartz (Phillips, 1999). The altered red-green-grey coloured rocks are massive and jointed (Sowerbutts, 2000). The varied colour partly reflects alteration but also reflects a range in the content of pinkish feldspar and dark green mafic microphenocrysts. This intrusion may be sill-like and up to 35 m thick, but it has local steep and cross-cutting margins, chilled against the Swanshaw Sandstone Formation where formerly exposed south of Barmody Hill [NS 3066 0688]. The intrusion is cut out to the east by the Kerse Loch Fault and offset by a smaller north-east-trending fault [NS 303 072]. Thin baryte veins, up to 15 cm thick, striking 356° and dipping 70° to the west cut the body in an old quarry at [NS 3028 0736].

A pale grey andesitic sill to the north of Craigfin, which was formerly quarried at [NS 302 065], is estimated to be about 40 m thick and was intruded at a lower level within the Swanshaw Sandstone. The rock alters to a brown to khaki colour, but small pink and clear feldspar phenocrysts are scattered in the fine-grained matrix. It also contains scattered small chlorite- and chalcedony-filled amygdales. Petrographically this sill is similar to the body on Barmody Hill to the north, and is described as an altered feldspar microphyric andesite with secondary quartz (Phillips, 1999, (S25229)).

A sill of andesitic composition is exposed north of Hollowshean [NS 244 061] on a ridge trending east-west, offset by north-westerly trending faults. A similarly trending sill (or leaf) crops out to the north, on Knox Hill [NS 237 063] (Phillips, 2001), and is of andesitic affinity (quartz-bearing andesite to microdiorite). This intrusion had previously been described as an Early Devonian plagioclase-phyric olivine and quartz dolerite (Tyrrell, 1914), Early Devonian quartz-dolerite (Eyles et al., 1949), or an unclassed mafic intrusion (Institute of Geological Sciences, 1978a).

A pale grey intermediate igneous rock (N3259) from Knox Hill Quarry [NS 237 063] is an altered, fine- to medium-grained, pilotaxitic, quartz-bearing andesite or microdiorite. The rock is composed of shape-aligned plagioclase, chlorite and quartz with minor to accessory carbonate, sericite, opaque minerals and apatite. Quartz is a common minor to accessory phase and occurs interstitial or intersertal to earlier formed plagioclase. This microtextural relationship suggests that quartz forms part of the primary igneous mineral assemblage. Chlorite (± opaque oxide) forms pseudomorphs after a ferromagnesian mineral; possibly mainly amphibole with some biotite. Rare carbonate and bowlingite pseudomorphs after pyroxene microphenocrysts also occur within this andesitic to microdioritic rock. It is petrographically similar to the andesitic rock (S25581) from near Hollowshean. Sample (S25580), also from Knox Hill Quarry, is a microdioritic rock, possibly a slightly coarser-grained variety of (N3259) and (S25581), and also contains minor K-feldspar (Phillips, 2001).

A porphyritic andesite sill or dyke-like body, which is considered to be Early Devonian in age and similar in composition to the sills described above, intrudes the Swanshaw Sandstone Formation near Kirkmichael Church [NS 3450 0891]. The andesite is grey with feldspar phenocrysts 1–2 mm in length. The intrusion is at least 10 m wide and has a steep margin to the west in contact with pinkish baked sandstone. Its eastern margin is faulted against purple-brown sandstone.

Microdioritic intrusions (P_D)

In the Straiton area, several microdioritic intrusions, mostly small sills and some dykes, are associated with the Early Devonian basalts and andesites of the Carrick Volcanic Formation. The sills commonly occur within the Swanshaw Sandstone Formation, up to 100 m stratigraphically below the base of the Carrick Volcanic Formation e.g. at [NS 3815 0175]. Many were previously termed 'plagiophyres' (Institute of Geological Sciences, 1978a). These bodies tend to be more compact and less vesicular than the basaltic andesite lavas or sheets. The microdioritic rocks are generally grey to grey-green in colour, containing intermediate plagioclase, pyroxene or hornblende phenocrysts in a fine-grained matrix of the same minerals together with some iron oxides.

East of Kirkmichael, an irregular microdioritic intrusion is exposed beside Dyrock Burn [NS 3515 0909] with an intrusion breccia in the Swanshaw Sandstone Formation at its margin. The intrusion has a grey to purplish grey, fine-grained feldspathic matrix containing altered dark green phenocryts up to 1 mm in length.

A microdioritic or 'porphyrite' sill, over 10 m thick, is exposed in a former quarry at [NS 3552 1067], south of Guiltree Hill. This microdioritic rock has feldspar laths set in a dark greenish grey matrix. The rock is blocky jointed with local spheroidal weathering. The sill intrudes the Swanshaw Sandstone Formation below the conglomerate correlated with the Isle Port Conglomerate Member, and leaves of the sill are intercalated with the formation to the east of Guiltreehill Farm [NS 3635 1072].

Kersantite (LK_D)

These lamprophyric rocks are compact, red or grey in colour, and are generally microporphyritic and altered. They contain biotite, altered pyroxene and olivine, plagioclase and iron oxide, together with alteration products including carbonate, chlorite, white mica, albite and opaque oxides. Two textural varieties of biotite have been recognised (Phillips, 1999) in these rocks: (a) early formed anhedral flakes locally weakly aligned to form a weakly developed pilotaxitic fabric, and, (b) later interstitial, sub-poikilitic plates, which in places, partially enclose earlier plagioclase. The kersantite intrusions are considered to be younger than the Knockskae dacite on evidence found west of this area. North of Knockskae, the kersantites occur mainly as sill-like bodies e.g. [NS 357 016] associated with the larger Knockskae dacite intusion. The sill-like, medium-grained kersanite intrusion near Craigenton at [NS 231 053] is up to 15 m thick and contains biotite and pyroxene.

Midland Valley Siluro-Devonian Felsic Intrusion Suite

The Midland Valley Siluro-Devonian Felsic Intrusion Suite ranges from intrusive microgranitic rocks and rhyolites to dacites, many of which are micro- or macroporphyritic. They are associated with the silicic component that evolved during the widespread Siluro-Devonian calc-alkaline magmatic episode. They are particularly common on the southern margin of the Midland Valley.

Dacite (RD_D)

The large intrusive body exposed on Knockskae [NS 362 011], Back Fell and extending westwards onto Clauchrie Hill [NS 346 016], is classified as dacite (Phillips, 1999) as it contains quartz and mainly altered plagioclase feldspar. It was formerly termed an acid porphyrite (Eyles et al., 1949). The intrusive rock is relatively massive and resistant and so forms the Knockskae–Glenalla Fell–Clauchrie Hill range extending a distance of 3.5 km. The body appears to be a hypabyssal laccolith intruding the Swanshaw Sandstone Formation. Eyles et al. (1949, fig. 4) interpreted the intrusion as a sill up to 700 m thick, however, modelled as a laccolith it may only be 125 m thick. Similar dacitic bodies are more common on the adjacent Sheet 14E in the ground north-west of the Southern Upland Fault.

The fine- to medium-grained microporphyritic rock varies from pale grey to pink with orange weathering zones. Pale weathered, small feldspar phenocrysts and biotite can be seen in hand specimen. In thin section, it can be seen to have a pilotaxitic texure. The rock tends to be highly altered and now comprises an assemblage of carbonate, quartz, albite, opaque oxides and clay minerals. In places, the cores of zoned plagioclase phenocrysts are preferentially altered. Chloritic (± opaque oxides) and/or carbonate pseudomorphs after possible amphibole are present in some thin sections. Small, commonly chloritised biotite flakes occur in the groundmass of the more evolved rocks. On the south side of Knockskae the rock is shattered by a fault and a hematite coating occurs on the joint surfaces. Adjacent to the intrusion the Swanshaw Sandstone Formation is baked to a pale buff colour.

Sample N2971 from a burn north-east of Drumyork Hill [NS 3418 0266] is of microporphyritic dacite containing plagioclase (up to 2.5 mm long) and amphibole phenocrysts in a fine-grained fabric of flow-aligned feldspar laths, with quartz, opaque oxides and accessory apatite in the groundmass. Alteration produced the accompanying cryptocrystalline chlorite and carbonate. Microscopic shear bands deforming the foliation are considered to have formed in response to ductile shearing during emplacement. Other microporphyritic dacite samples (N2972) and (N2973) from north of Clauchrie Hill contain biotite and possible amphibole phenocrysts and altered feldspar (now partly albitised plagioclase) set in a fine-grained groundmass, locally pilotaxitic, including irregular patches of very fine-grained quartz (Phillips, 2002).

Microgranitic rocks (F_D)

Microcrystalline rocks of granitic composition ('felsite') occur as a sill or laccolith north-west of Drummochreen Cairn [NS 275 053]. This body comprises pinkish, medium- or fine-grained, and altered felsic igneous rocks, but no petrographical analysis has been performed on them. They have previously been described as Early Devonian 'plagiophyres' (Institute of Geological Sciences, 1978a). This body is up to 50 m thick with a known strike length of about 3 km. It is slightly transgressive within the Swanshaw Sandstone Formation but in the east it exploits a horizon close to the unconformity with the Craighead Inlier.

A small, pale pink, aphyric microgranitic sill, about 1 m thick, is exposed [NS 3648 0218] in a burn north of Knockskae and is probably related to the Knockskae dacite intrusion.

North Britain Siluro-Devonian Plug and Vent Suite

Mochrum Hill Vent (Z_MH)

The prominent Mochrum Hill, west of Maybole [NS 265 101], is formed by an intrusive breccia with associated minor basaltic and basaltic andesite intrusions. This roughly elliptical body (about 1 km by 0.5 km) is composed mainly of poorly sorted greyish, buff and reddish purple, fine-grained basaltic and amydaloidal basaltic andesite fragments, set in a fine-sand to granule-grade, khaki-coloured or buff, volcaniclastic sandstone matrix. The matrix also contains clasts of fine-grained reddish sandstone, chert and quartz. Most of the angular to subrounded volcanic fragments are about 10–200 mm in length, and exceptionally up to 0.5 m; larger bodies of aphyric, greenish grey basalt and amydaloidal basalt may be blocks rather than intrusive bodies. Fresher basaltic rocks, exposed at the north-western and eastern margins of the body, are related intrusions. In the disused small quarry on the eastern side [NS 2715 1012], basalt contains small plagioclase and mafic phenocrysts. Amygdales are sparse and are filled with jasper and calcite. The basalt is locally spheroidally weathered and cut by carbonate- and hematite-bearing veinlets. It is described as a plagioclase-clinopyroxene microphyric basalt (Phillips, 1999). Two small dykes associated with the vent are exposed on the south side of the hill (Sowerbutts, 2000). One of these is also basaltic but the other is a pilotaxitic, plagioclase-microphyric basaltic andesite (Phillips, 1999). Most of the volcanic fragments within the vent are altered olivine-plagioclase-microphyric basalts, but they are mixed with olivine- and plagioclase-microphyric-basalts, and pilotaxitic and vesicular textures are variably developed. In thin section the volcaniclastic matrix to the breccia contains unsorted, commonly angular clasts of quartz and more rarely quartzite and sandstone, some of which is recystallised (?thermally metamorphosed). Just to the north-west of the body, country rock sandstones are interbedded with red siltstones that dip up to 40° into the vent. On the south side of Mochrum Hill [NS 2650 0990], near the contact with volcanic breccia, the Swanshaw sandstones are steeply dipping, baked and flaggy, indicating that the volcanic rocks disrupted and metamorphosed the nearby sandstones.

Small exposures on the southern side of the hill, for example at [NS 2645 0996], comprise a volcaniclastic conglomerate with a volcaniclastic ashy matrix. Clasts reach up to 0.50 m in diameter. It is uncertain whether these rocks represent a primary vent-fill (Eyles et al., 1949), or are epiclastic with reworked quartz clasts as is suggested by petrographical microscopy (Phillips, 1999; 2000). However, it is possible that the quartz clasts became disrupted and dispersed from sediments as the vent drilled through the country rock, and hence are true pyroclasts. A lithic-rich microbreccia from the vent (N2062) was examined (Phillips, 2000) and found to be clast- to matrix-supported, in a fine to medium sandstone matrix. The larger lithic clasts are angular to subangular in shape and range from 0.8 up to about 7.0 mm. These clasts are mainly composed of variably altered basaltic to andesitic volcanic rock fragments, which in some cases contain small quartz xenocrysts. The irregular lithic clasts are compositionally/petrographically similar, indicating that they were derived from the same source and had a limited amount of transport. The matrix appears to contain detrital grains ranging from angular to subrounded in shape and mainly composed of volcanic rock, with subordinate quartz and minor plagioclase. The volcanic fragments in the matrix are andesitic to dacitic in composition (i.e. more evolved than the larger clasts). Other detrital components of the matrix include opaque minerals, muscovite/white mica, apatite, chlorite, garnet and altered biotite. A minor component of the matrix appears to have been derived from degraded lithic clasts (Phillips, 2000).

Mochrum Hill has been interpreted previously as a volcanic vent of 'Lower Old Red Sandstone' age (Eyles et al., 1949) and possibly Early Devonian (Smith, 2000a). However, most of the vent rocks have probably been eroded away, leaving the ellipitical plan of the neck to the volcano. The vent rocks could be the source of the reworked volcaniclastic sandstones within the Swanshaw Sandstone Formation which crop out about 2 km to the north-east. These volcaniclastic beds lie at a slightly higher stratigraphical level than the country rocks surrounding the vent/neck. The sandstone formation is considered to be Siluro-Devonian in age and hence the Mochrum 'volcano' may be earlier than (and not directly related to) the overlying Carrick Volcanic Formation. Support for this hypothesis comes from the geochemical analysis of a basalt (N2155) sampled from the eastern side of Mochrum Hill, which was probably intruded into the neck, which has a content lower than Carrick Volcanic Formation samples (Phillips and Smith, 2008). It also has a different trace-element content to the Carrick volcanic rocks, for example lower Zr and Ce and higher Zn and Co values. A thin section of (N2155) shows an altered microporphyritic basaltic rock, containing rare, partially resorped (relict) quartz xenocrysts (Phillips, 2000).

Carlandcheek Vent (Z_CC)

At Carlandcheek [NS 2756 1817] near Bracken Bay, a 50 m wide zone of intrusive volcanic breccia is exposed (Sowerbutts, 1999). It consists of angular clasts of porphyritic, aphryic and amgdaloidal basaltic andesite, in a fine-grained, reddish grey igneous rock. In some patches, the more-massive, grey, aphyric basaltic andesite is dominant. The Carlandcheek exposures have previously been interpreted as 'Lower Old Red Sandstone' vent rocks, and the area identified as a possible supply point for the Carrick Volcanic Formation (Eyles et al. 1949). Another possibility, as the breccias are located along a probable fault line and next to a dyke, is that they could be an Early Devonian fault breccia surrounded by later intrusive material.

North Britain Siluro-Devonian Calc-alkaline Dyke Suite

Andesite (A^B)

Small dykes of porphyritic basaltic andesite or andesite, cutting the Mochrum Hill vent/neck, are exposed on the south side of the hill, (for example at [NS 2675 0975]). These igneous rock types are comparable in composition to those in the Carrick Volcanic Formation.

A small east-west-trending dyke or sill, exposed in the burn to the east of Kilhenzie Castle [NS 3096 0827], is at least 2 m thick. It is an 'andesite' with pink feldspar and dark green mafic phenocrysts, grading into a grey, finer-grained margin against baked and locally hornfelsed Swanshaw Sandstone Formation. Thirty metres south-east of Kilhenzie Castle [NS 3087 0821], a grey, fine-grained hornfels may be associated with another small andesite dyke.

Carboniferous to Permian intrusions

Early Carboniferous intrusions

Heads of Ayr Vent

The Heads of Ayr Vent (ZHA) is an isolated component of the widespread volcanic province which erupted in the Midland Valley of Scotland in early Carboniferous times. It is roughly contemporaneous with the Clyde Plateau Volcanic Formation (335+/−2 Ma; Monaghan and Pringle, 2004: Monaghan and Parrish, 2006).

It forms a well-exposed, prominent landmark (Plate 19) on the coast, 5 km south-west of Ayr. It is interpreted as a volcanic vent related to a moderately sized volcano. Geikie et al. (1869 p.17) and Geikie (1897) first recognised it as an early Carboniferous vent. The geology of the Heads of Ayr was briefly described by Burns (1888), who noted that the rocks at Greenan Castle are of similar composition. Detailed observations were made by Tyrrell (1920), Eyles et al. (1929) and Whyte (1963; 1992). It was described as part of the recent resurvey (Sowerbutts, 1999) and as a Geological Conservation review site (Williamson and Monaghan, 2003).

The Heads of Ayr Vent (Figure 16) cuts the Ballagan Formation, which becomes disrupted, faulted and folded adjacent to the volcanic rocks. Geometries previously observed at the western end of Heads of Ayr, on the wave cut platform (e.g. Eyles et al., 1929; Whyte, 1963) and new exposures caused by landsliding at the eastern end, substantiate the cross-cutting nature of the volcanic rocks. The contact is mainly near-vertical, cutting up through Ballagan Formation country rocks, although in detail it includes small sections sub-parallel to bedding for example at [NS 2931 1871]. The country rocks are baked, reddened, bleached and disturbed at the contact, and appear to have been fairly well lithified before the onset of Heads of Ayr volcanism. Recent palynology results from the Ballagan Formation below and the Lawmuir Formation above the Greenan Castle Pyroclastic Member [NS 3120 1933] constrain the age of this pyroclastic deposit to a time during the latest Tournasian to early Asbian (Stephenson et al., 2002). The pyroclastic member is commonly believed to be coeval with the Heads of Ayr volcanism (e.g. Whyte, 1963) and thus by inference, the Heads of Ayr Vent is also constrained to a latest Tournasian to early Asbian age. In areas to the north, the Clyde Plateau Volcanic Formation was being erupted during this timespan (Monro, 1999; MacPherson et al., 2000).

Six lithofacies (see below) were observed within the Heads of Ayr volcanic rocks (Sowerbutts, 1999). Of these, Lithofacies 4 (Intrusive monchiquite) forms a mappable unit and could be classified separately. The other lithofacies are complexly interbedded with each other, have gradational boundaries and are therefore not distinct, mappable units.

Lithofacies 1

Coarser lithic lapilli-tuffs. The most common lithofacies are crudely and massively bedded, poorly sorted, lithic lapilli-tuffs (Plate 20). Clasts are rounded to subangular, commonly 20 to 400 mm diameter, and composed of vesicular volcanic rocks, baked Ballagan Formation rocks, altered ultramafic rocks and carbonaceous wood pieces up to 0.6 m in length (Whyte, 1963). Volcanic clasts include juvenile or cognate vesicular basalts and accidental fragments of Devonian rocks. The lapilli-tuff matrix is commonly fine tuff and may constitute less than 5 per cent to about 40 per cent of the modal content (Phillips, 2000); in matrix-rich samples e.g. (S25749) lapilli are matrix-supported. Thin sections reveal the irregular nature of the clasts, the poorly sorted tuff matrix and calcite alteration (Phillips, 1999). No evidence was found of welding or rheomorphism in the lapilli-tuffs, suggesting the clasts were originally loose and had cooled because of a slow accumulation rate of ejecta from the volcano. Some of the carbonated serpentinite clasts contain the assemblage phlogopite-diopside-spinel. Several highly irregular volcanic lithic clasts have altered dark haloes (Phillips, 1999); these may be cooled margins. In some cases, the amgdaloidal lithic clasts are stretched due to deformation during extrusion (Phillips, 1999). Both these factors attest to the dynamic nature of extrusive volcanism at Heads of Ayr. Bedding is difficult to distinguish on the eastern side of the Heads of Ayr vent.

On the western side of Heads of Ayr, metre-scale massive beds are steeply dipping and folded (Whyte, 1963). Some larger pieces (one up to 20 m wide) of disrupted Ballagan Formation rocks are enclosed within the lithic lapilli-tuffs (Whyte, 1963). These accidental fragments along with the contact relationships imply that the Ballagan Formation was reasonably well lithified during the Heads of Ayr volcanism, and illustrate that the volcanic vent 'drilled' though the country rock. The presence of fossilised wood fragments, up to 0.6 m long (Whyte, 1963), suggests that the contemporary land surface was forested and that the volcanic event was coeval with other Clyde Plateau volcanism. Material was subaerially erupted with periods of tropical weathering between flows (Monro, 1999), rather than being derived from the Ballagan Formation or Clyde Sandstone Formation where there is no evidence that such vegetation existed in the district.

The Heads of Ayr lithic lapilli-tuffs could represent primary explosive block and ash deposits and/or epiclastic debris flows around a vent structure.

Lithofacies 2

Finer lapilli-tuffs. Common on the western side of Heads of Ayr are bedded, poorly sorted lapilli-tuffs, with volcanic clasts mainly less than 20 mm diameter. Normal and reverse grading with indistinct contacts is common, as are 'strings' of volcanic clasts. Some faint, low-angle erosional trough- and cross-bedded structures are observed. Rare larger blocks, some with underlying 'sag' structures are found. Some Heads of Ayr thin sections from this lithofacies are described as volcaniclastic sandstones by Phillips (1999). These are typically poorly sorted, heterolithic, lithic, matrix-rich volcaniclastic rocks with sedimentary rock fragments and quartz grains that are rounded to angular (Phillips, 1999). Phillips (1999) asserted that the slight rounding of the sedimentary rock fragments and quartz grains indicated limited sedimentary reworking. An alternative interpretation is that the clasts were derived from sedimentary rocks that the vent drilled through and dispersed.

The structures and petrography of these rocks suggest that they may represent surging pyroclastic flows (e.g. compare with examples in Fischer and Schminke, 1984) with some volcanic bombs. Alternatively they could be epiclastic, shallow subaqueous deposits. This lithofacies is probably what Whyte (1963) described as having 'false bedding due to flow/gas streaming'.

Lithofacies 3

Laminated tuffs. These thinly bedded deposits are present at the western end of Heads of Ayr. They are intimately interbedded with Lithofacies 2, in places with weakly developed troughs and erosional features. The rocks may have been deposited from surging pyroclastic flow, by ash fall or from epiclastic subaqueous deposition.

Lithofacies 4

Intrusive monchiquite. A group of eleven intrusive monchiquite bosses were mapped by Whyte (1964) in the central part of the Heads of Ayr. The largest of the bosses is about 35 m in length but others are much smaller. They consist of hard, grey-black basaltic rocks with microphenocysts, up to 2 mm in length, of greenish olivine and black augite (Tyrrell, 1920). The bosses in the east or north-eastern part of the group contain fresh olivine, whereas in the west or south-west the olivine phenocrysts are serpentinised. The serpentinisation was considered a late-magmatic process (Whyte, 1963). The olivine is magnesium-rich as is the whole rock, and a modal content of about 20 per cent olivine, 45 per cent augite, 20 per cent analcime/nepheline was given by Whyte (1963) with about 6 per cent labradorite, 6 per cent iron oxides and a few flakes of biotite. The monchiquite bodies were considered by Whyte (1963) to have been intruded after the volcanic activity forming the vent, and are possibly late Carboniferous to Permian in age, which is similar to other small monchiquite bodies in the district (see below). However, the breccias are so intimately mixed with the monchiquite that they appear penecontemporaneous.

Lithofacies 5

Intrusive breccias. Intrusive breccias (Whyte, 1963) form a zone which surrounds the monchiquite intrusions in the central part of the Heads of Ayr Vent. These are hard rocks, similar in composition to the lithic lapilli-tuffs, but with a greater percentage of sedimentary rock fragments (Whyte, 1963). The breccias have a monchiquite matrix close to the intrusions and gradually become more fragmentary farther away (Whyte, 1963). The gradation implies that the breccia and intrusions were contemporaneously produced as the infillings of small explosive pipes (Whyte, 1963).

Lithofacies 6

Brecciated lithic tuffs. Whyte (1963) identified 'intrusive lithic tuffs' in the central and western Heads of Ayr. These rocks cut across surrounding lithofacies and are composed almost entirely of country rock fragments with some volcanic cementing material. The intrusive breccias and tuffs are thought to have formed by brecciation of the underlying rocks due to minor explosions accompanying or preceding intrusion (Whyte, 1963).

Fisher and Schminke (1984) described some typical features of maars and tuff rings. These range in size from 0.2 to 2 km, and are characterised by slightly vesicular clasts; a common small grain size but with large lithic clasts including ultramafic rocks; poor sorting; penecontemporaneous deformation in associated sediments; high-energy transport due to water vapour, and horizontal transport due to surges and bomb sags. The lithofacies and geometry of pyroclastic lithofacies (1–3) in the Heads of Ayr Vent could therefore imply a small phreatomagmatic maar/tuff ring structure, with probable explosive flow and surge deposits plus epiclastic reworking. The scenario is similar to that of Whyte (1963) who envisaged a volcanic cone deposited both under water and subaerially, with small, fault-bounded basins or warps on the slopes of the cone. Later deformation caused the structures seen today (Whyte 1963). Presumably the small basins on the volcanic cone have formed above collapsed volcanic pipes, in the same way as the small circular basins (about 18 m in diameter) which occur in the Ballagan Formation country rocks up to 100 m away from the Heads of Ayr Vent.

Originally, the Heads of Ayr Vent supplied material at least as far as the Greenan Castle volcanic rocks (Whyte, 1963), and Lithofacies 1–3 are present at Greenan Castle. The intrusive lithofacies 4–6 are associated with a volcanic phase late in the Carboniferous history of the maar/tuff ring or with a subsequent (probably Permian) phase according to Whyte (1963). Similar late Carboniferous to Permian intrusive phases are inferred for many of the volcanic necks in east Fife (Forsyth and Chisholm, 1977; Williamson, 2003a). The age of initial development of the necks in east Fife is not known but may be contemporaneous with earlier volcanic beds in the local Namurian to Westphalian succession (Williamson, 2003a). However, the age span of the activity in the necks is not certain.

Late Carboniferous and early Permian intrusions

Carboniferous and Permian magmatism in Scotland has recently been reviewed by Read et al., (2002), Stephenson et al., (2003), Upton et al., (2004) and Macdonald and Fettes, (2007). The late Carboniferous to early Permian igneous rocks of the Midland Valley are typical of extension-related intraplate continental volcanism, but also have similarities to ocean-island basalts (Stephenson et al., 2003). Recent studies of ⁴⁰Ar/³⁹Ar geochronology of Carboniferous–Permian volcanism in the Midland Valley (Monaghan and Pringle, 2004) are also relevant to the timing of the intrusions.

A suite of alkali dolerite sills and dykes is exposed in a belt from Ayr south-eastwards towards Patna. The sills are clearly younger than the Coal Measures, and the monchiquites are related to the early Permian magmatism, but some olivine-dolerites and basalts could be late Carboniferous to early Permian or Palaeogene in age.

Analcime-dolerites (AD^A)

Several analcime-dolerite sills have been described as teschenitic (i.e. they include analcime). They are principally composed of plagioclase, clinopyroxene and olivine, however, many of the constituent minerals are altered. The largest of these sills lies below Ayr [NS 340 230] and extends north-westwards off-shore.

Two small, dark grey analcime-dolerite sills, probably leaves of the same body, intrude the Ballagan Formation south of Abbothill. The larger sill [NS 3688 1903] has a subophitic texture and is locally spheroidally weathered.

Olivine-dolerite (oD)

Parts of two olivine-dolerite sills crop out at the eastern margin of the district [NS 405 164] and [NS 405 157], but most of their crop lies on the adjacent sheet to the east. These dark grey mafic rocks contain calcic plagioclase, pyroxene, olivine and iron oxides. They are olivine-rich and characteristically contain minor nepheline. They were classified as theralites on the previous 1:50 000 Solid edition of Sheet 14W (Institute of Geological Sciences, 1978a) or 'kylites' in the previous memoir (Eyles et al., 1949).

Olivine basalt dykes (oB)

Olivine basalt dykes and minor intrusions of Carboniferous to early Permian age are apparently scarce in the Ayr district, although some could be included in those dykes of uncertain age (D). Small alkali basalt dykes, probably related to the monchiquite intrusions (Whyte, 1963), trend west-north-west and cut the Heads of Ayr Vent. They are considered to be late Carboniferous to early Permian in age.

Monchiquite (L^MO) and Lamprophyre (L)

A small dyke trending north to south, and classed as lamprophyre is exposed in a burn [NS 401 802] east of Loch Spallander.

A monchiquitic dyke at [NS 2995 1910], near Long Rue on the coast, trends from west to west-north-west and is described as altered clinopyroxene-olivine-plagioclase microphyric monchiquitic basalt (Phillips, 1999). It is believed to be part of the same swarm as the dyke exposed at Garryhorn Bridge [NS 2966 1384]. The latter microphyric basalt dyke is about 1 m thick. The rock is highly altered and replaced by carbonate minerals, but appears to have been olivine-rich and monchiquitic.

A monchiquite dyke, trending 115°, which crosses the River Doon at [NS 3782 1358], north-east of Skeldon House, has a fine-grained, pale grey altered matrix containing darker phenocrysts. A sample of the dyke (S24750) taken from the eastern side of the river is described as a microporphyritic olivine-clinopyroxene-bearing monchiquite. This sample contains a xenolith composed of polycrystalline green pyroxene (Phillips, 1999).

Similar monchiquitic lamprophyres are associated with early Permian vents and plugs (see below).

Early Permian vents and plugs (Z_V)

Numerous volcanic necks, and some plugs, have been related to the early Permian volcanic phase which fed the Mauchline Volcanic Formation but few lie within the Ayr district. East of Kirkmichael, vents form the hills of Carclout [NS 395 099] and Kirklafinn [NS 402 093], belonging to a group of vents centred farther east around Patna [NS 415 105]. Fragments of monchiquite or analcime basanite found within the vents resemble the lavas of the Mauchline Volcanic Formation. This is part of the alkali magmatism related to post-Variscan extension, which occurred in the central and western Midland Valley during the latest Carboniferous or early Permian, from 298.3 ± 1.3 to 292.1 ± 1.1 Ma (Monaghan and Pringle, 2004). This correlates well with similar activity observed across the north-west European foreland from 300 to 280 Ma.

The vent rocks were confirmed as early Permian by De Souza (1979) producing K-Ar determinations of 280–285 Ma on similar nearby vents. More recently, Monaghan and Pringle (2004) used ⁴⁰Ar/³⁹Ar geochronology to refine the relationships of the Carboniferous–Permian magmatism in the Midland Valley. A hornblende separate from an analcime-dolerite intrusion at Carskeoch (which lies just to the east of the district at [NS 411 093]) gave a concordant age of 295.2 ± 1.3 Ma, which is older than the K-Ar results and corresponds to earliest Permian (Gradstein et al., 2004). This intrusion is probably related to the early Permian vents and associated alkali olivine-microgabbro on Carclout Hill (see below).

A vent or neck about 250 m in diameter, occurring east of Prestwick [NS 368 261], was encountered during mining operations. Surface and mining evidence in the Mauchline area (Eyles et al., 1949) confirm that none of those vents are early Carboniferous in age. Furthermore, their petrology is distinct from the younger Paleocene igneous rocks (Eyles et al., 1929).

A massive, grey-brown mottled, intrusive volcaniclastic breccia crops out on Carclout Hill [NS 394 099]. The feature is approximately 350 m across from east to west. The breccia contains subrounded to angular volcanic and sedimentary rock clasts up to 0.2 m in length, set in a poorly sorted, coarse to granule-grade sandstone matrix. The volcanic clasts range from vesicular basalt to pilotaxitic porphyritic basalt, together with microgabbroic and gabbroic fragments.

The fresh, dark grey, olivine-microgabbro or nepheline monzogabbro 'essexite' body exposed in the former small quarry [NS 3932 0997] on Carclout Hill, is considered to be either a vent intrusion or possibly a large vent block. It was described as a small boss of the Crawfordjohn type of 'essexite' (Tyrrell, 1912 p.121). This distinctive titanaugite olivine-microgabbro contains orthoclase, minor nepheline, analcime, ilmenite, biotite, a little hornblende and apatite (Scott, 1915 p.458). Subsequently, Tyrrell (1928a, p.285) referred to the intrusion at Carclout as 'luscladite' because it was a type of 'olivine-theralite' containing little or no hornblende. This rock was analysed by Eyles et al. (1949 p.137–138) and at that time was classified as a 'theralitic essexite'. The large ophitic titanaugite crystals were noted as commonly edged with aegerine-augite, and the matrix contained labradorite laths (Guppy and Thomas, 1931, p.91). Sample (S25185) is best described as a medium- to coarse-grained augite olivine-microgabbro. Wallis (1989) analysed a hawaiite block (SW76) from Carclout Hill containing olivine and clinopyroxene phenocrysts. This rock was related to the Mauchline magmatism (Wallis, 1989).

A nearby vent breccia forms Kirklafinn and a disused pit at [NS 400 095] exposed massive, grey volcanic breccia that is part of this vent. The fine-grained matrix includes rounded quartz grains and blocks of olivine basalt and carbonated peridotite. Two small satellite vents of tuffaceous material lie to the north-east.

Another breccia is exposed about 100 m north-west of Boreland [NS 398 139], on the north bank of the River Doon. It forms a prominent feature about 200 m in length with a west-north-west alignment. The breccia mainly comprises mid-grey vesicular basalt and monchiquite, with some laminated calcareous siltstone and silty mudstone fragments, generally 10–20 mm but exceptionally 400 mm in length. The latter are slightly baked and sparry calcite infills the vugs. The sedimentary rock fragments are considered to be derived from the Ballagan Formation. The breccia is therefore thought to intrude the Ballagan Formation (compare Institute of Geological Sciences, 1978a). Some grey, coarsely crystalline mafic rock fragments have been described as 'carbonated peridotite' (Eyles et al., 1949), and may have come from deeper levels within the crust. The breccia is considered to be one of the early Permian volcanic necks grouped in the vicinity of Patna (Eyles et al., 1949), but its intrusion into the Ballagan Formation and its content of 'carbonated peridotite' is similar to the vent at the Heads of Ayr.

Basalt dykes in early Permian vents (oD^A_V)

South-east of Dalrymple, an alkali olivine-bearing dolerite dyke intrudes the early Permian volcanic vent breccia on Carclout Hill at [NS 3932 1000], and is most likely early Permian (see above). Another small olivine-bearing basalt dyke at [NS 3985 1385] intrudes the vent breccia north of the River Doon.

Monchiquites (L^MOV)

Several small lamprophyric intrusions are associated with the early Permian vents and plugs (see above) and considered to be of similar age.

A small plug of altered monchiquitic basalt is associated with vent breccia at Wallace's Seat [NS 3963 2288], and it is exposed on the right bank of the River Ayr. The Colvinston monchiquitic vent intrusion [NS 395 228] belongs to the early Permian Mauchline 'group' of lavas and plugs according to Wallis (1989), who analysed a sample (SW372) containing olivine phenocrysts. The vent intrusion is associated with two smaller, similar plugs (Eyles et al., 1949, p.106). These plugs contain irregular, variably orientated blocks of sandstone and mudstone up to 1 m long, considered to be similar to Coal Measures (Eyles et al., 1949). The highly xenolithic monchiquite vent intrusions are difficult to distinguish from compact vent breccias. Both types are considered to be early Permian in age as some cut the early Permian Mauchline lavas but not the overlying Mauchline Sandstone Formation.

Another monchiquitic basalt is shown intruding the Heads of Ayr Vent at [NS 287 188] and it is described together with that intrusion (p.53), where it is considered to be of Carboniferous age.

Palaeogene intrusions

The Ayr district lies near the south-eastern edge of the area affected by the Palaeogene magmatism, caused by extensive rifting prior to the opening of the Atlantic Ocean. A major hot spot or mantle plume produced the North Atlantic Igneous Superprovince (MacDonald and Fettes, 2007). The part of the superprovince within the British Isles is known as the Hebridean Igneous Province, formerly the 'British Tertiary Igneous Province' (Emeleus and Bell, 2005). The province was active between about 60.5 Ma and 55 Ma (Bell and Williamson, 2002), that is almost entirely during the Paleocene Epoch. Earlier tectonic events, such as the Permo-Triassic extension of the Mauchline Basin, probably focussed the upward migration of magma during intrusion of the Paleocene sill-complex in the Ayr district.

Two phases of mafic intrusion are evident and appear to have developed separately; the first was associated with the Prestwick–Mauchline Sill complex, and the second with dyke swarms centred on Mull and/or Arran (Cameron and Stephenson, 1985). Prior to Mykura's (1967) study, the age of the Prestwick–Mauchline Sill complex was debatable, either Permian or 'Tertiary' (Tyrrell, 1928; Eyles et al., 1949 and references therein) but the Paleocene age has been confirmed by palaeomagnetic measurements (Armstrong, 1957) and by dating (De Souza, 1979).

The dykes have both alkali and tholeiitic affinities and generally trend north-west to north-north-west. Several dykes are prominently exposed along the shore south of Ayr, and were described during the original geological survey (Geikie et al., 1869). The contact metamorphic effects of the 'Deil's Dyke' on a 'shale' xenolith was recorded by J Smith (1898a).

Although locally the intrusion of individual dykes is controlled largely by the structure of the country rocks, on the regional scale the trend of the dyke swarms can be attributed to the stress field at the time of intrusion (Vann, 1978). Most dykes in the Ayr district have alkaline compositions; tholeiitic dykes are less common and include a prominent tholeiitic andesite, the Cumbrae–Stevenston Dyke. The alkali dolerite dykes are considered to be earlier than the tholeiitic dykes. The former are thought to relate to the extrusion of the extensive lava fields in the Inner Hebrides and Antrim, and predate the central complexes and the associated intrusion of the tholeiitic dykes.

Arran is the closest Palaeogene centre to the Ayr district, which lies on the main dilational axis trending south-south-east from the Central Arran Ring-complex (Speight et al., 1982). Evidence from Arran indicates that the alkali dolerite dykes and sills are earlier than the North Arran Granite (dated at around 58.5 Ma ⁴⁰Ar/³⁹Ar, Chambers and Pringle, 2001) and the somewhat younger Central Arran Ring complex, all of which are cut by tholeiitic dykes.

Sills

In the Prestwick area, extensive outcrops of 'crinanitic' dolerite form the bulk of the Prestwick–Mauchline Sill-complex (aDP). This term was introduced by Mykura (1967) for the group of olivine-analcime sills between Mauchline [NS 516 253] and the coast near Bentfield [NS 340 249] (Cameron and Stephenson, 1985). The sill-complex, which has a maximum thickness of 60 m and may consist of between one and three leaves, was intruded into well consolidated strata. The leaves, up to 500 m apart, are connected by dykes or steeply inclined sheets. The transgressions of the sill elevated the body from the top of the Middle Coal Measures in the west, through the Upper Coal Measures and into the Mauchline Volcanic Formation on the Mauchline Sheet to the east. Transgressions of the sill combined with faulting and folding account for the complex outcrop pattern around Prestwick Airport [NS 365 268]. Its probable original extent was depicted by Mykura (1967, fig. 14) using outcrop and borehole evidence. On Sheet 14W the sill is intruded into the higher beds of the Middle Coal Measures; the large transgression up into the Mauchline Volcanic Formation occurs below Failford [NS 459 263] on Sheet 14E (Mykura, 1967).

Offshore in the Firth of Clyde, the sill-complex is exposed 5 km west-north-west of Meikle Craigs at Lady Isle [NS 275 293]. Geophysical surveys suggest that it extends discontinuously for at least 8 km farther west below the sea bed. These igneous bodies have been interpreted as laccoliths (McLean and Deegan, 1978). However, their discontinuous nature may simply be due to erosion of the sea floor, faulting and tilting of strata, transgression of the sill or exposure of different leaves of the sill. In any case it probably is related to the analcime-bearing olivine-dolerite sill-complex on Holy Island, around Lamlash Bay and Dippin on Arran (see below).

The age of the sill is considered to be Paleocene, since correlative rocks cut the early Permian Mauchline Volcanic Formation, and a K-Ar date of 58 ± 1.4 Ma was obtained from the sill near Howford Bridge [NS 516 254] in the Mauchline area (De Souza, 1979; Wallis, 1989). The petrographical similarity of the sill with those on Arran, and palaeomagnetic measurements (Armstrong, 1957), support the Paleocene dating. The intrusion of the sill-complex is an early feature of Paleocene history in the district, as it is followed by faulting and intrusion of tholeiitic dykes.

Numerous exposures of the sill-complex occur in the River Ayr, for example at Auchincruive House [NS 3909 2352], where it is termed the Auchincruive Sill (Mykura, 1967), and along the coast near Bentfield [NS 341 249]. Features of the sill include chilled margins, large xenoliths of sedimentary rock and brecciated margins. Late alkaline analcime-rich veins and segregations (analcime-syenite), up to about 0.5 m thick, have been recorded from the Prestwick–Auchincruive part of the complex (among others, such as Howford Bridge, occurring farther east) and their origin debated (Tyrrell, 1928b; Eyles et al., 1949). The analcime-rich veins (Sowerbutts and Smith, 1999) and olivine-rich veins with ophitic textures (Eyles et al., 1949) have been observed in the Auchincruive Sill at Auchincruive House. The pegmatitic facies are the product of late-stage crystallisation.

Analcime-bearing dolerites considered to belong to the same sill complex crop out at different stratigraphical levels, for example at Meikle Craigs [NS 328 288] and Ayr town centre [NS 3368 2225]. A thin section (S22509) from near the sill margin at the south-west corner of Meikle Craigs was a medium-grained (1–2 mm average grain size), aphyric teschenite. 'Crinanite' and teschenite are both varieties of analcime-olivine-dolerite, but teschenite contains more analcime and less olivine than 'crinanite'. The teschenite sample contained Ti-rich clinopyroxene, plagioclase and kaersutite (clinoamphibole), in an ophitic to subophitic texture with interstitial analcime. Accessory phases include biotite, opaque minerals and apatite. Brown to red-brown kaersutite is locally altered and hydrated to chlorite and blue-green amphibole; the rest of the rock contains some secondary sericitic white mica and very fine-grained chlorite (Phillips, 1999).

The comparable Palaeogene Dippin Sill on Arran [NS 050 222] is about 40 m thick. It also has a thick central 'crinanitic' unit with teschenite margins and patches of analcime-dolerite and basic pegmatitic rock (Gibb and Henderson, 1978). The latter authors postulated a mantle-derived basic magma held in a deep crustal reservoir, allowing fractionation before the magma was intruded at a shallow depth, initially forming the marginal facies and later the central bulk of the sill. As in the Ayr district, the analcime-bearing dolerites on Arran are earlier than quartz-bearing/tholeiitic dolerites.

The small intrusions of altered dolerite which occur within the Ayr Ell Coal, just south of Oswald Bridge [NS 3875 2304], have been described by Mykura (1965) as 'white trap', which is a term formerly used for intrusive igneous rock bleached through its contact with coal seams. The exposures, too small to show on the 1:50 000 map, are near the east bank of the River Ayr and comprise horizontally elongated ovoid masses of 'white trap', about 2 m long by 0.6 m high, within an earthy coked coal 0.79 m thick. The intrusions are likely to be Palaeogene in age as there are several nearby Palaeogene north-west-trending dykes, and the Prestwick–Mauchline Sill-complex crops out to the north. Similar minor intrusions of white trap have been recorded in boreholes and mines within south-west Ayrshire (Mykura, 1965).

Dykes

Palaeogene dykes are commonly 1–10 m wide and only the more persistent are over 10 m wide. They are near-vertical intrusions that in some cases are impersistent and are intruded in an en echelon pattern. Most of the dykes are classed as dolerites (sensu stricto, microgabbros, or gabbros where their grain size is greater than 2 mm) and where they are predominantly fine-grained (less than 0.25 mm grain-size) their basaltic nature has been indicated on the 1:10 000 scale maps. The rocks are characterisitically fresh and hard, with columnar jointing commonly developed perpendicular or parallel to fine-grained chilled margins. The dykes generally have a north-westerly trend and this is confirmed by offshore geophysical surveys. Dykes encountered underground in the Ayr Coalfield (presumed to be Palaeogene) confirmed this trend, but also proved trends in a west-north-west and westerly trend as well as a north-easterly set.

Different categories of Palaeogene dyke have been identified in the district and their petrology has been described (Phillips, 1999). These are alkali olivine-dolerite or basalt (oD^A_P), olivine-dolerite or basalt (oD_P), tholeiitic dolerite or basalt (D^T_P), tholeiitic andesite (A^T_P), quartz-dolerite (qD_P) and unclassed dolerite or basalt (D_P).

The alkaline dykes are olivine-dolerites or olivine basalts which typically contain ophitic, purplish titanaugite enclosing plagioclase laths, olivine and iron-titanium oxides. In addition, most of them contain small amounts of analcime or zeolites, so could be classed as 'crinanitic'. These are relatively common in the district and the olivine basalt and dolerite dykes are related.

Tholeiitic dykes are composed of labradorite laths and augite, with varying amounts of glassy groundmass, usually devitrified and darkened by finely disseminated iron oxides. They can include small amounts of free quartz (Eyles et al., 1949). Quartz-dolerite dykes are less common. Tholeiitic dykes are less common than alkaline dykes in the district, but are widespread and some are persistent. Evidence has been produced that they are younger than the alkaline dykes (Eyles et al., 1949). Many of the dykes considered to be Palaeogene in small, inland exposures are unclassed.

On the shoreline between Longhill Point and Bracken Bay, several of the Palaeogene dykes of quartz-dolerite, tholeiitic basalt, olivine-dolerite and basalt are well exposed trending north-west to north-north-west, for example Bracken Bay olivine-dolerite [NS 2800 1833]. The Deil's Dyke [NS 3065 1940] forms a prominent feature up to 15 m thick (Plate 21) and is of typical crinanitic composition, containing some analcime. On the south-east side, a baked raft of Ballagan Formation is incorporated, which includes globular spots (Smith, 1898a).

Likewise in the Culzean–Dipple area, alkaline, tholeiitic and unclassed dykes, observed in coastal or stream exposures, have a west-north-west to north-north-west trend. Chilled margins of the dykes are occasionally exposed, for example north of Dipple Factory [NS 199 035]. Direct evidence of the age of the dykes is limited; those with a north-westerly trend are probably related to either the Arran or Mull dyke swarms (e.g. Eyles et al., 1949) and therefore considered Palaeogene in age.

Alkaline dykes

The alkaline dykes may have been feeders to lava fields. Alkali olivine basalts of the Mull Plateau Magma-type are represented by the 'crinanites' and allied rocks which are particularly abundant in the South Arran Dyke-swarm.

Alkaline dykes occur throughout the Ayr district, but are particularly common south of Ayr. Most trend between west and west-north-west, but others trend north-west and a few north-north-east; some of these may be Permian in age so are unclassed (D on the geological map). Most of these dykes are relatively fine-grained and poor in analcime and zeolites, similar to the 'crinanitic' dykes on Arran. Rare coarser-grained ophitic types resemble the Prestwick–Mauchline Sill complex. A good example is the 10 m-thick dyke trending west-north-west on the shore at [NS 267 178] near Fisherton Cottage; it is rich in analcime and is cut by alkaline felsic veins.

Another alkali olivine-dolerite, over 7 m thick, trending north-westwards across the Water of Girvan at [NS 343 077], is considered to be Palaeogene in age. It intruded the Ballagan Formation and was classed as a 'crinanite' on the solid edition of the Ayr Sheet (Institute of Geological Sciences, 1978a). More recently Phillips (1999) described it as an olivine-dolerite. An alkali olivine basaltic dyke with small calcite amygdales, 1.5 m wide and trending 323°, crops out in Cloyntie Burn [NS 3330 0552]. In thin section (N4287), altered olivine and clinopyroxene phenocrysts can be seen in a plagioclase-rich matrix, which also contains small polycrystalline quartz xenoliths. The dyke is relatively altered but its orientation is typical of the Palaeogene dyke swarms.

In the Straiton–Spallander area, a north-west-trending dyke exposed in the south bank of the Water of Girvan north of Arnsow [NS 3516 0807] is an olivine-bearing alkali microgabbro thought to be of Palaeogene age. It is altered, amygdaloidal and was formerly classified as 'crinanitic' (Institute of Geological Sciences, 1978a; Eyles et al., 1949). A similar dyke of olivine-bearing basalt to microgabbro north-east of Troquhain [NS 3796 0960] has a north-north-west trend, and is also probably Palaeogene.

An altered olivine-analcime dolerite dyke, 5 m wide, is exposed intruding Ballagan Formation rocks near Skeldon House [NS 3779 1351]. The dyke was previously described as a camptonite, i.e. containing amphibole and analcime (Institute of Geological Sciences, 1978a), but this is difficult to confirm from thin section (Phillips, 1999). A dyke at [NS 3759 1341] to the south-west of Skeldon House is an alkali olivine-dolerite.

Tholeiitic dykes

Most of the tholeiitic dykes in central Ayrshire are narrow, up to 3 m wide, and cannot be traced far individually (Eyles et al., 1949). They usually have a north-west to north-north-westerly trend, and are well represented in the area between the Heads of Ayr and Kirkmichael. However, there are a few broader tholeiitic dykes which can be traced for considerable distances. An exceptionally persistent tholeiitic dyke near Bentfield [NS 339 247] was confirmed underground by mining records, and extends for about 4 km offshore as indicated by geophysical survey.

The Bracken Bay–Cairnhill–Straiton Dyke (Eyles et al., 1949, p.122) is another prominent dyke between 6 and 15 m wide. This quartz-bearing, tholeiitic microgabbro dyke trends north-westwards from Lambdoughty Glen [NS 395 057], where it is about 6 m thick. It has a distinctive abundant cryptocrystalline mesostasis. The dyke appears to continue for over 1 km north-westwards across Muirsmill Burn [NS 375 085], and thickens up to 15 m in a former quarry south of Cairnhill [NS 369 091]. Eyles et al., (1949, fig. 18) traced this dyke south-eastwards from Bracken Bay for over 20 km. The dyke cuts across north-north-easterly and east–west trending faults, but may be offset or deflected by the larger Kerse Loch Fault. It continues south-eastwards on to Sheet 14E (Mauchline) and turns to trend north-east along the Southern Upland Fault, 2 km east of Dalleagles [NS 593 106]. This deflection is discussed by Eyles et al. (1949). The dyke generally parallels, and is probably related to, the Cleveland–Armathwaite Dyke.

Quartz-dolerite dykes are relatively sparse and small in this district. A typical example is exposed near Otterden House [NS 308 144].

The regionally persistent Cumbrae–Stevenston Dyke is a tholeiitic andesite, up to 20 m wide, which lies close to the south-western edge of the Mull swarm. It extends north-west to Barassie [NS 322 333] on the Irvine Sheet (Monro, 1999) and south-east to Drongan [NS 446 187] on Sheet 14E (MacGregor in Eyles et al., 1949). It may extend farther south-east as the Cleveland–Armathwaite Dyke (Eyles et al., 1949). It is the most petrographically distinctive dyke of the Inner Hebridean Palaeogene swarm and the most siliceous. It was formerly termed 'cumbraite' (Tyrrell, 1917) but is essentially a tholeiitic andesite, characterised by abundant anorthitic plagioclase phenocrysts set in a groundmass of labradorite laths, enstatite and augite. Some patches of glass and quartz are present. It cuts the Prestwick–Mauchline Sill complex near St Quivox [NS 388 242]. The trace of the dyke towards and along the Southern Uplands Fault, and the possibility of it splitting, is discussed by MacGregor (in Eyles et al., 1949, p.121–123). Subsequent study (Macdonald et al., 1988) has shown that the Stevenston–Coylton (–Drongan) segment of the dyke is not comagmatic with the Cleveland Dyke (sensu stricto). The Stevenston–Coylton segment contains 6–10 per cent plagioclase phenocrysts and pyroxene phenocrysts are rare compared to the Cleveland Dyke (sensu stricto), which averages 3–4 per cent plagioclase. The Stevenston–Coylton segment shows internal differentiation, the central facies being more evolved than the margins. Both this segment and the Cleveland Dyke could have been emplaced as single, separate pulses of magma from the Mull centre. It has been established that for the Mull Dyke swarm, with increasing distance from the central complex, the number of individual dykes decreases but individual dykes are thicker and more widely spaced (Jolly and Sanderson, 1995). This is probably the reason why only one persistent dyke of this type is found in the district.

A summary of Palaeogene events is as follows:

Emplacement of the Prestwick–Mauchline Sill- complex.
Formation of step-faults marginal to the Mauchline Basin.
Continued movement along major north-east-trending faults (probably continued throughout and overlapped with dyke intrusion).
Emplacement of alkaline dykes followed by tholeiitic dykes.

Chapter 9 Structure

The Ayr district lies on the south-western edge of the Midland Valley terrane in Scotland. The Midland Valley is considered a distinct terrane (Bluck, 2002) lying between the Grampian Highland and the Southern Uplands terranes, but it may be partly overthrust by the latter. The Midland Valley terrane has clearly been subject to three main orogenic phases, firstly in the Mid Ordovician to mid Silurian (Caledonian), then in the Mid Devonian ('Acadian') followed in the late Carboniferous by the Variscan event. The limited exposure in the district does not allow the structure of the area to be fully elucidated, but structural observations during resurvey inform the regional understanding of the long period of structural evolution in the terrane.

The evidence for earlier, pre-Llanvirn tectonic events (i.e. the obduction of the Ballantrae Ophiolite complex and accretion of the hidden arc-basement of the Midland Valley to the edge of Laurentia) can only be deduced from elsewhere in the Midland Valley, particularly south of Girvan. The latter obduction and accretion caused the Grampian Orogeny in the Scottish Highlands.

The Caledonian orogenic events that occurred during the closure of the Iapetus Ocean are evident in the deformation and uplift of the Caradocian to Wenlock age rocks of the district, and also the equivalent age rocks in the Southern Uplands.

Additionally, since the Midland Valley terrane has a basement considered to be an amalgam of various strike-slip wedges at the margin of the Laurentian continent, it has been subject to subsequent minor transpressional and transtensional tectonic phases, many reactivating the major Caledonoid features. The unconformity observed between the Llandovery–Wenlock rocks of the Craighead Inlier and the Siluro-Devonian Swanshaw Sandstone Formation has been attributed to sinistral transpression (Smith, 1995) at the end of the Caledonian events.

Folding about axes of east-north-east trend, and reactivation of Caledonoid north-east trending structures with throw down to the north-west, affected the Siluro-Devonian rocks of the southern Midland Valley and are associated with Acadian deformation (Smith, 1999a). Siluro-Devonian and Carboniferous volcanic episodes had more local effects on the structure of the district.

Structures affecting Carboniferous and earliest Permian rocks in the Midland Valley with east-trending and north-trending orientations, as well as the reactivation of north-easterly Caledonian structures, notably the Dailly Syncline, have been attributed to polyphase Variscan tectonism (Read 1988; Rippon et al., 1996; Smith, 1999a).

Later periods of regional tension and rifting, occurring during early Permian and Palaeogene times, also tend to be associated with magmatism.

Deformation phases

Pre-Llanvirn deformation

The obduction of the Ballantrae Ophiolite complex (exposed in the Girvan area) during Arenig times, and circumstantial evidence of an earlier Ordovician volcanic arc, form the case for the Grampian Orogeny affecting the Midland Valley terrane. The circumstantial evidence is provided by the presence of arc-derived detritus in the Mid Ordovician inliers around Girvan. The detritus was deposited after the Grampian Orogeny and includes igneous clasts, dated isotopically as Ordovician, derived from a proximal source to the north that is assumed to be a volcanic arc, similar to the Lough Nafooey arc in western Ireland. The distal part of the forearc basin is not exposed; it was probably overthrust by the Southern Uplands (Bluck, 1983).

Around Girvan, the Ballantrae Ophiolite complex contains graptolites which indicate that sediments within it were deposited during the early–mid Arenig. The complex was part of a larger ophiolite obduction event which occurred at about 470 Ma, about the same time as the Highland Border and Tyrone Igneous complexes. Since the latter ophiolites lie on the northern margin of the Midland Valley terrane, it is thought that this terrane was a separate rifted fragment of Laurentia, with a volcanic/plutonic arc developing prior to the lower Arenig. Precambrian–Ordovician magmatism in the Midland Valley is inferred from dated mafic to granitic clasts from the Kilranny and Benan conglomerates (Longman et al., 1979). A cryptic 'flysch' sequence of Ordovician age has also been identified in the north-east of the terrane (Haughton and Halliday, 1991), which was fed north-westwards into the Siluro-Devonian basins near Stonehaven around [NO 880 860], possibly derived from an analogous arc system.

It is not certain whether the Craighead lavas are part of the Ballantrae Ophiolite complex or a later, early Caradoc volcanic suite (Floyd, 1999), but they lie unconformably below the mid Caradoc sedimentary rocks in the Craighead Inlier which were involved in Caledonian deformation (see below).

Caledonian deformation

Because of the extensive cover of Upper Palaeozoic rocks in the Midland Valley terrane, only limited areas of Ordovician and Silurian strata, affected by the Caledonian Orogeny crop out, and these occur in the southern part of this district. However, they are important in unravelling the early history of the Midland Valley terrane, and its role in linking the orthotectonic Grampian Highland and paratectonic Southern Uplands terranes as the Iapetus Ocean closed at a destructive plate margin.

The Lower Palaeozoic rocks exposed in the Craighead Inlier, the Girvan main outcrop and smaller inliers to the south-east, now appear to occupy tracts bound by Caledonoid faults (Figure 17), but they were all part of an active shelf margin on the southern side of the Laurentian continent, facing Iapetus to the south.

The post-Llanvirn sequence rests unconformably on the eroded Ballantrae Ophiolite complex which is considered to be a fragment of oceanic crust obducted during the Grampian Orogeny (see above). During mid Caradoc times sedimentary basins developed, controlled by contemporaneous faulting and accumulating conglomerates that fine to the south-east into thick greywacke sandstone sequences. The sequences are considered to form the cover rocks to the accreted forearc basin and more extensive obducted rocks than presently exposed. Contemporaneous fault controlled topographical highs are inferred from the areas where shallow water clastics and fossiliferous limestones (e.g. Craighead Limestone) accumulated. The Craighead Limestone 'reef' developed over the edge of the block (rift shoulder), against the precursor to the Kerse Loch Fault. Syndepositional slumping occurred in the basin farther north during Ashgill times.

The basin fills became progressively younger to the north-west (e.g. the Benan Conglomerate is older than the Craighead Limestone). A proximal forearc scenario is indicated by the bouldery nature of the conglomerates, which show a cyclicity into finer sediment suggesting regular (intermittent) basement faulting (Bluck and Ingham, 1992). The Ordovician faunas in the area show strong Laurentian affinities, but this provinciality decreases upwards in the succession as the Iapetus Ocean was eliminated. After a period of post-Ashgill deformation (possibly minor uplift due to magmatism, for instance at 443.3 +/−6.4 Ma in biotite-garnet-bearing granite according to Haughton and Halliday, 1991) or erosion due to a fall in sea level during the Hirantian Stage, a Silurian (Llandovery–Wenlock) succession was deposited, again in a rifted marginal setting. Deposition continued into early Wenlock times (up to around 427 Ma) at least in the main Girvan outcrop, so the main Caledonian deformation affecting the Midland Valley terrane was post-Wenlock. The deformation appears to have occurred at fairly shallow depths, with no accompanying penetrative cleavage. Studies of clay minerals and the colour alteration indices of conodonts from near Girvan (value 1–2.5) indicate that the cover sequence suffered weak burial metamorphism at temperatures no greater than 150°C (Oliver et al., 1984). Acritarch colouration from the Straiton Subgroup (Llandovery–Wenlock) indicates that the beds are of low thermal maturation and are of immature status as an oil source (Dorning, 1982). This is in marked contrast to the metamorphism of the Lower Palaeozoic rocks in the Southern Uplands (zeolite to prehnite-pumpellyite facies, temperatures of 250–400°C), which is consistent with the model in which the Southern Uplands is an accretionary prism thrust up onto the Midland Valley terrane late in the Caledonian Orogeny (Bluck, 1983; Stone and Merriman, 2004).

The observed deformation style is different in the different Lower Palaeozoic inliers, so the Craighead and the Girvan main inliers are described separately. The difference is probably due to the development of fault-bounded tracts during Caledonian deformation, which are now represented at rockhead by the Kerse Loch, Straiton/Bargany and Knockinculloch faults (Figure 17). Besides dip-slip, these faults are likely to have suffered strike-slip movements, and these were considered to be the reason why inliers along the southern Midland Valley had differences in their sedimentary fill and relationships to the overlying Siluro-Devonian strata (Smith, 1995). The faults may have a listric form, similar to the tract-bounding faults in the Southern Uplands. In their basin model for the Silurian of the Midland Valley and Ireland, Williams and Harper (1988) suggested an inter-arc environment controlled by oblique-slip faulting. A phase of sinistral transpression in post-Wenlock to pre-Early Devonian times was postulated by Smith (1995).

Caledonian deformation of the Craighead Inlier (north of the Kerse Loch Fault)

The Lower Palaeozoic succession exposed north of the Kerse Loch Fault lies in a lenticular area (in all 12 km by 2 km) termed the Craighead Inlier, forming an anticline trending 060° and plunging gently (about 15°) north-eastwards. The fold is asymmetrical, with beds dipping about 30° north-north-east below Siluro-Devonian rocks on the northern limb and up to 75° south-east on the southern limb where the inlier is faulted against Carboniferous strata. At the end of the Ordovician a period of transpression/uplift occurred, during which part of the Ordovician sequence was eroded down to the Arenig volcanic and associated plutonic rocks. In mid Rhudannian times (around 440 Ma) the Mulloch Hill Conglomerate was deposited, followed by a transgressive sequence. The postulated hiatus (around 434 Ma) below the Lower Camregan Grits Formation may be influenced by tectonic factors, since graptolitic shales below are suddenly succeeded by the shallow water grits. The tectonic shuffling reported in the Pencleuch graptolitic shales (Floyd and Williams, 2003) is likely to be later; related to the growth of the Craighead anticline in the footwall of the Kerse Loch Fault. The pattern of faulting at [NS 250 025] in the Shalloch Formation, which is subparallel to the Kerse Loch Fault, indicates some sinistral movement related to the latter fault. The fan of faulting around the hinge suggests faults related to later relaxation of the anticline.

The south-east limb of the anticline is cut out by the Kerse Loch Fault zone, whereas the Siluro-Devonian Lanark Group succession overlies its north-west limb unconformably. After the main Caledonian fold phase the Lower Palaeozoic rocks were affected by several steeply dipping 'dip' faults whose trends vary from north to south in the west, to north-east to south-west in the eastern part of the outcrop at or near the anticlinal hinge. An earlier set of minor faults was recognised by Harper (1982), which are chiefly restricted to the conglomerates in the Auldthorns and Mulloch Hill Conglomerate formations. This phase of faulting is thought to have occurred during the later stages of folding, with rafts of conglomerate being thrust into the surrounding, less competent mudstones in the hinge region of the anticline (as a result of accommodation in the hinge).

Caledonian deformation in the Straiton area (Girvan main outcrop)

The Llandovery–Wenlock rocks in the Straiton area are exposed in a belt north-west of the Straiton Fault. They appear to lie in steeply stacked to overturned thrust sheets formed during the post-Wenlock Caledonian Orogeny (see Floyd, 1999), and are predominantly younging to the north.

The steeply inclined Silurian strata of the Straiton area are inferred to have been thrust sheets steepened by crustal shortening. They are considered by Floyd (1999) to be the youngest part of the autochonous/parautochonous cover on the basement that includes the Ballantrae Ophiolite complex (Floyd, 1999). Four major northward-directed thrust faults bearing strata of Ordovician age are presently exposed to the south within the Carrick district of the Midland Valley terrane (Williams, 1962; Floyd, 1999). These older strata lie on low-angle, south-dipping thrust faults of which the Pyet Thrust is the lowest. The Silurian strata of the Straiton area were probably steepened and overturned as they were overthurst toward the north-west from the Southern Uplands terrane. This deformation affects the strata up to early Wenlock age but is overlain unconformably by the Lanark Group.

The main Girvan outcrop north-west of the Straiton Fault is strongly tectonised, particularly at its southern margin, and the strata are generally overturned. There are older strata excised by the fault near-parallel to strike as inferred from the absence of lower units of the Drumyork Flags Formation (and lower parts of the Dailly Subgroup) recognised in the Carrick district (Floyd, 1999), about 4 km along strike to the south-west.

By comparison with the model for the Midland Valley Terrane described from the Carrick district to the south (Floyd, 1999), the structural pattern of the Llandovery–Wenlock strata suggests that they belong to a steep thrust stack which is overturned to the north-west. Since the sequence was deposited on an actively subsiding, rifted shelf, at the southern edge of Laurentia, some of the rifts may have been reactivated when the Southern Uplands was finally thrust over the Midland Valley. This reactivation has been related to sinistral transpression occurring within the southern part of the Midland Valley at that time (Smith, 1995).

The strike of the Llandovery–Wenlock strata overall is about 070°, and most dips vary from vertical to about 60° to the south-east. There are few minor folds or flexures on the steeply dipping strata, but there is no evidence of more than one phase of folding. Within the Drumyork Flags Formation at [NS 3335 0248], locally north-west dipping beds (about 60°), younging north-west, indicate minor upright folding within the overturned stack. There are, however, numerous sheared and crushed zones close to the Straiton Fault and locally [NS 3767 0412] a crenulation cleavage has developed dipping 60° to the south-east. These features could have formed as a result of reactivation of the earlier thrusts or the Straiton Fault. The Kerse Loch Fault also has a probable post-Caledonian history (see below).

Late Silurian–Early Devonian events

During Siluro-Devonian times, a period of tension and formation of fault-defined basins during sinistral strike-slip (Smith, 1995; 1999a) accommodated the accumulation of the Lanark Group. The source for the Greywacke Conglomerate may have lain to the south-east within the Midland Valley terrane (Bluck, 1983), which has since been overthrust by the Southern Uplands or covered by younger strata.

Limestone clasts from the Isle Port Conglomerate Member (Swanshaw Sandstone Formation) are of latest Llanvirn to mid Cardoc age and their derivation from the south-east, suggests that they may have been sourced from the Stinchar or Craighead limestones. Conodonts in the clasts have been heated to 100–200°C (Conodont Alteration Index 3) (Armstrong and Owen, 2000). Oliver et al. (1984) recorded values of 60–140°C (CAI 2) from Silurian strata near Girvan, but noted that older Ordovician conodonts in the cover sequence of the Ballantrae complex have values between 1 and 2.5 CAI (Bergstrom, 1980 in Oliver et al, 1984). The elevated CAI values in the Isle Port Conglomerate clasts were used by Armstrong and Owen (2000) to infer a southerly source (now concealed) within the Midland Valley terrane. Although these temperature values could be consistent with localised heating from igneous activity, Armstrong and Owen (2000) cite faunal differences and the results of paleaogeographical analysis as favouring their interpretation.

Locally small synsedimentary faults are observed within the Isle Port Conglomerate Member, near Croy [NS 243 128]. These minor, normal, north-east-trending synsedimentary faults have throws in the order of 0.15 m.

Elsewhere local deformation is related to the penecontemporaneous volcanic eruptions, for instance, close to the Mochrum Hill Vent [NS 265 101], where steeper dips are observed in the strata adjacent to the vent, dipping moderately (up to 40°) into the vent. The steeper dips around in the volcaniclastic sandstones of Brockloch Glen [NS 295 117] could be due to the development of Lower Devonian vents or faults accommodating the venting.

Mid Devonian events

In Mid Devonian times a more widespread tectonic event, which can be equated with the Acadian Orogeny, affected the Lanark Group in the southern Midland Valley. Acadian deformation in the Midland Valley (Friend et al., 2000) would appear to have occurred during late Emsian times (about 400–394 Ma), followed by a period of nondeposition and erosion until Famennian times.

Strictly, the Acadian deformation event affects Lower Palaeozoic and Lower Devonian strata south of the Iapetus suture (Dewey and Strachan, 2003). Dewey and Strachan found it difficult to reconcile this deformation event in the Midland Valley with the 'flat-lying' and almost undeformed Early Devonian above the Southern Uplands terrane. However, Early Devonian strata in the latter terrane are tilted and unconformably overlain by Upper Devonian beds around [NT 900 600] in the Eyemouth district (Greig, 1988). It appears therefore that the effects of the 'soft' nature of the Avalonia–Laurentia collision are observed in the Southern Uplands and Midland Valley terranes. In the southern part of the Ayr district, this collision produced open folds north-east of Maybole around [NS 310 115], but farther south near Straiton around [NS 370 030], tighter folds with overturned limbs were controlled by the earlier Caledonian 'grain' and fault structures.

North of the Kerse Loch Fault

The effects of the Acadian (and subsequent Variscan) orogeny are minimal on the coast near Croy [NS 246 126] and the overall dip of the Lanark Group strata is very gentle (2–4°) northwards, increasing to 15–20° around Dunure. North-east of Dunure around [NS 280 175], the Acadian event produced the unconformity and change in dip between the Lower Devonian Carrick Volcanic Formation (10–20°) and the Upper Devonian and lower Carboniferous successions (5–10°).

Inland east of Maybole [NS 314 101], a gently northward dipping succession is inferred from the limited exposures in the Swanshaw Sandstone Formation.

In the Turnberry area, Swanshaw Sandstone Formation strata dip at 8–20°, swinging from south-westerly dip directions near Ballochneil [NS 227 065] to north-westerly near Laigh McGownston [NS 221 035], defining a broad synclinal fold with an approximate west-north-west axis running from Turnberry [NS 205 055] to Chapelton Farm [NS 240 045]. A complementary anticlinal trace runs west-north-west through Morriston [NS 2315 0855]. These folds are impersistent structures, possibly related to transpression between the Kerse Loch and Maidens faults.

Dip directions also define some smaller scale folds at Chapelton Burn [NS 229 049] and upper Chapelton Burn [NS 258 056]. These minor folds have axes on an east-north-east to north-east trend, and could be related to Acadian deformation. Faults are rarely exposed or inferred in the Swanshaw Sandstone Formation. However, a near-vertical, north-trending fault in Chapelton Burn [NS 2189 0467] is evident at a one metre wide brecciated zone, which appears to control the westward extent of the Chapelton Burn Member (and could post-date Acadian deformation).

At Carlandcheek [NS 2763 1818], and possibly for a few kilometres inland, the geometries suggest onlap of the Stratheden Group onto the steep topography of the Carrick Volcanic Formation, probably across a north-north-west trending palaeofault that was active during the Acadian.

Acadian deformation only weakly affects the Siluro-Devonian rocks south of Dalrymple around [NS 352 114], forming open folds with axes of east to east-north-east trend. A weak open syncline with an easterly axial trace, which is developed in the Lanark Group rocks south of Cassington [NS 352 114] and Montgomerieston [NS 370 118], is considered to be Acadian. This is because it has a gentle easterly plunge, which is not evident in the limited exposures of Carboniferous rocks in the fault blocks in the east of the sheet.

Straiton area (south of the Kerse Loch Fault)

Acadian folding in the Swanshaw Sandstone Formation south of Straiton [NS 370 030] occurs on axial traces trending about 072°, and with fold wavelengths of about 1–2 km. This wavelength pattern is not apparent in the Lanark Group rocks to the north-west of the Kerse Loch Fault, and may reflect the fact that they lay to the north-west of the thrust front which affected the Lower Palaeozoic rocks (compare with the Craighead Anticline). Close to the Straiton Fault the Swanshaw Sandstone strata are vertical to slightly overturned, suggesting that the fault had an element of reverse movement during Acadian or later compression. Where the massive pile of the Carrick Volcanic Formation overlies the sandstones, folding was harder to accommodate and locally faulting is more common, for example south-east of Craig around [NS 393 021].

Late Devonian events

After the Acadian deformation, there is evidence in the north of the Midland Valley for Upper Devonian deposition in a transtensional environment (Bluck, 2000). In this district the Stratheden Group appears to have been deposited against the palaeofault at Carlandcheek, but it is uncertain if it was active in the Famennian.

Syn-Carboniferous events

During the deposition of the Inverclyde Group the variation in thickness of the formations may have been influenced by sinistral transtension, for example along the Kerse Loch Fault. The increased thickness of the Inverclyde Group on the south-east side (hanging wall) of the Kerse Loch Fault suggests that it was acting as a growth fault from early in the Carboniferous (Figure 12).

North of the fault, volcanism occurred contemporaneously with the deformation of the Strathclyde Group (exemplified by the Heads of Ayr Vent), and concomitant tectonics, which had become predominantly dextral (Read et al., 2002), may have caused the unconformity below (and possibly the one above) the Greenan Castle Pyroclastic Member (Clyde Plateau Volcanic Formation).

North of the Kerse Loch Fault, and extending up to the Inchgotrick Fault, is the Mid Ayrshire Block (Read et al., 2002). Within this block, a persistent palaeohigh developed around Ayr and is probably the cause of intra-Clackmannan Group unconformities in that area. The renewed volcanism resulting in the Troon Volcanic Member was apparently accompanied by movement on faults such as the Inchgotrick Fault. This occurred during a continuation of the dextral-oblique slip regime, which was accompanied by the development of sub-basins.

Variscan deformation (end Carboniferous)

The main deformation affecting the Carboniferous strata occurred at the end of the Variscan (post-Bolsovian) Orogeny which in the western part of the Midland Valley tightens structures overprinting Caledonian features. This is the case, for example, with the Dailly Syncline [NS 265 024] (Monaghan, 2004). Any continued dextral movements appear to have been taken up on the Caledonoid bounding faults. This contrasts with the well-developed north–south axes of folds in the eastern part of the Midland Valley (Read et al., 2002). Many of the faults in the Carboniferous rocks are the result of Variscan movements and are considered in this section. The larger faults which commonly grew during the deposition of the Carboniferous strata are considered in a separate section on faulting. To the north of the Kerse Loch Fault, the Carboniferous strata are gently dipping in open, generally north-west trending 'basins and domes'. It is not easy to distinguish syn-Carboniferous deformation from the Variscan events. Much of the folding in Carboniferous rocks appears to be Variscan, although the very gentle dips eastwards into the Mauchline Basin were enhanced after Permian deposition.

Maybole–Kirkoswald area

The major north-east trending Kerse Loch Fault in the Craighead area has Lower Palaeozoic rocks cropping out in its footwall block and mid Carboniferous rocks in the hanging-wall block. The throw on the Kerse Loch Fault is estimated as a minimum of two kilometres south of Maybole around [NS 275 038]. The smaller faults subparallel to the Kerse Loch Fault, trending north-east to north-north-east, affect mainly Siluro-Devonian rocks, but probably have a history into lower Carboniferous times on the evidence of the small outlier of Ballagan Formation occurring north-west of the Kerse Loch Fault.

South of Kirkoswald, for example at [NS 240 063], the interpretation presented on the old maps (Second Survey, Ayr County series, 1928) taken in combination with the magnetometer survey results (Raines et al., 1997) suggest that the andesitic intrusion is cut by several north-west to west trending normal faults. These faults are inferred to have throws over about 20–40 m and with such trends, could relate to Carboniferous tectonism.

Dailly area

The main faults crossing the Dailly area are the generally north-easterly trending Kerse Loch and Bargany faults (Figure 12) which bound the Carboniferous outcrop. The Variscan Orogeny is responsible for tightening the Dailly Syncline, the north-eastern end of which can be discerned in the vicinity of Little Craigfin [NS 303 506]. The Dailly Syncline has a low plunge to the south-west and its trace is roughly parallel to the Kerse Loch Fault (Figure 12). The syncline is also cut by subsidiary faults, subparallel to the Kerse Loch Fault. Since there is no complementary anticline lying to the south-east, and the north-western limb of the syncline is less than a kilometre wide compared to 3 km for the south-eastern limb, it is thought that the fold is mainly the result of drag (or synsedimentary growth since the Clackmannan Group is thicker here than to the north) on the Kerse Loch Fault. Although there is little exposure in the Crosshill area [NS 325 065], the broad outcrop pattern of the Ballagan Formation is interpreted as the result of a thick Ballagan succession lying in the north-eastwards continuation of the Dailly Syncline, which is opening out in that direction.

These structures are interpreted to be reactivated features that were generated by the Caledonian Orogeny, and have a long history of movement. The north-east trending Bargany Fault, with Siluro-Devonian strata in its footwall, forms the south-eastern side of the synclinal graben structure in which the Carboniferous strata of the Dailly area are preserved. The Kerse Loch Fault forms the north-western side to the graben. The trace of the fault is not straight, trending from east-north-east to north-north-east, with several en échelon strands. The Ordovician–Silurian Craighead Anticline lies in the footwall of the Kerse Loch Fault. The north-east trending Carboniferous Dailly Syncline lies in the hanging-wall block of the Kerse Loch Fault. The axis of the syncline is curvilinear, such that it closes to the north-east where it plunges up to 24° to the south-west. The syncline closes to the south-west on the adjacent Carrick Sheet to the south, where it trends 050° and plunges up to 10° north-east (Floyd, 1999).

A subsidiary north-east trending anticline, plunging gently south-west, lies on the northern side of the syncline where the Kerse Loch Fault has an east-north-east trend and an en échelon step [NS 2555 0238]. The north-eastern end of this anticline is cut out by the main Kerse Loch Fault, and a portion of Clyde Sandstone Formation appears to have been offset sinistrally at least 500 m in a fault wedge. Since dextral movement is considered to have been dominant from Mid to Late Mississippian times into the Pennsylvanian, this may indicate sinistral movements continued into Tournaisian times.

Dips within the Carboniferous strata around the Dailly Syncline are commonly about 30°, but on the north-western side of the asymmetric fold, dips between 40 and 70° are more commonly observed. Clearly, the fold has been tightened in post-Namurian times. Borehole and mining records show that the Limestone Coal Formation strata thicken and become more coal-dominated from the south-western end to the centre of the syncline (Eyles et al., 1949), indicating that some synsedimentary basin development occurred. However, Mykura (1967) noted that major thickness variations were not related to the current shape of the basin. Field mapping demonstrates that the Lawmuir Formation thickens from approximately 30–50 m on the north-west and south-western sides of the syncline to about 100 m on the south-east and north-east (see Floyd 1999; Smith, 2003).

The Carboniferous strata of the Dailly area are cut by smaller faults with a several orientations (Figure 12). Some structures are subparallel to the main graben-bounding faults of north-east trend and are probably related to them, for example the antithetic fault bounding late Carboniferous strata near Roan of Craigoch [NS 2954 0429]. Another set of normal faults trend from west- to north-west and are well documented by mine plans, dipping at approximately 70°. The Lawmuir Formation is interpreted as thickening over one of these west-north-west-trending faults south of Drumburle [NS 2881 0279] but this geometry is very poorly constrained, and the apparent thickening could relate to faulting and folding in the hanging wall of the fault.

The thickness of the Ballagan Formation increases north-eastwards from Dailly, but its base is truncated by the cross-cutting Bargany Fault to the south.

Smith (2003) documented the evidence that thickening early Carboniferous successions between Dailly and Crosshill occupied a releasing bend on the Kerse Loch Fault during sinistral oblique slip. This tectonic framework agrees with local fault measurements and published studies from other parts of the Midland Valley (e.g. Smith, 1995; Rippon et al., 1996). This sinistral tectonic regime had continued on from the Mid Devonian Acadian events. However, by Mid to Late Mississippian times, the regime in the Midland Valley appears to have been dominated by dextral oblique-slip (e.g. Read, 1988; Underhill et al., 2008). In the Dailly area, this change in tectonic regime is perhaps reflected in the presence of the Clyde Sandstone Formation–Lawmuir Formation unconformity.

In the main Ayrshire Coalfield to the north of the Dailly area, the Late Mississippian to Mid Pennsylvanian dextral oblique-slip regime meant that faults such as the north-east to easterly trending Kerse Loch were active, resulting in significant syndepositional stratal growth (Eyles et al, 1949; Monro, 1999). Smith (2003) documented post-Mid Mississippian dextral strike slip measured in a fault close to the Kerse Loch Fault, and interpreted stratal thickness variations around the Dailly Syncline as a result of drag or growth on the fault, linked to dextral oblique slip.

In summary, evidence from the Dailly area supports the late Visean to late Carboniferous dextral oblique-slip hypothesis, with movement on the Kerse Loch and Bargany faults and the possible development of a proto-Dailly syncline. For example, syndepositional thickness variations in the Lawmuir and Limestone Coal formations could be related to growth fold development on a releasing bend of the Kerse Loch Fault, differential subsidence between the Bargany and Kerse Loch structures, and/or associated with coeval movement on west- to west-north-west trending faults at the releasing bend (Figure 12). The small subsidiary anticline on the north-western side of the syncline occurs at a local restraining bend of the fault in the dextral oblique-slip system.

In common with fold structures across the Midland Valley, the Dailly Syncline is interpreted to have been tightened by late Carboniferous to Permian end-Varsican tectonism.

Ayr–Prestwick area

Within the Ayrshire Coalfield around Ayr, the overall dip of the strata is gentle towards the north, but dips are variable due to sub-kilometre scale dome and basin folds. The majority of measureable faults are high angle (60–90°) and normal, and considered to be postdepositional. Where there is sufficient data, larger faults appear to cut the entire Middle Coal Measures succession. Numerous faults create a complex dip and outcrop pattern and appear to crosscut and offset the dome and basin structures, implying that the gentle folding preceded faulting. Bedding dips are commonly between 5–10° though locally reach 25–30° adjacent to faults.

In the field, small normal faults (throws of tens of centimetres) are commonly exposed (e.g. Laigland [NS 3835 2242]), whereas larger faults are rarely exposed. However, one major normal fault, downthrowing Middle against Lower Coal Measures, is exposed in the north bank of the River Ayr [NS 4007 2170].

Data from mine abandonment plans and field observations indicate that three main orientations of faults are present. Most common are faults trending north-west to west-north-west, with throws of a few metres to tens of metres and spacings of approximately half a kilometre. Mykura (1967) suggested that these structures moved in the late Carboniferous, preceding the deposition of the Mauchline Volcanic Formation. In places, north-west-trending faults cut older faults striking north-east along reactivated Caledonoid trends (e.g. Mossblown at [NS 3987 2487]). Some north-east-trending faults (e.g. the Inchgotrick Fault in the north) and north-westerly faults, which in between Monkton [NS 360 280] and Rosemount [NS 375 291] indisputably cut the Prestwick–Mauchline Sill (Mykura, 1967 p.92) are now known to be Palaeogene and must therefore have moved after 55 Ma. A third set of faults observed near Prestwick Airport (e.g. at [NS 3621 2606]), trending north-north-west, cut and offset north-west to west-north-westerly structures. The faults trending north-north-west are high angle structures, traceable for 3–4 km along strike, with throws up to 70 m. They appear to be the youngest structures, orientated subparallel to the Palaeogene dyke trend with the north-west-trending faults curving towards them, perhaps indicating some oblique slip movement (e.g. at [NS 3621 2606]).

East-trending faults of similar age are identifiable for a few kilometres along their length, with throws of about 30 m either to the north or south (e.g. near Brieryside [NS 374 275]). Abundant smaller faults with throws less than 5 m are locally observed in faulted zones a few hundreds of metres wide, for example at Brickrow [NS 382 249].

Domes and basins in the Inverclyde Group may be due to early Carboniferous movement (possibly related to the eruption of the Clyde Plateau lavas farther north). This affected the area around Ayr creating a palaeohigh across which Mid Mississippian strata were either not deposited or are severely condensed. On the coast, open folds in the Ballagan Formation (Plate 22) plunge gently north-west.

Around Martnaham Loch, the broad dome and basin post-Westphalian folds have wavelengths in the order of one kilometre, and the anticlinal cores of Kinnesswood Formation are inferred from changes in dip. In general the fold axes have a north-east to north-north-east trend. On the south-east flank of the anticline passing through High Balsarroch [NS 364 153] dips are locally 30–40° to the south, whereas to the north dips are between 10° and 24° northwards.

Straiton area

Around Straiton, Carboniferous strata have been affected by reactivation of the Caledonoid Straiton Fault. North of the fault, Carboniferous strata are generally gently dipping but, owing to the interruption by faults, few folds can be defined at any scale. Minor folds were, however, observed in the March Burn [NS 378 073] section, with hinges plunging at 5° eastwards to 088°.

Dalrymple area

The effects of syn-Carboniferous deformation and/or the Variscan Orogeny around Dalrymple are generally limited to normal faulting, although dextral strike slip movements may have deformed the strata (Read, 1988). Within the Carboniferous strata, new biostratigraphical data has led to the inference of faults with north-westerly trends. One passes through Kennedy's Dub [NS 398 138] and another just north-east of Netherton [NS 373 136]. Overall these faults appear to form a graben within the Inverclyde Group, which could have developed as a result of dextral movement on the releasing bend of the Kerse Loch Fault. North-west trending faults downthrowing to the north-east are evident in the field to the east of both Ballycoach [NS 362 117] and Montgomerieston [NS 369 125]. Less common, but probably of similar age, are north-north-easterly trending faults such as the one between Ballycoach and Montgomerieston.

Syn- to post-Permian events

The open north-west-trending syncline forming the Mauchline Basin developed during or after Permian times. The Clackmannan Group strata south of the Littlemill Fault lie on the western limb of a related subsidiary syncline (Mykura, 1967). Although faults trending north-west are associated with the development of the Mauchline Basin, such as the Mossbog Fault on Sheet 14E, these are difficult to differentiate from those which also affect the Palaeogene intrusions. A set of north-east trending faults which intersect these faults are also considered to be post-Permian. Mykura (1967) also considered that the Lanark Group forming the Carrick Hills was the core to a complementary anticline plunging gently north-westwards. Offshore the Carrick Fault (see below) bounds a Permo-Triassic basin, and may have a syn-rift history. Within the Carrick Basin, a north-north-east-trending synclinal structure has been postulated (McLean and Deegan, 1978). This trend is oblique to that of the Mauchline Basin, but it may be a 'growth syncline' in the hanging wall of the Carrick Fault. Within the Carboniferous rocks offshore in Ayr Bay, several impersistent folds are interpreted as having north-north-east trends in the southern part, which swing to trend north-easterly farther to the north-east (McLean and Deegan, 1978). It is not clear if these are Variscan or later structures, although they are congruous with those within the Permo-Triassic rocks.

Palaeogene to Neogene events

Tension in a north-east to south-west direction associated with initial rifting of the North Atlantic allowed Paleocene mafic intrusion over a period of about 5 Ma, punctuated by faults affecting the Prestwick–Mauchline sill (Mykura, 1967). The normal faults affecting the sill have various trends, including east-north-east and west-north-west suggesting a period of north–south tension before returning to north-east to south-west tension when the tholeiitic dykes were intruded. A prolonged period of uplift and erosion appears to have continued from Eocene into Neogene times in this district. No Neogene faults have been detected and there are no historical records of earthquakes affecting the Ayr district.

Faulting

Nearly all the major north-east trending faults have acted in syn- or post-Carboniferous times, but some probably had a longer history starting with the Caledonian Orogeny. Other faults trending north-north-west and north-north-east probably have a history ranging from Permian times into the Palaeogene. The named faults on the Bedrock edition are described briefly below together with small, but significant faults.

Inchgotrick Fault

In the northern part of the district the normal Inchgotrick Fault (Figure 17) is interpreted to extend south-westwards offshore from south of Meikle Craigs [NS 325 285] where it downthrows the Prestwick–Mauchline Sill complex against the Scottish Middle Coal Measures Formation. The Inchgotrick Fault must have had a significant downthrow in Palaeogene times in order to have affected the sill complex, yet it is crossed without offset by the Palaeogene Cumbrae–Stevenston dyke. The fault extends north-east across the adjacent Irvine sheet, but trends more easterly in the Kilmarnock district, where the throw increases as indicated by the downthrow of Scottish Lower Coal Measures against Lower Devonian lavas. The Inchgotrick Fault appears to be intersected offshore by later Palaeogene (or Neogene) movement on the major north-north-east-trending Carrick Fault, which has been mapped roughly parallel to the coast in the west of the sheet. The Carrick Fault downthrows to the west-north-west, and the throw increases to the south where it bounds a Permo-Triassic offshore basin, so it may have an earlier rift history dating from that period. It is also associated with fracture zones up to 500 m wide along parts of its length (British Geological Survey, 1985a).

Maidens Fault

Onshore north-west of Maidens [NS 212 084], the north-north-east-trending Maidens Fault downthrows Carrick Volcanic Formation to the west of Swanshaw Sandstone Formation. This fault may be related to Acadian or Carboniferous deformation, but if it extends below superficial deposits into Turnberry Bay, it could be a splay of the larger Carrick Fault and have moved syn- or post- Permo-Triassic times (but probably not since the intrusion of Paleocene dykes as they are not offset by the fault). Another fault, also with a north-east to northerly trend, downthrows the Carrick Volcanic Formation against the Swanshaw Sandstone Formation near Culzean Castle [NS 233 100]. Assuming that the Carrick Volcanic Formation lies at approximately the same horizon above the Swanshaw Sandstone Formation, the downthrow on the Maidens Fault would be in the order of 300 m if the fault was normal, but it may only be tens of metres if sinistral strike slip is the main component of the fault. An alternative interpretation that could be supported by the dips of the Carrick Volcanic Formation, would be to extend the Swanshaw Sandstone Formation–Carrick Volcanic Formation unfaulted contact further to the north-east (at both Culzean and Maidens) and decrease the offset and throw of the faults.

Kerse Loch Fault

The major Kerse Loch Fault originated in pre-Carboniferous times (probably Caledonian as it bounds the southern side of the Craighead Anticline) and was active intermittently during the Carboniferous, affecting strata as young as Westphalian (Bolsovian on the Cumnock Sheet). In the Kirkmichael area the Kerse Loch Fault has a general north-easterly trend inherited from Caledonian structures. In this area the fault is not exposed, but mapped with a curvature that is concave on its downthrow side to the south-east. During Pennsylvanian times it is thought to have been active (Read, 1988), and at that time may have had a dextral strike-slip component. During these times therefore, the curvature on the fault is interpreted to form a constricting bend on its south-east side in the vicinity of Crosshill, but there is no evidence at outcrop due to lack of exposure. Near Little Craigfin [NS 3023 0542] in Lawmuir Formation sandstone, a small fault strikes 004° dipping 74° to the west with slickencrysts indicating dextral movement in Visean times or later. Near the Water of Girvan, another small fault in Lawmuir Formation sandstone strikes 338° and dips 70° to the north-east. However, earlier in the Carboniferous it is possible that a sinistral regime still prevailed, continuing on from the Mid Devonian Acadian Orogeny (Smith, 1995). This would concur with the evidence of sinistral movements in the Loch Spallander area (Smith, 2001), and account for the normal faults within Lanark Group rocks in the releasing bend of the Kerse Loch Fault around Craigfin (Figure 17). These small north-east to north-north-east-trending normal faults cut the Swanshaw Sandstone Formation and the igneous sills in the Springgarden Bridge [NS 302 072] to Carsloe Glen [NS 305 064] areas.

In exposures in Dyrock Burn [NS 3526 0911], this fault is associated with a minor fault striking 076° and dipping 80° north-west with sinistral movement indicated in near horizontal slickensides. Minor (probably later) north-westerly trending faults strike 347° and dip 65° north-east or strike 337° and dip 76° north-east.

Littlemill Fault

The Littlemill Fault at the eastern edge of the district near Shankston Loch appears to be an antithetic splay of the Kerse Loch Fault and downthrows Scottish Lower Coal Measures to the north, between the faults. There is evidence of a growth syncline developed between the Kerse Loch and Littlemill faults (Mykura, 1967), but this is more evident on the adjacent Sheet 14E.

In the Dailly area oblique movement on the Kerse Loch Fault is associated with the growth syncline which tightened to become the Dailly Syncline (see above).

Cloncaird Fault

The Cloncaird Fault, a splay of the Kerse Loch Fault, which lies 1 to 2 km south-east of it, crosses the Water of Girvan [NS 358 072] and repeats the downthrow of the Kinnesswood Formation to the south-east of the Swanshaw Sandstone Formation (Figure 17). The Cloncaird Fault is interpreted to intersect the east-south-east trending Blairquhan Fault (Smith, 2001).

Lambdoughty Fault

The fault affecting lower Carboniferous rocks and running east-north-easterly north of Loch Spallander [NS 393 093], is an extension of the Drumgrange Fault established in the Patna area to the east (Mykura, 1967; Institute of Geological Sciences, 1976) and downthrowing to the north-west. The north-north-west trending Lambdoughty Fault (Figure 17), south of Loch Spallander, is interpreted to downthrow Ballagan Formation to the west, as the sandstone succession to the east is inferred to be Kinnesswood Formation. This fault is not exposed in the area and may become insignificant along the southern part of its trace. Its northern end is interpreted to be cut by the westward extension of the Drumgrange Fault. North of the latter, the unnamed fault crossing Backglen Burn [NS 383 098], downthrows the Ballagan and Clyde Sandstone formations to the east.

Blairquhan Fault

Near Knockgardner, the Kinnesswood Formation is unconformable on the Silurian inlier, but at the eastern end of the inlier [NS 380 050] the Blairquhan Fault throws the Kinnesswood Formation down against the Silurian strata. To the west of Blairquhan, opposing dips in Kinnesswood Formation strata occur across the fault.

Some minor post-Devonian faults strike about north-east to north-north-east within the Swanshaw Sandstone Formation, and cut Early Devonian andesitic intrusions (e.g. at [NS 296 065]). Another small north-north-east-trending fault offsets the conglomerate unit within the Swanshaw Sandstone Formation west of Kilhenzie Castle [NS 305 082]. Some of these faults throw to the south-east; others to the north-west with downthrows of tens of metres.

Cloyntie Fault

Near the southern margin of the sheet [NS 332 054] a significant north-north-westerly-trending fault, the Cloyntie Fault, is inferred to downthrow the Ballagan Formation against the Kinnesswood Formation near Cloyntie.

Planar normal fault in Bracken Bay

In Bracken Bay [NS 279 182], a planar normal fault dipping at 60–70° and striking east to east-north-east can be observed bringing the Ballagan Formation and Stratheden Group into contact. Granulation seams can be seen in the footwall sandstones and dip variability is noted in the Ballagan Formation hanging wall block. The throw is of the order of 50–100 m. The earlier edition of the map (Institute of Geological Sciences, 1978a) continues the fault eastwards. At the edge of Bracken Bay the fault was measured as striking at 254°, and therefore the interpretation here is that the fault tips out laterally and the Kinnesswood and Ballagan Formations are conformable to the east.

Minor steep-dipping dextral transcurrent faults west of Beoch

Within the Carrick Volcanic Formation west of Beoch [NS 284 144], there are minor steep-dipping dextral transcurrent faults trending 041° and 355°. The latter offset the local trap topography and the junction between the Swanshaw Sandstone and Carrick Volcanic formations. Other faults affecting the volcanic formation strike east-north-east and west-north-west. The timing of these faults is debatable but the dextral transcurrent set is probably Mid Mississippian to Pennsylvanian in age (Read, 1988, 1989; Underhill et al., 2008).

Straiton Fault

South of Straiton, the north-east trending Straiton Fault (Figure 17), has some north-westerly downthrow inferred to be post-Carboniferous, but its main displacement is an earlier Caledonian event, which throws the Lower Devonian strata down to the south-east against the Silurian rocks of the Main Girvan inlier. The continuation of the Straiton Fault to the north-east is thought to be the Headmark Fault (Simpson and MacGregor, 1932). Since the latter fault has a post-Carboniferous downthrow to the south-east (i.e. opposite to the Straiton Fault), it must be a scissor fault, or involve strike-slip movement in post-Carboniferous times. The Straiton Fault may have developed above a Caledonian thrust stack, dipping south-east between Caradoc and Llandovery to Wenlock strata, and reactivated as a result of tension in Siluro-Devonian times. It may link with the Camregan Fault to the south-west below the Lanark Group on the Carrick Sheet (British Geological Survey, 1995). East of the Blairquahan Fault, it appears to have acted as a normal fault downthrowing Kinnesswood strata to the north-west in post-Carboniferous times.

Bargany Fault

In the south of the district, the Bargany Fault lies sub-parallel to, and just to the north-west of the Straiton Fault, and extends south on to the adjacent Carrick Sheet (British Geological Survey, 1995), acting as the south-eastern limit to the Carboniferous strata in the 'Dailly Graben'. The Bargany Fault is antithetic to the Kerse Loch Fault, but may be the result of reactivation in an extensional regime (at least into Carboniferous times) of an earlier Caledonian fault.

Knockinculloch Fault

The Knockinculloch Fault in the south-east of the district has a relatively small downthrow to the south-east within the Lanark Group rocks, but it has a Caledonoid trend and extends south-westwards into the adjacent Carrick district, where it separates tracts with different Ordovician strata. It may therefore have had a Caledonian history before being reactivated, probably during the Mid Devonian (Acadian).

Chapter 10 Concealed geology

The concealed geology of the Ayr district has partly been elucidated by the use of seismic, gravity, ground magnetic and aeromagnetic methods. Early work applying gravity anomalies to the geological structure of the district (McLean, 1966) identified a high around Kirkoswald, with lows around Dailly and towards the Mauchline Basin. A summary of earlier offshore geophysical surveys in the Firth of Clyde is given in McLean and Deegan (1978).

Crustal structure

The basement structure of the district is inferred from the overall structure ascribed to the Midland Valley Basement Terrane (MVBT), the deep structure of which is largely known through the LISPB seismic refraction profile (Bamford et al., 1978). This shows that the crust of the MVBT has a three-layer structure, in contrast to the two layer structure in the rest of the UK, with a relatively deep Mohorovičić discontinuity ('Moho') at about 35 km depth, marking the boundary between the crust and the upper mantle. The MVBT seismic refraction profiles across the central Midland Valley (Dentith and Hall, 1989) confirmed the presence of the high velocity refractor, at a depth of about 8 km, that is typical of the basement of the Midland Valley and is considered to represent the top of a granulite gneiss basement layer.

Based on these seismic data and on potential-field modelling, Kimbell et al. (2006) have published a network of regional interpretative profiles across southern Scotland and northern England. Profile Two of this set, which passes north-west to south-east through the centre of the Mauchline Basin, lies some 16 km to the north-west of the district, and it is considered that the deep crustal structure modelled on this profile is similar to that beneath the district.

Bouger anomaly data

Onshore, the available gravity measurements form part of the UK onshore gravity databank that is maintained by BGS. These stations have an average spatial density of one per square kilometre, although locally there may be larger voids in the coverage, mostly due to the lack of vehicular access (for example, near Blacktop Hill [NS 2689 1540]). The offshore part of the district is very poorly covered by sea-bottom measurements that were made in 1964/5. Gravity values are linked to the British National Gravity Reference Net 1973 (NGRN 73), which is itself linked to the International Gravity Standardisation Net 1971 (IGSN 71). Where necessary, terrain corrections have been applied and in general extend to a radius of 48.6 km.

For data reduced using a Bouguer density of 2700 kg.m⁻³, the highest anomaly values, of over 26 mGal, occur near Hollowshean [NS 2417 0600] where the culmination of an area of high values occurs over the Ordovician–Devonian rocks in the south-west of the district (Figure 18). This zone of high anomalies extends southwards to link with anomalies near Girvan, while to the north and north-east, ridges of elevated values extend outwards from the regional high. To the south-east of the zone of high anomalies, a steep north-east trending gradient (Figure 18) is associated with the Kerse Loch Fault (KLF), while the anomaly is bounded on its western and north-western flanks by a steep gradient associated with the offshore Carrick Fault. To the south-east of the KLF, a north-east trending Bouguer anomaly trough is seen over the Dailly Syncline, which in turn is bounded on its southern margin by the north-east trending gravity gradient associated with the Straiton Fault.

The Ayrshire Coalfield is marked by an east–west trending trough of relatively low gravity anomalies, which reach a minimum value of 13 mGal at Annbank [NS 4091 2332]. Within the trough, anomalies gradually increase westwards towards the coast, where they reach 15 mGal. The paucity of gravity stations offshore of Ayr precludes meaningful interpretation of the anomalies.

Aeromagnetic data

The aeromagnetic data (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19) for the district comprise three generations of survey as outlined in (Table 5). Firstly, the entire district, both onshore and offshore, is covered by the 1959 survey flown by Canadian Aeroservices Ltd (CA59) that now forms part of the national aeromagnetic databank (note that this is of limited use for detailed interpretation). Secondly, the offshore and near-onshore parts of the district are covered by a commercial high resolution survey acquired in 1994 by a consortium of JEBCO Seismic Ltd, ARK Geophysics Ltd and Atlantic Reconnaissance Ltd (left hand side of (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19)). Finally, the north-west of the district is covered by BGS high resolution aeromagnetic data (HIRES) acquired in 2004 (Busby et al., 2009).

The high resolution data were processed to enhance subtle near surface features, and the resulting anomalies, together with the unprocessed data, were loaded into a customised GIS where it could be overlain on the mapped geology. Interpretation of the data (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19) was made in the GIS by identifying, digitising and classifying features.

A prominent south-south-west trending magnetic lineation extending offshore from [NS 3309 2871] marks the offshore part of the Inchgotrick Fault (IF). To the north of this, an east–west-trending magnetic high appears to be associated with the offshore extension of a splay of the IF, along which is a chain of minor intrusions, such as that at Lady Isle [NS 2749 2933]. Prominent east-south-east-trending linear anomalies that lie between the offshore IF and the coast probably represent igneous layering within a buried sill.

North-west trending linear anomalies within the district (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19) correspond with dykes that are exposed either locally (especially at the coast), are known from underground workings or may be totally concealed. Busby et al (2009) interpreted a north-west-trending linear negative anomaly about 3 km south-west of the Cumbrae–Stevenson Dyke as the Cleveland Dyke. Comparison of the geophysical anomaly with the mapped geology suggests that this dyke is also known from underground workings near the racecourse at Ayr [NS 3580 2220]. Two dimensional modelling of the dykes suggests that their width increases with depth (Busby et al, 2009). Carboniferous intrusive rocks of the Prestwick–Mauchline Sill complex generally exhibit a less coherent anomaly pattern, although there are areas, for example near East Sanquhar [NS 3673 2433] and near Bentfield [NS 3415 2484], where these rocks are associated with anomalies.

Around Brown Carrick Hill [NS 2920 1620], the Lower Devonian Carrick Hill Volcanic Formation (CRKV) is associated with short-wavelength magnetic anomalies, and is cut by two north-west trending dykes that crop out on the coast at Bracken Bay [NS 2760 1821] and Deil's Dyke [NS 3063 1942]. The first of these dykes shows apparent dextral displacement at Sauchrie Burn [NS 2989 1568], where a north-east trending basalt dyke is mapped. The pyroclastic rocks of the Heads of Ayr Vent [NS 289 187] exhibit a minor anomaly.

In the south-eastern corner of the district, the long-wavelength negative magnetic anomaly (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19) forms part of the north-east trending regional anomaly that is associated with the Southern Upland Fault. In the south-west of the district, a north-east trending magnetic high near High Craighead [NS 2285 0130] is associated with Ordovician basaltic lavas, possibly of the Balcreuchan Group, while the magnetic high 500 m north-west of Knockbrake [NS 2904 0746] is related to intrusive Devonian andesites. A low-amplitude high, extending westwards from this anomaly to the mapped andesites and basalts at Hollowshean [NS 2408 0603], suggests that these intrusive rocks are continuous at shallow depth. This area was studied in detail by ground magnetometry in the Milton Wellfield field survey (Raines et al., 1997). Although not associated with a significant magnetic anomaly, the Mochrum Hill Vent [NS 265 100] lies on a north-west trending linear anomaly that is due to a normally magnetised dyke extending offshore from Culzean Bay [NS 245 123]. The dyke cropping out on the coast at [NS 2450 1408] is associated with linear anomalies that trend south-eastwards towards Maybole [NS 291 106]. A positive magnetic anomaly at Culzean [NS 230 102], which extends about one kilometre offshore, is associated with the CRKV.

Subsurface modelling

Various bedrock horizons have been contoured in the Ayr Coalfield where available borehole and mining data allow a subsurface 3D model to be created. (Figure 20a) and (Figure 20b) show the contours for the Ayr Hard Coal and the base of the Lower Limestone Formation respectively, illustrating the effects of faulting which offsets the contoured surfaces (e.g. the Inchgotrick Fault at [NS 336 290]). The figures also show the rapid deepening of the units into the Mauchline Basin. To the west of this basin, the contouring indicates that the geological units are generally flat-lying, with gentle folding on north-west trending axes possibly superimposed on north-east-trending Caledonoid axes. Three-dimensional block models of the Ayrshire Coalfield have been made (Figure 21).

Chapter 11 Quaternary

Pleistocene

Lowland Scotland has been glaciated on several occasions during Pleistocene times (i.e. the last 2.6 Ma). However, erosion during the last main late Devensian (Dimlington) glaciation (Table 6) removed earlier glacial and interglacial sediments from this district that had been deposited during the early Quaternary, and also earlier Neogene sediments. However, opencasting for coal at Sourlie [NS 338 414], in the Irvine district, exposed organic sediments from the last interglacial dated approximately between 33 500 and 29 000 14C BP (Bos et al., 2004).

The late Devensian glaciation lasted from about 29 000 to 14 500 years ago. A return to very cold conditions in Scotland, as represented by the Loch Lomond Stadial (12 700–11 500 BP) caused an expansion of the Highland Ice-Sheet, but this did not extend into the Ayr district.

Older and harder rocks tend to resist erosion, and as a result, the movement of the ice-sheets created ice-moulded bedrock landforms, some of which are preserved and provide an indication of the glacial history. Late Quaternary superficial deposits cover much of the district (Figure 22a), particularly the low ground, and the distribution and nature of the deposits is shown on the 1:50 000 Simplified Bedrock and Superficial Sheet 14W (Ayr) published in 2009.

Earlier models of Pleistocene glaciation in the Ayr district have been published (Smith, 1898b; Charlesworth, 1925; Eyles et al., 1949). Smith's observations, including finding shell fragments within the deposits, led him to conclude that various boulder clays had been deposited in water rather than as a result of 'land-ice' (Smith, 1898b). As Peacock (1995) comments, Smith's views were those of a careful observer, and that even now the resolution of the land ice versus marine origin for several glacial deposits has still not been satisfactorily resolved in some instances. The model put forward by Eyles et al., (1949) involved the transport of frozen shelly clay masses from offshore, essentially as erratics within till. Subsequent work on the Pleistocene glaciation in central Ayrshire by Holden (1977) could find no evidence to indicate any glaciation before approximately 27 500 years BP when the last ice-sheets (late Devensian) submerged the area. More recently, Finlayson et al. (2010) have remodelled the late Devensian ice-sheet in west central Scotland using NEXTMap remote sensing and three dimensional evidence from borehole data.

During the latest resurvey a lithodemic classification was maintained, but more detailed work in the Clyde area and farther north in Ayrshire allows some Quaternary lithostratigraphy to be established in the Ayr district (Table 6).

The lithostatigraphy is mainly based on work by the British Geological Survey (McMillan et al., 2005), which made adaptations to Sutherland (1999). In this district the Quaternary deposits can be classified into glacigenic and non-glacigenic groups (Table 6). These groups interdigitate on a regional scale, and this may also be the case with the Midland Valley and Southern Uplands glacigenic subgroups within this district, but during the recent resurvey no sections were observed which distinguished the subgroups. The Wilderness Till Formation (Browne and McMillan, 1989) is known to be present at least in the north of the district by extension from the Irvine, Kilmarnock and Glasgow districts. Presently there is no evidence of pre-Dimlington Stadial deposits onshore within the district, although borehole records of the thicker superficial deposits around Ayr encountered several layers of till separated by sand or gravel. Offshore, below thin modern day sediment, clays belonging to the Jura Formation (Davies et al., 1984) cover earlier formations. The few offshore boreholes in the area penetrated acoustically layered silty clay (about 20 m) overlying sandy till. The till or diamicton is locally thin (about 3.5 m) resting on bedrock.

Detailed studies of radiocarbon dating and event stratigraphy in the Midland Valley, and other parts of the British Isles (Lowe et al., 1999), has produced an overall chronology for Devensian to Holocene times.

Devensian glaciation

From about 29 000 14C years BP (Bos et al., 2004), ice caps sourced in the western Highlands and the south-west Southern Uplands gradually coalesced, forming an ice-sheet which at its maximum extent (about 26 000 years BP) covered most of northern Britain. During this time the centre of the ice-sheet migrated, altering the direction of ice flow across the Ayr district. Ice movements were complicated by the interplay of ice converging from offshore in the Firth of Clyde and the Southern Uplands to the south (Eyles et al., 1949). There is evidence for an early phase of ice flow from the west, followed by a northward extension of the ice sourced in the Southern Uplands, and then a retreat from east to west (Eyles et al., 1949). Most of the superficial deposits date from the last major glaciation, the main late Devensian glaciation of the Dimlington Stadial. An extensive diamicton was deposited as a lodgement till beneath the ice-sheet. Carboniferous sedimentary rocks, which are less resistant to erosion and usually occupy low ground, tend to be well covered by till; whereas the more resistant igneous rocks and older rock are commonly free of superficial cover. Ice-moulded bedrock, crags and tails, drumlins and glacial striae indicate that the dominant direction of ice flow was from the west, in the Firth of Clyde. Locally in the valley of the Water of Girvan in the south of the district, the ice flow was from the south, sourced in the Southern Uplands. Post-glacial (latest Devensian) deposits around Ayr include glaciofluvial and raised beach deposits.

By the time of the Windermere Interstadial (14 500–12 700 BP) the area was probably free of ice. The relatively subdued relief of the district may explain why the presumed periglacial effects of the succeeding Loch Lomond Stadial are inconspicuous, with no ice wedges or solifluction deposits identified as belonging to that interval.

Evidence for late Devensian ice movements

The NEXTMap image of the Ayr district (Figure 22b) shows a topography dominated by an easterly trending pattern of drumlins and ice-moulded bedrock. These landforms indicate ice flow onshore from the Firth of Clyde. North of Ayr and into the Irvine district, the trend swings to a north-easterly flow.

The asymmetry of crag and tail features (e.g. Mossblown Farm [NS 394 249]) indicate that the late Devensian ice-sheet flowed from west to east.

In the south-east of the district, there is evidence that the latest ice-sheet was sourced from the south-east and east, as indicated by the drumlin orientations and their asymmetry, the numerous Loch Doon granite erratics and the glacial striae in the adjacent hills. The general softness of the sedimentary rocks means that few rock outcrops or glacial striae are exposed in the vicinity of Crosshill and Kirkmichael.

A glacial striation, recorded on the old geological fieldslips in an old 'whinstone' quarry at [NS 3065 0689], trends westwards and another south-west of the church at Kirkmichael at [NS 3446 0893] also trends east–west. A striation was observed on red sandstone in the railway cutting at [NS 308 073], trending 300° and indicating ice flow to the north-west. This supports the flow of ice sourced in the Southern Uplands. In the illustrations of ice currents shown by Eyles et al. (1949), at the glacial maximum, the southern limit of Highland Ice (from the north and west) lay just south of Maybole. At a later stage the southerly derived ice pushed north of Maybole, as indicated by the Loch Doon granite and Lower Palaeozoic greywacke erratics overlying the shelly 'Highland' till. The general east–west drumlin trend is generally attributed to the Highland ice coming inland from the Ayrshire coast at an early stage, but the glacial advances are difficult to distinguish in this area and the east–west trend could also be attributed to the later stage of advancing Southern Uplands ice. This is because in this area of generally low ground, later ice flow from the southern hills may have been deflected into a more westerly direction by Highland ice lying to the north. In the Eyles et al. (1949) model, the general glacial retreat was from the east to the west and south-west. Glacial drainage channels are considered to be few and inconspicuous due to the prevalence of stagnant meltwater along the ice margin.

The late Devensian ice-sheet scoured the hills south of Straiton and produced such features as striae and roche moutonnée (Institute of Geological Sciences, 1978b). They indicate local ice flow from the south-south-east or south-east, (indicating that Southern Uplands ice was dominant in this area) probably throughout the late Devensian. The Water of Girvan flows in a northerly trending U-shaped valley south of Straiton [NS 386 032].

Farther north, a crag and tail feature developed at Carclout Hill [NS 395 100], due to the resistant nature of the volcanic vent forming the hill. The feature is oriented east-south-east, which is consistent with the drumlin orientation in the vicinity. The steep end of the hill is to the west and may indicate that ice flow was from the west, although Eyles et al. (1949) interpreted the flow to be west-north-westerly, consistent with their model of ice flow dominantly from the Southern Uplands. The geological resurvey of the area in the 1920s (Ayr 45NE New Series 6” to 1 mile map) recorded striae on the north side of Carclout Hill trending 100°.

Previously measured glacial striae indicate south-westwards ice flow near Dunure [NS 263 163] and the junction of the Southern Uplands, with Highland ice on the Carrick Hills (Eyles et al., 1949). Glacial landforms including striking crag and tail features near Dunduff Castle [NS 2700 1635] are also indicative of south-west-directed ice flow.

South of the Carrick Hills, the volcanic rocks are striated [NS 2824 1436] with east-north-easterly trends, but striations are not common as much of the volcanic bedrock is altered and rubbly.

Meltwater channels

Glacial meltwater channels are relatively small and scattered in this district. They appear to be draining former ice-dammed lakes rather than forming subglacial drainage.

Channels running east–west on the hills near Craigfin appear between resistant rock ridges scoured by ice-sheets and enhanced by the glacial meltwater run-off. There are three main subparallel channels, which occur at [NS 301 059], [NS 301 061] and [NS 301 062], cutting into the rockhead. The development of these channels would accord with the retreat of the ice to the west over the hills and with the meltwaters flowing down into the Water of Girvan catchment.

A small, but distinct channel extends north-westwards from Drumore Loch [NS 339 099] at the northern margin of the sheet. This suggests that here, ice lying to the south-east blocked meltwater from flowing in that direction. Other meltwater channels may have cut through the till between drumlins but have been subsequently infilled with silt and sand.

The small channel to the north-west of Aitkenhead Toll [NS 3444 0825] is associated with glaciofluvial sand and gravel deposits. It may have drained water from wasting ice around Maybole into the Water of Girvan.

Around Straiton, east of Cloncaird Mains [NS 367 075], an overflow channel empties north into Kelsie Burn. It probably indicates that a body of ice retreated southwards up the Water of Girvan valley.

Meltwater channels around Dunduff [NS 2715 1621] appear to have drained to the north-east. The channels are commonly a few to 10 m deep and cut into the Carrick Volcanic Formation. Small drainage channels developed in the Carrick Hills, for example [NS 267 134], vary in direction probably due to remnant 'caps' of ice draining off the hills.

Meltwater overflow channels trending north-east and south to the west of Guiltree Hill [NS 353 109], may have drained stagnant ice formerly covering the hill. The overflow channel mapped as trending northwards to the west of Montgomerieston [NS 364 125] during the previous survey (Institute of Geological Sciences, 1978b), is not obvious as it is occupied by a broad alluvial flat. However, the channel extends southwards towards Ballycoach [NS 363 119] (Eyles et al., 1949, p130).

In some areas around Dailly, the streamlined topography of the till has been dissected by hollows interpreted to be glacial meltwater channels. The most impressive of these is a westwards-oriented U-shaped hollow north of Maitland [NS 2930 0285], about 10 m deep, a few tens of metres wide and about 700 m long. A series of smaller hollows are oriented westwards and southwards between Moorston [NS 2826 0200] and Whitehill [NS 2875 0148] and towards the sands and gravels of the Balcamie eskers, indicating drainage towards the glaciofluvial deposits.

Till deposits

Most glacial tills in the district are massive diamictons interpreted as lodgement tills formed beneath the ice- sheets. Lodgment till is generally a stiff, compact diamicton containing facetted rock fragments of variable size set in a sandy, silty clay matrix. The clasts comprise a variety of scattered boulders, cobbles and pebbles within the matrix. Separate till bodies could not be distinguished within the district. Eyles et al. (1949, p.127) note that in general the texture of the till matrix changes from clay to a loam or sand upwards. One explanation they give is a change from the Highland ice-sheet to Southern Uplands ice-sheet, but there is no apparent break in the deposition of the till.

Ayr–Prestwick area

The majority of the countryside east of Ayr and Prestwick comprises small rounded hills or drumlins of variable thicknesses (up to 30 m) of reddish brown lodgement till, which commonly has a sandy clay matrix containing a diverse pebble assemblage. Local sand and gravel interbeds or lenses about 0.6 m thick are recorded at depth in boreholes near Prestwick Airport (e.g. SE 1008 at [NS 3695 2614]).

The drumlins are consistently oriented from east to north-east, although a coherent drumlin shape may be lost due to erosion by subsequent drainage. Drumlin features are tens to 500 m wide and up to a few kilometres long with heights of 10–30 m (e.g. Low Wardneuk [NS 387 290]). Smaller elongate mounds and depressions in the same orientation are in places superimposed on larger features (e.g. around Brocklehill Farm [NS 3970 2375]) or occur in isolation. This till is equivalent to the Wilderness Till Formation (Table 6) described in the Irvine district (Monro, 1999) and elsewhere around the Clyde (Hall et al., 1998).

A section in Raith Burn [NS 4043 2722] reveals a pocket of grey-brown till with shell fragments, about 1 m thick, which overlies 0.8 m of gravel above a red-brown till. Since the grey-brown till appears to be local, it is interpreted as an 'exotic' block deposited on (or within) the red-brown till. Other grey till exposures containing marine shells are recorded east of Monkton [NS 371 280]. This grey till with marine shells mapped by Mykura (Geological Survey of Great Britain (Scotland), 1966a) in the Pow Burn near the A77 [NS 371 280] could not be confirmed as it is no longer exposed, but he recorded purplish sandy till overlying purplish grey till in sections up to 3.3 m thick. It was along this burn that clay with Arctic shells was recorded in the original survey (Geikie et al., 1869; Smith, 1898b).

In the former mineral railway cutting south of the River Ayr opposite Auchencruive at about [NS 390 232], Smith (1898b) recorded a few shelly fragments and also inferred an interbed of sand within the till. The original survey (Geikie et al., 1869) noted that this cutting exposed 'stiff red boulder clay (upper)', so it is not certain if the shells were contained within this till. Interbedded sand and gravel is recorded within brownish till at the east end of Auchencruive policies (Smith, 1898b).

Site investigations (SE 3445: Borehole 2 [NS 3423 1666] and Trial Pit 5 [NS 3426 1674]) along the A77, encountered stiff brown till containing local thin laminations of soft brown clay and local pockets of stiff laminated clay overlying slightly laminated till. No shells were recorded so it is not known if these are glaciomarine or glaciolacustrine clays.

Borehole records from the vicinity of Ayr and Prestwick indicate that locally the till deposit is up to 13.2 m thick (NS32SW/4), and several layers of till are separated by sand or gravel beds up to a few metres thick. This may imply phases of ice retreat, glaciofluvial deposition and renewed ice-sheet growth, or fluvial deposition beneath the ice-sheet. Whether this all took place in the late Devensian is not known.

Around Martnaham Loch area [NS 390 166] the till is commonly moulded into drumlins 0.3–1 km long and 100–500 m wide, with a relief of about 10–30 m. The drumlins are generally oriented east-north-easterly in this area with steeper, broader slopes to the east-north-east, but drumlin axes vary to trend easterly, east-south-easterly and north-easterly. Some of the mounds are complex with smaller subparallel flutes on the flanks, perhaps due to different phases of ice flow. Ice flow was generally from the west, but the north-easterly trending drumlins may have been moulded in the later stages of glaciation by ice movement from the north-east (Eyles et al., 1949). The till is greyish or reddish brown, partly depending on the colour of the underlying bedrock which has been locally incorporated. In several site investigations it was noted that the till was brownish near the surface becoming grey downwards, and in places contained lenses of sand and soft clay. In the Kincaidston area around [NS 353 192], boreholes penetrated 5.6 m to over 9 m of till. In the vicinity of Ailsa Hospital [NS 360 187], over 10 m of till was encountered, without reaching rockhead, although about 50 m to the south another borehole penetrated 7.7 m of till including 0.25 m of grey laminated silty clay just above rockhead.

Over much of the Culzean–Dipple area, a red-brown diamicton of stiff clay and pebbles overlies rockhead. This deposit is commonly formed into east-north-easterly to easterly oriented drumlins, or gently undulating ground, and is interpreted as a lodgement till. Drumlins in this area are commonly elongate, about 1 km in length and 250 m wide. Biotite granite boulders, thought to come from the Loch Doon intrusion to the south-east (Eyles et al., 1949), are occasionally found, weathered out of the till (e.g. Lady Burn [NS 223 027]). A glacial striation recorded on the second survey County Series map, east of Culzean Castle at [NS 237 102] implies a flow of ice towards the west.

East of Maybole around [NS 326 102], the diamicton becomes sandy and the drumlins are commonly oriented east–west.

Around Heads of Ayr [NS 290 180], a red-brown diamicton contains clasts up to 0.15 m including coal, sandstone, Palaeogene igneous rocks, porphyry, phyllite, green lithic arenite, Ballagan Formation, schist and biotite granite. Clast provenance indicates the presence of detritus eroded by both Highlands and Southern Uplands ice-sheets, in agreement with previous work (Eyles et al., 1949).

Around Croy, till is generally lacking over the hills formed by the resistant Lower Devonian volcanic rocks, but tends to cover the softer sandstone strata. Over the sandstones, the till is locally thin, red-brown and sandy.

On the west coast north of Croy [NS 246 136], cliffs with locally developed landslides expose over 6 m of dark brown till containing clasts of local rocks such as basaltic andesite and red-brown sandstone, as well as vein quartz, coarse-grained quartzose sandstone and greywacke sandstone. Shell fragments are sparsely distributed within the till but indicate movement of the ice-sheet onshore, and this is considered, together with the presence of clasts from the Scottish Highlands, as evidence for the north-westerly derivation of the detritus in the till. East of the hills (e.g. west of Howmoor Cottage [NS 277 120]), in the lee of rock mounds, tails of till have been deposited (for example at [NS 280 120]) by easterly moving ice flow. However, in places (e.g. [NS 294 142]), the long axes and asymmetry of small drumlins confirm ice flow toward the west-south-west.

Lodgement till is widespread around Kirkmichael, commonly in the form of drumlins 250–500 m in length. They are oriented east-south-east to easterly south of the Water of Girvan, but to the north of the river the drumlins are elongated east–west to north-easterly. The till colour varies from reddish brown to grey-brown. In places, such as near Maybole (for example [NS 297 086]), the lodgement till is thin to absent and hummocky glacial deposits rest on rockhead. Farther south, in the railway cutting [NS 3046 0570], west of Carsloe, up to 10 m of till was encountered. East of Crosshill [NS 337 065], easterly trending drumlins are formed of till in the order of 20 m thick, although some drumlins may be rock cored. Borehole evidence at Littleton Farm [NS 313 088] indicates that above rockhead, till 8.5 m thick overlies a soft sandy clay 4 m thick. At Society Street (SE 3440 [NS 3016 0979]) in Maybole, site investigation boreholes penetrated a stiff to very stiff, brown, silty, sandy till, up to 9.3 m thick, resting on brown, medium to fine-grained sandstone.

Over the more resistant igneous rocks forming Kildoon Hill [NS 295 072] and Craigfin Wood [NS 300 059], till is thin to absent. When ice flowed from west to east over these hills it appears to have deposited 'tails' which merge with the adjacent drumlins [NS 309 061]. Near Ashfield [NS 339 051] the till is also thin where the hills, formed of resistant Kinnesswood Formation, begin to rise to the south-east.

The widespread presence of pale biotite-granite erratic boulders, presumed to come from the Loch Doon pluton, was noted in the previous surveys from the Water of Girvan to beside Drumore Loch around [NS 340 097].

North of Dalrymple around [NS 360 155], the drumlin field contains drumlins in which the main orientation of the long axes is east to south-east. Drumlin long axes of north-east trend are observed locally in the area south of Dalrymple to Burnbank [NS 353 129] and these may reflect interference of the ice flow around the resistant rock-cored hills. Exposures of the till are few around Dalrymple, partly because the River Doon has a broad meander belt. The exposures of till typically show a brown weathering top to a stiff, grey lodgement till comprising a silty sandy clay matrix to a variety of stones and boulders. The thickness of till is variable but over 10 m is reported in the banks of a burn south of the River Doon [NS 389 140]. No shelly tills or interbedded sands and gravels have been observed in this area. In borehole (NS31SE/10) sunk near Barnshean Loch [NS 380 114] in the 19th century, just over 30 m of 'surface clay' was recorded above rockhead. At the Dalrymple Sewage Works (SE 4566) [NS 3536 1407]), below 13.7 m alluvium, firm to stiff, grey-brown interlaminated clay and sand with some coarse to fine, subrounded to angular gravel passes down into dark brown till. This suggests that glaciolacustrine deposits may lie under the broad belt of alluvium near Dalrymple.

Over much of the Dailly area, a red-brown diamicton of stiff clay with subrounded pebbles, or sandy clay with pebbles overlies bedrock. This deposit is commonly moulded into east-north-easterly to easterly oriented drumlins or gently undulating ground, and is interpreted as a lodgement till.

Other areas underlain by till lithologies have an undulating, rather than streamlined appearance. In some cases this could be due to complex drainage dissection of the till for example north of Moorston [NS 2820 0240], in others it may be related to mining-related subsidence, shafts or pits, for example east of Bargany Mains [NS 2515 0100].

North of the sand and gravel of the Balcamie eskers at [NS 2820 0170], an unusual, flat area, underlain by a till lithology of sandy clay with pebbles, may represent an area of till soliflucted from the surrounding drumlinoid topography.

At the Loch Spallander dam site [NS 385 083], borehole evidence indicates that till thickness varies fairly rapidly from about 3 m to over 18 m. Here the top metre proved generally soft, grey or brown sandy clay containing gravel and boulders, overlying a stiff to very stiff till which became sandy where it rested on sandstone at rockhead.

Near Straiton [NS 378 065], drumlins are developed on the lower ground over the thicker till cover north-west of the Straiton Fault. They trend east-south-easterly to easterly, in contrast to the more north-north-westerly alignment of the glacial striae on the higher ground to the south-east. This difference may relate to early and late stages of the glaciations, or a change in ice flow direction as it descended from the higher to lower ground and was influenced by transverse, more westerly flow in the lowlands. The till cover is minimal on the higher ground south of Straiton, because glacial erosion was dominant. Erratic boulders of Loch Doon granite are present, particularly west and south-west of Straiton, and indicate that the dominant ice flow in this area was from the south-east. No evidence was found around Straiton of marine shells or 'Highland' metamorphic erratic boulders.

Hummocky glacial deposits

The broad hummocky topography around Maybole, for example near Tunnoch Farm [NS 308 097], is produced by a diamicton of angular sandstone fragments set in a poorly sorted sandy matrix. These hummocky deposits appear not to have travelled far and were deposited from a stagnant remnant of the late Devensian ice-sheet. They were referred to as 'angular moraine' in the previous memoir (Eyles et al., 1949, Plate viii).

The unit is difficult to classify, and in one scenario Eyles et al. (1949) considered it to be 'boulder clay' possibly deposited by a re-advance of the ice-sheet. In another scenario, they considered it could be the result of ice wasting locally since, as they pointed out, it lacks rounded striated clasts and the sandstone debris is mostly local. In places, wispy lenses of sand and gravel were reported within the deposits, which are interpreted to indicate evidence that some water sorting is involved with the formation of the deposits. Where the deposits are less hummocky and appear to contain more sand and gravel south of Maybole [NS 302 089], a sand- and gravel-rich category has been mapped separately. North of Tunnoch [NS 308 098], hummocky glacial deposits appear to pass laterally into glaciofluvial sands and gravels, which may be associated with wash out from the wasting ice mass.

Around Maybole, there is a definite junction between the sandy lodgement till which forms smooth, gently undulating slopes and drumlins, and an overlying unit that is commonly formed into hummocky and moundy ground. These mounds, 10–100 m diameter and tens of metres high, are composed of clayey sand with scattered, poorly sorted clasts. Faint bedding and a variation in clast concentration are sometimes observed. Clasts are commonly locally-derived Devonian sandstone or igneous rock. Some exposures (e.g. north-north-east of Thornbroke [NS 2930 0883]) contain numerous angular coaly clasts. In the area north of Leffinwyne (for example at [NS 2670 0770]) the moundy ground is peppered with irregular peat or bog filled lows which are interpreted as kettleholes. In flatter areas south-west of Fordhouse, stream sections [NS 2968 0888] expose a lower unit of fairly compact, red-brown clayey sand, overlain by looser, friable, silty sand with clasts including coaly material, phyllite, baked sandstone and altered igneous rock. Within the looser sandy silt, there are lenses of clean, well-sorted gravel and coarse sand. The gravel composition indicates derivation from the local Siluro-Devonian micaceous sandstones. Some gravel lenses are laminated and faintly cross-bedded. Taken together, the angular sandstone-bearing diamictons passing up to laminated and cross-bedded gravels, are interpreted as hummocky glacial deposits passing upwards into dominantly glaciofluvial deposits.

Similar uneven and moundy ground, interpreted as hummocky glacial deposits, extends as far west as Kirkoswald [NS 245 080]. Exposures showed that the moundy ground is formed from angular pebbles and cobbles of sandstone in a sandy, clayey matrix (Sowerbutts, 2000).

Previously these deposits were mapped as 'moraine' (Institute of Geological Sciences, 1978b). Eyles et al. (1949) suggested that 'the formation of the Maybole–Straiton moraine may even have followed upon a slight north-easterly re-advance of the (Southern Uplands) ice lobe, suggested by the fact that 'two boulder clays are distinguishable over a considerable area between Maybole and Kirkoswald' (Eyles et al., 1949, p.133). Alternatively, they also suggest that the morainic material accumulated and remained on the surface of the ice, and therefore escaped attrition and striation.

The moundy deposits are here interpreted to represent very local, poorly sorted morainic deposits from a stagnant remnant of the late Devensian ice sheet, followed by washed out and water-sorted material from the same melting ice remnant (Smith, 1999b; 2000a).

Glaciofluvial sand and gravel

The main deposits of glaciofluvial sand and gravel in the district lie in a belt extending from Maybole to Straiton.

Glaciofluvial sand and gravel near Straiton is mainly found flanking the Water of Girvan and deposited on till, for example near Milton Mill [NS 375 053]. Also around Straiton, a thin sandy layer including large boulders of Loch Doon Granite overlies the till (Geikie et al., 1869). These deposits are thought to represent superficial and englacial material washed out as the ice retreated up the valley of the Girvan Water. The sand and gravel is only locally moundy, and eskers and meltwater channels are poorly developed in this area; possibly because the meltwaters seeped away or deposited sheets of sand. These flat spreads have in places been incised by later development of drainage and left terraces above the present river valleys.

In the Kirkmichael area, these deposits are mainly of an uneven to hummocky nature and considered to be ice contact deposits. A belt of glaciofluvial sand and gravel up to 500 m wide lies just to the south of Kirkmichael [NS 344 085], trending north-westwards from the Water of Girvan to near Chapelton Loch. This belt forms a conspicuous feature and appears to cut across the roughly east–west drumlin trend. It may be related to a drainage system flowing south-east on (or beneath) the remnant ice which deposited the hummocky glacial deposits around Maybole; or it may represent a drainage system flowing north-west into the River Doon catchment from the Water of Girvan, which may have been blocked to the south-west near Crosshill. The sand and gravel lies in irregular elongate mounds about 10 m high and up to 100 m in length. Narrow, but sharp, esker-like ridges are developed within the deposits near Chapelton and around Kirkmichael House [NS 342 085], where the ridge is about 800 m in length. Several former pits in the sand and gravel (e.g. at [NS 343 083]) near Kirkmichael House exposed fine- to coarse-grained, structureless gravels. In a former pit east of Broom Knowe [NS 326 097], up to 7.6 m of coarse, sandy, locally cross-bedded gravel, was recorded by the previous survey as containing many Devonian igneous rock type pebbles. At the margins of a former pit near Chapelton [NS 3258 0939], several sections in variably stratified and cross-bedded coarse sands and gravels were seen in coarsening up and channelled beds.

A moundy ridge of sand and gravel deposits lies at the margin of the hummocky glacial deposits [NS 295 111] north-west of Maybole. In a former quarry, faces are up to 5 m high and expose stratified and crossbedded brown sands, some of which are coarse to granular and include pebbles up to 0.1 m in length. The foresets dip to the north indicating palaeoflows to the north, possibly washed out as material from the stagnant ice which deposited the hummocky glacial deposits.

Another belt of sand and gravel lies in the Auchenwynd to Littleton area south of Maybole. A borehole at Littleton Farm [NS 31322 08789] penetrated 4.27 m of sand and gravel above till. Elsewhere only small patches of sand and gravel are mapped; some flanking the Water of Girvan (e.g. at [NS 326 074]) and the Dalhowan Burn [NS 330 059]. The sand and gravel is generally poorly sorted but lenses of sand at least 0.5 m thick are exposed. Small sand and gravel mounds associated with the alluvium near Crosshill (e.g. at [NS 3275 0620]), indicate that glaciofluvial drainage from Dalhowan [NS 330 059] flowed north-west towards the Water of Girvan. A mound of sand and gravel up to 140 m in length at [NS 328 057] flanks the drumlin west of Cloyntie.

Near Dalrymple, mounds and small beaded eskers of fluvioglacial sand and gravel occur on either side of the River Doon. The best example of an esker in this area is the one south-east of Netherton [NS 373 135]. Sand and gravel deposits are also mapped along the tributary valleys at Veenston Glen [NS 372 133] and near Hollybush [NS 392 148]. One isolated mound of sand and gravel lies to the east of Dalrymple Bridge [NS 359 143] and another south-east of Barbieston Bridge [NS 367 143]. A pit excavated in the centre of the elongate mound at Veenston Glen [NS 372 133] shows flaser cross-bedded, medium-grained sands dipping away from the axis of the mound. Some sand lenses contain numerous coaly fragments and gravel beds are present towards the top of the mound.

On the southern side of Martanham Loch [NS 390 167] several small mounds and elongate ridges of sand and gravel are probably associated with glacial meltwater drainage into the loch. A borehole on one mound penetrated 4.5 m of sand and gravel before reaching till. An isolated sand and gravel ridge lies on the north-west side of a drumlin at Braston [NS 362 193].

Near Culzean Bay, at Balchriston [NS 248 114], a spread of sand and gravel forms an outwash terrace or delta flowing to the north.

In the Dailly area, north and west of Balcamie (e.g. [NS 2772 0145]) sand and gravel are formed into steep-sided ridges, hollows and moundy ground. Ridges and hollows are sinous, oriented westwards and commonly about 10–20 m wide and 5–10 m high. A small excavation across one of the ridges shows it to be formed of cross-bedded, subrounded to rounded pebble gravel with poor to good sorting. These deposits are interpreted as waterlain, glaciofluvial deposits. The ridges have the characteristic features of eskers (termed here the 'Balcamie eskers'), inferring an ice-contact genesis for these sands and gravels. The deposits were probably formed during the melting and retreat of the late Devensian ice-sheet and may be linked with the meltwater channels observed in the adjacent lodgement till.

Buried channels or bedrock depressions

These are not well quantified in this district but near the coast the larger valleys are probably underlain by deeper 'channels' or rock tunnels formed by earlier glacial erosion. For instance, a borehole on the esplanade near Doonfoot encountered at least 20 m of alluvium (i.e. down to about 19 m below Ordnance Datum).

The present course of the Water of Girvan was probably the approximate preglacial course, but there may be associated buried 'channels' in the vicinity that cut more directly north-west to the coast. South-east of Maybole, some borehole data indicates a general north-westerly trending buried 'channel' within the subcrop of the Swanshaw Sandstone Formation shown by north-eastward fall in rockhead depth from Kilhenzie Castle [NS 308 082], where rockhead is exposed at 73 m above Ordnance Datum (OD), to Littleton Farm where it is down to about 41 m above OD and further north-east to, near Heart Loch, where rockhead is at about 50 m above OD.

Raised beaches (late Devensian)

Broad coastal terrace features can be traced along the peri-urban eastern fringes of Ayr and Prestwick extending inland up to 4 km from the present coastline. The highest feature is interpreted as the top of a late glacial raised beach deposit (Geological Survey of Great Britain, 1966a and b) that cuts into the till deposits to form a flat to gently sloping bench comprised of sand and gravel. This feature can be traced at 25–30 m above OD and in places there is a back feature indicating the former coastline, although this is not distinct because the till into which it is incised is relatively soft. In some areas, several benches can be identified, for example at Thornyflat Farm [NS 3690 2285], and in others the lowest part of the terrace becomes dissected by drainage channels. Red, parallel-bedded sands, gravels and thin pebbly beds of the highest raised beach deposit are exposed in Ladykirk Burn at [NS 3803 2606] and [NS 3824 2622]. These sediments represent relatively high energy coastal deposits.

Raised beach deposits of late Devensian age may include marine laminated silty clays at depth in the former estuaries of the Ayr and Doon rivers (cf. the River Irvine, Boyd, 1986a).

Little or no clay has yet been recognised at levels comparable with those of the beach deposits. However, in a former section near Prestwick Airport at [NS 3736 2679], in the third (uppermost) raised beach, 2 m of brown sandy clay with gravelly lenses overlay 0.3 m of grey plastic clay with silt laminae, above a grey, poorly bedded plastic clay on till. No shells were recorded from these deposits. Also near the airport, evidence from SE 1008 [NS 3697 2618] indicates that dark brown, plastic sandy clay forms lenses up to 2 m thick within and below sand and gravel forming the third raised beach. Eyles et al. (1949, p.135) described a section in Pow Burn [NS 372 277] through the deposits of the uppermost raised beach as comprising coarse gravel (1.3 m) on laminated sand with pebbles and boulders (1.3 m) on dark brown, laminated sandy clay (0.9–1.8 m+).

In the vicinity of Belmont, Ayr, boreholes (e.g. at [NS 3412 2034]) in this raised beach encountered loose to soft sand with clay layers down to 8.45 m, on till. At Lochside, north of the River Ayr, boreholes (e.g. SE 4552; Borehole 12, [NS 3448 2249]) on the uppermost beach proved about 1–3 m sand and gravel directly on rockhead, suggesting a rock platform or reef on the late glacial beach.

A borehole (NS31NW/2), [NS 3055 1846]) at the holiday village south of Ayr penetrated sandy clay to a depth of 1.22 m on the uppermost raised beach resting above till.

Just inland of Culzean Bay, remnant patches of the raised beach at 20–30 m [NS 247 119] and [NS 246 143] are present but no littoral deposits are exposed. The level of this beach above sea level suggests that its age is late Devensian.

Around Turnberry [NS 210 070], a second (intermediate) raised beach platform is observed lying at about 15 m above OD and is considered to be late Devensian. The lowest beach (Holocene) here is probably covered by blown sand.

Late Devensian to Holocene age glaciomarine to glaciolacustrine deposits

East of the industrial complex at Dipple, a pond indicates the former presence of a clay pit [NS 204 027] which was used to supply Dipple Tile Works. No exposures of this previously worked clay are currently present, but Robertson and Crosskey (1874) described the outcrop in the pit. In summary, a succession of clays are present (surface level about 10–15 m above OD). The upper 2–3 m of clay below the surface contains fossils such as Cypridae, which are a freshwater ostracod family. At depths of 3–4 m, grey clay contains predominantly brackish (tidal) fossil ostracods such as Leptocythere castanea (Sars) together with marine gastropods Utriculus obtusus (Montagu) and Omalogyra atomus (Philippi). At 4–5 m depth, the clays contain mainly marine forms including Mytilus edulis Linnaeus, Mya truncata Linnaeus and marine gastropods with few fresh or brackish water types (Robertson and Crosskey, 1874). Besides shelly fossils and ostracods, a few foraminifera were recorded (Robertson and Crosskey, 1874). Even the uppermost clays contained almost fully marine foraminifera (e.g Quinqueloculina seminulum), which were possibly deposited during storms, but none of the fossils are of 'a very Arctic character' (Smith, 1898b, p.6).

This isolated patch of clay is tentatively interpreted as the deposit of a quiet marine embayment between two topographically higher drumlins, probably at the time that the highest raised beach was deposited in the late Devensian. During the late Devensian, as the land rose and relative sea level fell, the area would have become progressively brackish and then nonmarine (lacustrine).

Late Devensian to Holocene raised beach deposits

In several places along the coast (e.g. Turnberry), three levels of raised beach deposits can be observed. These are approximately 7, 15 and 25 m above OD, and for convenience are here labelled as 1, 2 and 3 respectively (without implying that only the three beach levels exist, as accurate altimetric levelling has not been undertaken). According to Smith (1997, fig. 12.8), the Holocene main postglacial shoreline has been modelled at about 10 m above OD in this area. These flat to gently sloping areas of ground are separated by breaks in slope. In some places they are covered and obscured by blown sand, for example south of Dipple factory [NS 199 023]. The deposits commonly overlie platforms of till and back features to the raised beaches are usually in till. The raised beach deposits are composed of sand and gravel (e.g. [NS 219 083]) with rare patches of clay (e.g. [NS 207 025]). The deposits are commonly only 1–2 m thick but reach 20 m in and around Ayr. These sediments overlie bedrock or up to 10 m of till (Sowerbutts and Smith, 1999). They are interpreted to have formed between the late Devensian (raised beach 3, about 25 m above OD) and the Holocene (raised beach 1, about 7 m above OD) (Plate 23) as relative sea level fell due to isostatic rebound of the land after glaciation.

In the first (lowest) raised beach at Doonfoot [NS 3217 1927], a temporary excavation in well sorted sand was at least 1 m deep. Sections recorded (Geological Survey of Great Britain (Scotland), 1966b) at [NS 352 275] and [NS 352 274] in the first (lowest) raised beach indicated the presence of purple or grey clay beds up to 0.6 m thick, within (and below) beach sands. Site investigation boreholes in Ayr (e.g. SE 4551 Borehole 3 [NS 3438 2172]) indicate that the lowest raised beach comprises loose to medium dense sand with gravel; this deposit may overlie earlier alluvium deposited from the River Ayr, as grey silty sand was encountered at a depth of 4.7 m above a layer of peat (over 0.2 m thick) at a depth of 5.8 m.

Evidence from the Irvine River, to the north of the district, indicates that around the beginning of the Holocene sea level lay at around 2 m above OD, and that around 5000 BP it had risen to about 10–13 m above OD (Boyd, 1986b) before retreating again.

Along rocky parts of the coastline such as Greenan Castle and Deil's Dyke (Plate 23) or Barwhin Point to Culzean, small patches of raised beach are found amongst the rock outcrops (e.g. at [NS 229 100]), where raised sea cliffs and caves can be observed.

Just south of Balkenna Bridge [NS 2018 0455], raised beach pebble–cobble gravels appear to be cut by sand-filled channels. These channels probably link with the hollows interpreted as drainage channels approximately a hundred metres to the east, and may represent older drainage systems cutting across the third raised beach terrace.

Flat and gently sloping coastal terraces, sometimes bounded by steep rocky cliffs, can be seen at three levels along the coast. The terraces are most clearly defined south-west of Dunure and occur at approximately 25 m (80 ft), 15 m (40 ft) and 6 m (20 ft) OD. By correlation with the terraces observed around Ayr (Sowerbutts and Smith, 1999) the 25 m and 15 m terraces are probably late Devensian, and that around 6 m is probably Holocene. Associated features include uplifted sea stacks within the terraces and raised sea caves (e.g. in the south-west part of Bracken Bay). Rare exposures show the coastal terraces to be composed of rounded, clast-supported, pebble–boulder gravels (e.g. at [NS 2937 1853]), and better sorted sands (e.g. Carwinshoch Burn [NS 2898 1831]).

South of Dunure, narrow raised platforms along the coast lie at 3–12 m above OD generally, with the foot of a cliff as a back feature. The platform is usually covered by beach sand and gravel. The slightly lower wave cut rock platform is covered at high tide. Adjacent to the more resistant Carrick Volcanic Formation, the raised beach platform is less well developed and includes former sea stacks at Katie Gray's Rocks [NS 245 139]. The raised beaches are more continuous, about 40–100 m wide, in the softer Swanshaw Sandstone Formation sandstones, and less well developed again where the Isle Port Conglomerate Member meets the coast. Locally, between Rancleugh and Goatsgreen cottages [NS 246 118] and [NS 244 111], blown sand and alluvial cones have accumulated on the raised beach. The age of the main raised beach has been discussed (Monro, 1999) and is considered to be Holocene, as isostatic recovery of the land led to a lowering of sea level and no till is found on the terraces or rock platforms.

Holocene

In the milder climate, which prevailed from sometime after 11 600 BP, alluvium was deposited along watercourses and in hollows. Temperate vegetation became established but no substantial areas of lowland peat accumulated in the area. However, in the course of draining hollows near the southern boundary of Maybole Parish in the 19th century, below about 3 m of soil and peat, biogenic marl beds were encountered which contained the head of Cervus megaceros Hart (elk), the head of a variety of Bos taurus Linnaeus, horns of Cervus elephas Linnaeus and shells including Limnea and Planorbis (sic, in The New Statistical Account of Scotland, 1845).

Along the coast, isostatic rebound continued to create a raised beach terrace. Freeze–thaw processes promoted local land slides and development of talus against rock cliffs. Made ground, cuttings and fill are the result of man's activity.

Alluvium

Alluvial deposits are very variable and not always extensively developed near former or existing watercourses or hollows. Layers of peat may be intercalated or buried within soft silt and clay, or sand and gravel. In postglacial times alluvium has infilled several of the hollows between drumlins (e.g. at [NS 393 156]). Many of the flat spreads of alluvium (hundreds of metres wide) appear too large for the present watercourse and probably overlie late Devensian or early Holocene lake deposits. In the low ground, across the central part of the district, belts of alluvium trend north-easterly to easterly along depressions between drumlins, and so partly reflect the earlier ice flow. Commonly, exposures of brown silt and clay lie above interbeds and lenses of sand and gravel occurring in sections 1–3 m thick.

A belt of sandy alluvium, 100–500 m wide lies along the floodplain of the River Ayr, and smaller flats lie along the meandering Water of Coyle (e.g. [NS 400 211]). Locally the alluvium is cobble- or pebble-rich and terraced. The terraces are, however, fairly local, and for clarity only the larger ones are shown on the 1:50 000 Superficial Deposits map. Linear or curvilinear bars can be occasionally observed on flat terraces, for example south of Auchincruive House [NS 387 232]. Some boggy areas of alluvium exist in hollows between drumlins. A broad area of alluvial sand and silt south of Mossblown [NS 402 245] has been partly covered by made ground. Relevant borehole records show that alluvial sediments can reach up to 4 m in thickness.

South of Kirkmichael, flat spreads of brown silt, sand and gravel are found along the Water of Girvan and its tributaries such as the Barlewan Burn around [NS 318 077]. These spreads reach up to 700 m wide west of Crosshill around [NS 317 066] and are considered to cover or be part of ice-dammed lakes that were ponded up as the main ice-sheet retreated westwards (Eyles et al., 1949). The deposits are variable, ranging from gravel to clay with peat layers. In this area, the course of the Water of Girvan appears to be partly controlled by the solid geology in that it meanders across the 'softer' lower Carboniferous rocks (and keeps south of the Kerse Loch Fault). This was probably its approximate preglacial course, but there may be associated buried 'channels' in the vicinity.

Patches of alluvium near streams south-west of Maybole are commonly composed of pebbly gravel (e.g. [NS 2605 0706]). These commonly reach a few metres in thickness. Other isolated patches of alluvium tend to be sandy, silty or peaty, and probably accumulated in hollows formed after the late Devensian ice wasted in areas such as Heart Loch [NS 311 094] and Chapelton Loch [NS 322 097]. Some low, flat areas of alluvium with little or no associated modern drainage (e.g. south-east Cultezeoun [NS 285 094]), might relate to glaciofluvial meltwater deposition.

Boreholes that penetrated below made ground at a site [NS 301 093] south of Maybole, encountered patches of very soft peat up to 0.9 m thick above 0.2 m sand and gravel lenses, on over 8 m (unbottomed) of soft to very soft, brown, slightly clayey, sandy silt.

Near its mouth the River Doon is flanked to the west by a Holocene raised beach, and formerly it appears to have taken a course farther north, near Seafield [NS 328 202]. Whether this change was natural or due to artificial diversion of its course is debatable (The New Statistical Account of Scotland, 1845). The development of alluvial spreads along the River Doon is variable, at the New Bridge of Doon [NS 3315 1795] it is confined by bedrock to the river bed, but 500 m to the east the meander belt is about 250 m wide.

The broad spread of clayey alluvium along the Doon valley in the vicinity of Dalrymple is up to 650 m wide, and surrounds a till drumlin [NS 356 144] on which Dalrymple church is built. Boreholes (SE 4566) at the sewage works encountered 15.45 m of alluvial sand, gravel and cobbles above till. Where the meander belt of the River Doon is narrower, near Skeldon House [NS 375 136] and Kennedy's Dub [NS 398 138], alluvium rests directly on rockhead. Other watercourses such as Primpton Burn [NS 377 147] and Dryrock Burn [NS 388 108] are flanked by alluvial belts of varying widths, mainly comprising silt, sand and gravel. Isolated hollows with flat alluvial deposits, (e.g. at [NS 352 118]), may contain glaciolacustrine sediment at depth deposited by glacial meltwaters. These deposits tend to be brown sandy silts or clays.

West of Loch Spallander (e.g. north of Cairnhill [NS 369 097]) and north of Orchard Farm [NS 355 097], the larger spreads may also cover, or be part of, ice-dammed lakes formed during late Devensian times.

Farther south-east, in the Straiton area, alluvial deposits mainly comprise a mixture of silt, sand and gravel. In places, minor lenses of clay and/or peat are present. An extensive belt is present along the Water of Girvan at [NS 386 015], which appears to infill a U-shaped valley developed by ice flow from the south. North of Blairquhan [NS 364 059], the old river channels and small ox bow lakes are still evident in the alluvial belt of the Water of Girvan.

Near Dailly, alluvial, red-brown sand is extensive along the Water of Girvan. One or two metres of sand are exposed at some localities and the total thickness of the alluvium could be substantially greater. A striking oxbow lake lies within the alluvium east of Dalquaharran Castle [NS 2735 0220].

River terrace deposits

River terrace deposits occur where older alluvial spreads are cut through or bypassed by more recent alluvium. Generally terraces are not extensive in this area, either due to common reworking of the alluvial belt by the river or the building of levees to prevent flooding.

Beside the River Ayr, limited terraces are preserved near Tarholm at [NS 395 225] and beside the River Doon at, for example [NS 341 167]. Near Dalrymple, small areas of river terrace can be recognised south of the River Doon [NS 371 139] and south-east of Dryrock Burn [NS 379 102]. No sections were exposed in the deposits but they are thought to comprise silts, sands and gravels similar to the adjacent alluvium.

Alluvial terraces developed along the Water of Girvan are larger but variable in extent owing to common reworking of the alluvial belt during the Holocene. Some of the largest terraces lie to the north of Blairquhan [NS 364 062] and are 5 m or less above the present floodplain. The terraces are mainly composed of sand and gravel with some silty beds. South of Aitkenhead Toll [NS 345 079], two levels of alluvial terrace occur along the Water of Girvan above the present alluvial flat. The lower terrace lies at about 49 m above OD and the upper terrace is about 52 m above OD, although it is obscured by the development of the former creamery.

Near Lindsayston Farm [NS 2860 0093] and east of Dailly village [NS 2745 0108], river terrace deposits occur along the Lindsayston Burn.

Alluvial fan deposits

These deposits form cones where active, commonly seasonal tributaries, deposited their sediment load onto the floodplain, for example near the Water of Girvan at [NS 370 064] and [NS 392 008]. Alluvial fan deposits mainly comprise poorly sorted sand and gravel with some silt and clay. Small alluvial fans, mainly comprising sandy gravel, are present, for example near Barlewan Burn [NS 314 076] and south-east of Longhill at [NS 346 052]. The alluvial fan deposits south-west of Dalrymple [NS 354 135] comprise variable silts, sands and gravels which merge north-westwards into the alluvial deposits of the River Doon.

Two lobes of alluvial fan material are found adjacent to Dalchomie Farm [NS 2787 0865] and pass laterally into angular boulders and pebble head deposit on the adjacent steep slope. Alluvial cones have built up over the raised beach at the mouth of the Rancleugh Burn [NS 246 117] and further south at Goatsgreen Cottage [NS 244 111].

In the south of the district sand- and gravel-dominated alluvial fans were observed north of Quarrelhill Burn [NS 2535 0242] and near Lindsayston Farm [NS 2846 0075].

Lacustrine deposits

In the Dalrymple area, small flattish areas of lacustrine deposits occur adjacent to Loch Croot [NS 378 119], Barnshean [NS 377 112] and Shankston [NS 392 116] lochs. These variable deposits comprise interlaminated sand, gravel and silt with peaty layers. Well-bedded lacustrine sands and sandy clays are also found surrounding parts of Mochrum [NS 266 091] and Craigdow [NS 261 064] lochs. These deposits are all considered to be Holocene in age.

Peat

Peat, or partially decomposed vegetation, accumulated during the Holocene, mainly in the hilly areas in isolated poorly drained hollows. It is usually mapped where it is about 1 m thick, but a thin cover may be found on remote hill slopes.

Peat may have once been more extensive over poorly drained low ground around Ayr, such as alluvial flats and raised beaches, as remnants up to 1.6 m thick were encountered on the upper (third) raised beach to the west of Ayr race course (SE 4553: Borehole 1 [NS 3456 2218]). This peat is now covered by fill.

On the higher ground in the Carrick Hills, peat is sparse above the Carrick Volcanic Formation, probably because the acid water associated with the development of peat is neutralised by the weathered alkaline volcanic rocks. More commonly, peaty alluvium and peat infill the low ground between hills, such as on Red Moss [NS 271 129], and the bases of glacial meltwater channels (e.g. at [NS 2740 1587]).

In contrast, on the more acidic dacite intrusion of Knockskae and Back Fell, peat has accumulated on gentle slopes [NS 351 018] as well as in hollows [NS 355 010] and [NS 363 011]. Farther west, in poorly drained hollows over till, broader accumulations of peat occur at, for example, Barony Moss [NS 321 010].

Patches of boggy peaty ground are developed on the poorly drained parts of the hummocky glacial deposits south-west of Maybole. Generally the peat (e.g. east of Glenluie) is quite thin and is probably interbedded with, or passes laterally into, alluvial deposits. Peat deposits in hollows in till around Craigdow Loch are more extensive and probably thicker (e.g. [NS 260 056]). South-east of Maybole, 0.9 m peat was encountered in a site investigation at [NS 301 093] although it proved to be an impersistent lens. Thicker patches of peat occur south of Maybole associated with the spreads of alluvium, such as around Heart [NS 311 094] and Chapelton [NS 322 097] lochs. Records indicate that 2.4 to 3 m of soil and peat locally overlie white or pale grey biogenic marl containing organic remains, including the head of a giant elk (see above).

Blown sand

A belt of blown sand, up to 1.5 km wide, stretches from the high tide mark inland around Prestwick [NS 350 260], encroaching on the Holocene and late Devensian raised beach deposits. The loose sand is locally cross-bedded and may contain frosted sand grains, as well as organic and shell fragments. Sand dunes are evident on the golf courses next to the shore at Prestwick and Ayr, and they appear to have an overall shore-parallel trend. Dunes are essentially stabilised inland, but can be identified due to short wavelength and locally large undulations in built-up areas. Their distribution does not appear to have changed much in the 140 years since they were first mapped by the BGS.

A borehole near the mouth of the River Doon penetrated 2 m of fine- and medium-grained, light brown blown sand.

Farther south, above the high water mark near Turnberry at [NS 200 057], a strip of dunes on top of the Holocene (No. 1) raised beach is common, and in places the dunes can reach 5–10 m in height and extend inland for up to half a kilometre, encroaching onto raised beach No. 2.

A narrow strip of blown sand covers the raised beach feature (at about 10 m above OD) on the coast west of Balchriston [NS 2447 1140]. It appears to have blown onshore from the recent sand beach but, judging from survey maps, it has not developed any further in over one hundred years.

Landslides

East of Ayr, indistinct landslides with slight backscars and uneven/mounded toes can be identified on several steep slopes (e.g. north of Bridgend Mains [NS 3975 2140]).

Small areas of landslide have been mapped near Croy along parts of the cliffs which lie at the back of the Holocene (No. 1) raised beach. The slide at [NS 243 109] includes large blocks of sandstone from the cliff above, but to the north [NS 246 135] the slides are entirely within till.

Talus

At the base of exposed rock faces such as Corbie Crag [NS 389 018], and south of Cawin Hill [NS 372 027], areas of talus or scree have accumulated, generally as a result of rock falls due to freeze-thaw processes. These form unstable, loosely clast-supported aprons of angular rock fragments, together with blocks of local rock types.

Marine beach and tidal flat deposits

The nature of the beaches along the Ayrshire coast is closely related to the bedrock geology. The igneous rocks and locally massive sandstone beds produce rugged shorelines backed by cliffs, whereas the softer Carboniferous mudstones and Coal Measures form gently shelving coastlands.

Beach or shoreface deposits along the coastline vary from long, narrow (about 100 m) expanses of golden sand (e.g. north of Turnberry [NS 199 050]) to gravel and sandy flats in amongst rocky outcrops (e.g. Dowhill Port, north of Dipple factory [NS 200 031]). At Maidens, the harbour area [NS 211 080] contains muddy, tidally influenced beach and tidal flat deposits.

Near Prestwick the beach is extensive, relatively wide (about 250 m) and gently shoaling. Coal and industrial waste constitute a proportion of the beach sediment, although most of the modern beach sand seems to be derived from erosion of the lower part of the raised beach deposits (Deegan et al., 1973).

Offshore superficial deposits

Information about the Quaternary deposits offshore is limited and mainly comes from surveys made by the Institute of Geological Sciences (British Geological Survey) in the 1970s. The work culminated with the publication of the Clyde Sheet sea bed sediments and Quaternary geology map at 1:250 000 scale (British Geological Survey, 1985b). The sea floor off Ayr generally shelves gradually to a depth of about 70 m below OD at the western limit of the sheet.

The Quaternary deposits (Figure 23) below the recent sea bed sediment are generally thin (0–30 m thick), and are only over 40 m thick in narrow channels at the western edge of the district. Silty clays up to about 30 m thick form the bulk of the Quaternary deposits. These silty clays are referred to the Jura Formation (Davies et al., 1984), which is assumed to cover much of the sea floor (except for tidally scoured areas of bedrock). In parts of the inner shelf off the west of Scotland, the Jura Formation overlies the late Devensian Barra Formation (Davies et al., 1984), but the latter has not been identified in this district. Offshore boreholes put down into bedrock proved that the clays overlie an impersistent layer of brown to greyish brown sandy till or diamicton, up to about 10 m thick. This may correlate with the late Devensian Hebrides Formation (Davies et al., 1984).

Jura Formation

This formation is acoustically well stratified with close, parallel horizontal to inclined reflectors on seismic profiles (Davies et al., 1984). This layering does not appear to relate to lithological changes, because drilled offshore boreholes record the sediment type as fairly homogeneous clay or silty clay (Figure 23). It is thought that the seismic reflectors relate to changes in water content, compaction or other physical properties (British Geological Survey, 1985b). Lack of resolution makes the layering difficult to identify in sequences thinner than 20 m which is the basis for the line within the formation on the inset map of offshore superficial sediment.

The formation is considered to be late Devensian to Holocene in age, and partly correlates with the 'Clyde Beds' (Clyde Clay Formation) onshore (Davies et al., 1984). The very soft to firm silty clays contain scattered gravel, probably representing dropstones. Lateral changes in seismic facies, localised erosive horizons that are attributed to tidal scour, and multiple infill of small basins, are stated to illustrate the complex nature of this formation (Davies et al., 1984). The lack of resolution in the seismic profiles prevents mapping the subsurface Quaternary units, but details of borehole drilled through them are given below.

In offshore borehole 70/15 (or 55-05/91 at Lat. 55° 25.06' N Long. 4° 49.50'), the topmost 17 m of silty clay is calcareous, grey-brown with some fine sand. The underlying poorly sorted, silty, sandy clay with pebbles and cobbles and some shell material, was described as possibly 'boulder clay' (Chesher et al., 1972). Borehole 72/02 (or 55-05/99 at Lat. 55° 30.01' N Long. 4° 48.33') penetrated 22 m of dark brown clay, passing down into dark grey clay with black organic material (British Geological Survey, 1985b). The clay deposit is taken to be the Jura Formation overlying 8 m of brown and greyish brown sandy till on bedrock. If the Jura Formation partly equates to the Clyde Clay Formation, then the underlying till may correlate with the Wilderness Till Formation onshore.

Nearshore boreholes put down to check the coal mining off Prestwick (Lawrie and Anderson, 1953) found that Prestwick Bay was a rock-basin in late glacial times, probably scoured by ice. During the warm period (Windermere Interstadial) after the main glaciation, the rock floor was covered by clay brought in by a river (a precursor of the Pow Burn) entering from the north-east. The clay contains a rich fauna of ostracods (e.g. Cythere lutea Muller, Cytheropteron latissimum (Norman) and Cytheridea papillosa Bosquet) and foraminifera (e.g. Quinqueloculina seminulum (Linnaeus), Protelphidium orbiculare (Bardy) and Pyrgo depressa (d'Orbigny)). Most of these species (a longer list is held with BGS Edinburgh Palaeontology department) still live in the North Atlantic but some are definitely of Arctic provenance. Later, probably in the Holocene (after the Loch Lomond Stadial), sand was deposited above the clay and it contains an ostracod and foraminifera assemblage without Arctic species and some species of a more temperate habit. It was considered by Anderson (Lawrie and Anderson, 1953) that later in the Holocene the sand was elevated above sea level and blown into dune-fields by the south-west winds. Subsequent submergence converted the bay back into a sea-filled basin.

Recent sea-bed sediments

Surveys of the recent sea-bed sediments showed that they fined from complex areas of sand and gravel in Ayr Bay to a blanket of mud about 8 km westwards from the shore (British Geological Survey, 1985b). The sea-bed sediments occur as a thin layer above the Jura Formation.

The distribution and type of offshore surface deposits in the district vary from gravel and sand to muds, ranging from coarse littoral through transitional facies, to deep silty clay facies (Deegan et al., 1973). Gravelly sediments were found only in Ayr Bay close to sea-bed rock exposures. The gravel matrix is a well sorted, lithic, very coarse to granule-grade sand, in places, containing, large proportions of the algae Lithothamnium or shells. In the more widespread littoral sands, local variation was noted; for instance near the Heads of Ayr, a high percentage of the clasts are mafic minerals of which up to 5 per cent are magnetic lithic types and magnetite. In general, the sands contain quartz together with some pink orthoclase and garnet, as well as shell, coal and slag debris. In the deeper waters the deep silty clay facies is characterised by cohesive silty clays with generally less than 1 per cent sand content. Even in this 1 per cent sand portion the detrital quartz fraction is small, as skeletal carbonate and indeterminate organic detritus make up the bulk of the sand portion.

Artifically modified ground

Human modification of the landscape is widespread in the coalfields and built-up areas of the district. Substantial areas of made ground, worked ground and infilled ground are shown on the Superficial Deposits and Simplified Bedrock edition of the Ayr Sheet. However, such ground is not always capable of being mapped at 1:50 000 scale, and on many developed sites ground has been landscaped, excavated or covered by rock or earth. Landscaped ground is common in urban areas and also parks, gardens and golf courses. It is not practical to delineate landscaped ground at the 1:50 000 scale. The disturbed ground shown on the 1:50 000 map between Auchalton [NS 335 035] and Balgreggan [NS 348 048] are areas in which excavation of limestone and made ground are closely associated.

Made ground

North and east of Ayr, most made ground shown is due to bings (spoil heaps) of coal waste from the former Auchencruive [NS 385 249], Enterkine [NS 406 236] and Tofts [NS 400 264] deep mines. Many of these mounds have since been landscaped but rise up to 25 m from the original ground surface, for example, that east of Mossblown [NS 407 247]. Embankments of made ground are shown along built–up parts of railways, roads and as levees to the Water of Girvan. Made ground is also mapped at coastal defences and reservoir dams and is widespread at Prestwick Airport. Site investigations near the River Ayr in Ayr town centre indicate areas of made ground (e.g. at [NS 341 217]), and many sites too small to show on 1:50 000 scale are covered by 1–3 m of fill, ranging from sand, gravel and clay to ash and masonary rubble.

Around the Dailly Coalfield, made ground is commonly in the form of bings, many of which have been landscaped and forested (e.g. around Dalquaharran Mains [NS 2740 0290]).

Worked ground and infilled ground

Worked ground includes former quarries, excavations for constructions and cuttings for road and rail.

The former opencast site at Glenshalloch Wood [NS 2965 0497] is an example of worked ground, with infilled ground in a quarry nearby [NS 2975 0478].

Infilled gound is shown where former quarries and opencast excavations have been wholly or partially filled. The fill may either be natural materials or rubbish (e.g. Holmston quarry [NS 3640 2085]).

Disturbed ground

Disturbed ground relating to mining (and probably subsidence, since coal has been worked in the Dailly area since the 15th century) at coal crops, shafts and adits is also observed at some localities (e.g. Burning Hills [NS 2780 0325]) and near Bargany Mains [NS 2490 0100]).

South-west of Blairquhan, three areas of disturbed ground are shown in the 1:50 000 Superficial Deposits map at [NS 347 049], [NS 343 046] and [NS 635 035]. These are areas of complex quarrying together with collapsed and dumped waste material from the working of pedogenic limestone beds within the Kinnesswood Formation, which was partly overlain by till.

Oval depressions several tens of metres across around Mount Stairs [NS 3920 2275] have grown since the 1960s on evidence from aerial photography. These may be due to the collapse of old mine workings (Sowerbutts and Smith, 1999).

Contaminated land

The monitoring and remediation of contaminated land is an undertaking of local government. Some information is found in the maps and records of historical mineral and industrial workings held by the BGS. These include some landfill areas. As of 2005, a landfill in Maybole was being monitored by South Ayrshire Council for gas migration and groundwater pollution. Former gas works in South Ayrshire such as the one at South Harbour, Ayr have also been assessed.

In the former coalfields around Dailly, shallow (less than 30 m deep) mining, shafts and adits should be considered a possibility in any development.

Chapter 12 Economic geology

In the past, the coalfield areas of Ayrshire were the main targets for economic minerals. Details of the economic geology and its history for the Ayr and Prestwick area up to the early part of the 20th century are given in Eyles et al. (1930) and those for the Dailly to Patna area in Simpson and MacGregor (1932).

Energy resources

As with much of central Scotland, Carboniferous coal exists in the Namurian Limestone Coal Formation and the Westphalian Scottish Coal Measures Group.

In former times the Ayr district was a coal producer, working Westphalian coals from the Ayr coalfield, and working Namurian coals in the Dailly coalfield. However, no coal is presently worked by mining or by opencasting, and most of the accessible seams are worked out.

Westphalian coals

Coal was mined from the Coal Measures in the Ayr–Prestwick area for hundreds of years until the large collieries (e.g. Auchincruive [NS 387 248], Enterkine [NS 410 238]; (Figure 24)) closed in the late 1960s and early 1970s. The earliest mines were small, shallow pits and adits, dug from the surface, up until the 1900s. Deeper, shortwall and stoop and room workings in the early 20th century were followed by modern deep and extensive longwall mines. Underground mining extended offshore from Prestwick for about 2 km westwards and potential resources remain offshore.

Economic coals were only found in the Middle Coal Measures; the Upper and Lower Coal measures contain no economic coals in the Ayr area, nor does the Limestone Coal Formation. The main target was the Ayr Hard (or Splint) Coal, which was about 1 to 1.5 m thick commonly in two leaves. Significant overlying seams were the Ayr Jewel (average 0.48 m, but usually too thin to work), Ayr Diamond (0.6 m), Ayr Major, Three Dirt or Soft (1.6 m in 4 leaves), Crawfordston (1.65 m) and Ayr Ell (0.64 m) coals. The latter, although a good housecoal, was too thin to work in this district. All the worked seams attained their maximum development in the Annbank area on the eastern edge of the sheet. Underground mining was never easy in the Ayr coalfield, as the strata are offset by numerous faults and intruded by sills and dykes. Locally some coal seams have been burnt or coked by intrusive sills and in places are associated with white trap (Mykura, 1965). This obviously affects the quality and rank of the nearby coals.

The shallower pits closer to Ayr (Figure 24), such as Kerr [NS 356 231] and Auchincruive Holm [NS 385 229], have not been worked since about 1869 (Eyles et al., 1930), and the Ayr Hard (about 1.2 m thick or less) is generally considered to have been worked out in that area by 1930.

The colliery [NS 365 259] at Auchencruive Nos 4 and 5 (known as Glenburn) was sunk in 1912 and finally closed in 1973. Much of its annual production of 256 000 tons came from below the sea, mining the Ayr Hard Coal which occurred at a depth of 110–180 m below sea level and dipped 2–3° inland (Lawrie and Anderson, 1953). The larger colliery [NS 387 248] at Auchencruive Nos. 1, 2 and 3 near Mossblown closed in 1960. Coal was exported mainly to Ireland via the harbour at Ayr. Smaller pits were sunk at Tofts (west of Bennals [NS 400 264]), Drumley [NS 407 249] and Annbank [NS 406 229]. The one at Drumley (Ayr Nos. 1 & 2) opened in the 1890s. It worked four seams including the Ayr Diamond Coal by longwall methods. In 1898, ignition of a rare outburst of firedamp (methane) in the Ayr Diamond Coal caused the death of seven miners. Little damage was done to the workings but the pit had ceased production by nationalisation in 1947. It became a ventilation shaft and escape route from Enterkine Nos 9 and 10 ([NS 410 238] closed 1959).

Namurian coals

In the Dailly area, the Limestone Coal Formation lies in a syncline which is cut off to the north-west by the Kerse Loch Fault and extends south-westwards onto Sheet 8W (Carrick). Records show that coal has been extracted in the area from at least since 1415 (Simpson and MacGregor, 1932), and mining continued at depth until 1967. The area was attractive to miners because although dips on the syncline margin can be quite steep, there are several thick coal seams occurring close together at outcrop and continuing to depths of a few hundred metres. The main named seams mined were the Hartley (Dailly) or Rotten (about 2 m), Craigie (Dailly) (1.22 m in leaves), Corral (Dailly) (1.37 m in leaves), Parrot (Dailly) (about 2 m), Ell (Dailly) or Ground (about 0.9 m) and Main (Dailly) (over 2 m) (Simpson and MacGregor, 1932; Mykura, 1967). They were of variable thickness and quality and commonly split into leaves, but in the section at Maxwell Pit, from the Hartley Coal to the thin coal above the Main, coal constituted 31 per cent of the total thickness of strata. North of Dailly, coals come to crop at the 'Burning Hills' [NS 2780 0325], which were on fire for many years (after a pit collapsed over an underground furnace in 1849). Numerous disused pits and shafts occur close to the coal crops. Earlier named collieries were Meadowmuir [NS 2515 0127], Kilgrammie [NS 2572 0192], Holehouse [NS 2643 0273], Drumochreen [NS 2801 0358] and Dalzellowlie [NS 2958 0444]. Coal was latterly mined from the Killochan Colliery [NS 2475 0143] shafts, Maxwell [NS 2747 0279] and Dalquharran [NS 2665 0182] adits. The Dailly area was examined for open-cast extraction, but apart from the Glenshalloch site [NS 2965 0497] which was active during the late 1960s or early 1970s, the area was found uneconomic.

The legacy of mining near Dailly has left potential hazards for land use due to the presence of pit shafts and shallow mineworkings. At the time of the last survey, many of the coal bings had been landscaped and forested. Previously there was a mine water pollution incident into the Water of Girvan in 1979.

Although thin coal seams of 'good' and inferior 'splinty' coal, up to 0.79 m thick, were proved in boreholes south of Martnaham Loch around [NS 3958 1675], they were variable in thickness and quality. These coals in the Limestone Coal Formation (and any concealed east of Ayr) were not considered worth mining (Eyles et al., 1930, pp.53–55). No economic coals were discovered within the formation on the eastern margin of the sheet.

Little is known about the coalbed methane potential of this area or suitability for underground coal gasification. The limited number of thicker seams combined with geological problems such as burning by sills and dykes may restrict the interest in prospecting. Coal rank is high volatile but with coking coal where thermally metamorphosed. Measured methane content in Scottish Coal Measures in the adjacent Mauchline area were quite low at about 2–3 m³ per tonne.

Peat

Peat may be used as a fuel or as a horticultural product. It has accumulated in poorly-drained hollows mainly in remote hilly areas south of Ayr. These are not considered to be economic resources.

Bulk minerals

Hard-rock aggregate

The larger dolerite intrusions could continue to provide good sources of hard-rock aggregate. In general they tend to be laterally persistent and of even quality, however, the quartz dolerites in this district have rarely been quarried in the past which tends to indicate that they are not the best. Quarries were opened up in the thick feldspar-phyric andesitic sills south of Maybole at [NS 3066 0688] and [NS 3013 0647]. The andesitic rocks from Knockhill Quarry [NS 2935 0680] and Knox Hill Quarry [NS 237 063] once worked variably altered rocks of secondary quality. Other hard igneous rocks such as the microdiorite [NS 355 106] near Guiltree Farm and the rhyolitic rock near Ballycoach [NS 362 115] have formerly been quarried on a small scale.

Former small quarries (e.g. Markland, [NS 381 170], now infilled) worked the Troon Volcanic Member. These basaltic lavas are too variable in quality to be economic at the present time. In the past the Carrick Volcanic Formation was quarried where the lava flows were more compact such as at Howmoor Quarry [NS 278 119]. The quality of these basalts is variable and unlikely to meet modern specifications for aggregates.

The Permian volcanic breccia at [NS 401 095], north-east of Loch Spallander, has more recently been quarried.

Sand and gravel

No sand and gravel is presently worked in the district (at 2009). Sand was formerly worked from pits in dune sands north-west of Monkton [NS 342 280]. Former pits in glaciofluvial sand and gravel at Lyonston [NS 313 102] and Fairyknowe [NS 318 104] were recorded by Goodlet (1970). Future resources may be found in the glaciofluvial deposits, blown sand, alluvial sands or coastal terraces of the district. Resources of gravel offshore in Ayr Bay have been examined in the past (Deegan et al., 1973). The marine gravels are locally mud-free, well shaped and subrounded, comprising up to 50 per cent of the deposit. However, the shrinkage potential is likely to be variable; in places the lava clasts are fresh but in others they are extremely decomposed (Deegan et al., 1973).

The glaciofluvial sand and gravels are partly sorted and of variable thickness as a result of meltwaters leaving eskers and kames in contact with the former ice-sheet. Examples of former pits in these deposits are found near Kirkmichael House [NS 343 083] and near Veenston Glen [NS 372 133]. Mounds of sand and gravel up to 4.5 m thick occur south of Martnaham Loch (e.g. at [NS 392 164]) but are too small and variable in quality to be economic resources. The presence of some coaly fragments within the deposits is deleterious to its quality in certain uses. The deposits are relatively small and only viable for local use.

Clay

Alluvial clays were worked for tile-making from the 19th century into the 20th century as agricultural drainage was improved. Many were still working at the time of the original survey in the 1860s, but had closed by the 1900s. Former workings are recorded in alluvial clay on sand and gravel near Whitehill Cottage [NS 399 170], south of Martnaham Loch. There are records of tile-making in clays 250 m north-west of Bank Farm at [NS 356 193] and near Knockjarder [NS 362 151] where the clay was blue in colour. Orange-brown alluvial clay was formerly extracted for tile-making from the alluvial spread east of the River Doon at [NS 3465 1695]. The former Macilveenston Tile works at [NS 363 137], south of Veenston Cottage, was also situated on River Doon alluvium. Farther east, the former Gateside (Glenside) tile works at [NS 366 092] may partly have worked clay from till; as did the old tile works at [NS 387 056] near Glenhead.

Near the coast, the Dipple Tile Works [NS 204 027], dating from 1837, extracted clay from a 3 to 5 m thick glaciolacustrine deposit, situated in a hollow between the middle and upper raised beaches. This clay contained a marine fauna at the base and a freshwater one towards the top. In 1842 in Kirkoswald Parish alone, about 1 million tiles were produced annually from three tile works (Simpson and MacGregor, 1932).

Clay has also been extracted from small quarries in the past. Shale and fireclay were extracted from seams in the Coal Measures from colleries such as the Enterkine Pits. The former Annbank brickworks were situated beside the railway to the south of Mossblown and produced 30 000 bricks a day.

Around Dailly fireclays from the Limestone Coal Formation were formerly worked for fire-bricks and ordinary bricks from the Kilgrammie Engine Pit [NS 2572 0192], Dalquharran [NS 2665 0182] and the Killochan [NS 2475 0143] pits (Simpson and MacGregor, 1932 p.54).

Freestone and sandstone

In the 19th century freestone from the thick sandstone unit near the top of the Limestone Coal Formation, just below the Index Limestone, was quarried for use in the Dailly–Straiton area (Eyles et al., 1949). Local sandstones in the Ballagan (west of Kirkbride House) and Kinnesswood formations have also been extracted in the past.

Culzean Castle is said (Smith, 1895) to be built of local fine-grained sandstone (Swanshaw Sandstone Formation) but its exact source is unknown. Other interbedded sandstones and siltstones, such as the beds east of Knockgardner, were quarried on a small scale in the past for local use. The latter quarry is now a Geological Consevation site.

Limestone

Only very small pits in the fine-grained dolomitic limestones or cementstones (Ballagan Formation) were worked, mainly for agricultural lime in the 18th and 19th centuries. Within the Kinnesswood Formation, such as near Threethorns [NS 357 044], Balgreggan [NS 357 049], Auchalton [NS 335 036] (Plate 24) and Lannielane [NS 313 018], quarries in pedogenic limestone (cornstone) beds, up to 4 m thick, were worked for lime from at least 1793 until 1860. There is evidence that the Todglen Limeworks [NS 344 046] continued into the 20th century. Similar pits situated south of Doonfoot, such as at [NS 3214 1817], are long abandoned.

The limestone at Lannielane Limeworks [NS 313 018] was fairly pure calcite, with only small rhombs of dolomite or ankerite scattered in the fine-grained calcite, which have subsequently been replaced by chert. Scattered grains of quartz, feldspar and probably limonitic clay granules were identified in samples (Muir and Hardie, 1956).

The Hurlet (Patna) Limestone, in the Lower Limestone Formation, does not appear to have been worked at the eastern edge of the sheet. In the Dailly area both the Hurlet (Captain's Bridge) and the uppermost limestone (Hosie) in the Lower Limestone Formation were once quarried for agricultural lime, although they are of poor quality.

The Ordovician limestone at Craighead [NS 234 014] was also extensively worked in the past, not only for lime but also for road metal and surface dressing (Eyles et al., 1949). It is essentially a calcite limestone with a little magnesium oxide, clay and quartz content (Muir and Hardie, 1956).

Semiprecious stones

Agate

Minor amounts of agate or variegated chalcedony (Heddle, 1901 p.76; Macpherson, 1989) are collected for cutting and polishing from the geodes weathered out of the Carrick Volcanic Formation, either from pebbly beaches or ploughed fields in the Dunure area. They formed as amygdales infilling vesicles, generally occurring within the more siliceous lavas. The abraded examples (rounded on the shore) have been termed 'Scotch pebbles'. A comprehensive collection was made and described by J. Smith (1910) of the semi-precious stones and geodes he found as the coastal railway from Ayr to Girvan was excavated (Plate 25a), (Plate 25b).

Besides concentric banded, straight banded (or onyx) and moss agate, he described carnelian, jasper, chalcedony, opal, amethyst and quartz. They commonly occur as compound amygdales of more than one variety of silica, and some also include minerals such as calcite, celadonite and delessite. Bright red jasper was recorded as various shaped masses, up to 0.15 m across, within white crystalline limestone lenses (Smith, 1909) forming part of the Carrick Volcanic Formation.

John Smith's agate collection is curated by BGS Edinburgh. Wintersgill (2004) documented this collection which is important in the understanding of agates and their formation. This is because, although Smith's theories of agate formation have been proven incorrect by modern analysis, he is regarded as a pioneer in the study of agates.

An up-to-date description and list of localities in Ayrshire is given by Macpherson (1989). In the latter booklet a so-called 'scenic' agate collected from the Heads of Ayr is illustrated which contains much green celadonite.

Studies of the oxygen isotopes from Dunure and Turnberry (Fallick et al., 1985) support the idea of a low temperature (about 50°C) origin for the agates from a fluid having at least a component of meteoric water.

Metalliferous minerals

No currently economic resources of metalliferous minerals are known. Former reports of metalliferous minerals are outlined below.

Lead and barite

Galena was reported from a disused quarry south of Montgomerieston [NS 3702 1218] (and also at Guiltreehill) as being worked prior to 1845 (Eyles et al., 1949, p.19; Simpson and MacGregor, 1932, p.156) and the ore was stated to contain a high proportion of silver.

At present only thin veins of pale pink barytes are exposed in the quarry. The veins, up to 5 cm wide, contain barite and calcite with minor amounts of galena and probable chalcopyrite. The veins are nearly vertical and strike N066°. Other thin, calcite-coated vertical joints in the quarry strike eastward. They all cross-cut greenish grey volcaniclastic sandstones intercalated within the Lower Devonian Carrick Volcanic Formation.

At St Murray's sandstone quarry [NS 305 113] near Maybole, thin veins of lead ore are recorded and coatings of manganese oxide occur on joint surfaces.

Three or four thin (about 1 cm) sulphide veins were found by the River Doon at [NS 3395 1635] near Nether Auchendrane, within east–west fractures in the Carrick Volcanic Formation. Minerals identified by x-ray powder analysis include barite, marcasite, pyrite and chalcopyrite with secondary malachite. Barite and chalcopyrite also occur in some veins farther south, although they are dominantly calcite and barite with a tiny amount of disseminated chalcopyrite and some larger cubes of galena. The latter veins were once worked but they are unlikely to be of any economic importance.

Ironstone

The Dalmellington (or Lugar) Blackband Ironstone was extensively worked in the 19th century (Macgregor et al., 1920) as a source of iron. The seam lies near the base of the Lower Coal Measures and its average thickness was about 0.23 m. No economic resources remain.

Gold

In the recent past a report has been published (Leake et al., 1997) on the potential for gold mineralisation in Permo-Triassic redbeds and their contacts with underlying rocks (in this case the Scottish Upper Coal Measures at the edge of the Mauchline Basin).

Water supply

Abundant potable surface water is supplied by the reservoir scheme at Loch Bradan, west of Loch Doon around [NX 430 970]. This loch supplies most of South Ayrshire, including Ayr, Prestwick, Turnberry, Dalrymple, Maybole and farther afield. The scheme includes water diverted from the River Stinchar. Currently (in 2009) the scheme is being upgraded to increase its capacity from 95 megalitres per day to 120.

Smaller surface reservoirs such as Loch Spallander and near Raith [NS 395 271] and Carcluie [NS 342 155] also provide supplies of water.

For more remote localities, there is potential for obtaining groundwater mainly from aquifers in strata such as the Kinnesswood or Swanshaw Sandstone formations. Older sandstones and interbedded mudstones were indurated during the Caledonian events, so in such rocks, groundwater may only occur in shallow cracks and joints.

Water bores up to 1968 listed in the Well Catalogue series (Carter and McAdam, 1968) indicate that most of the bores around Ayr were in Carboniferous strata, mainly Coal Measures, and those around Maybole were in Swanshaw Sandstone Formation (Lower Old Red Sandstone). Successful boreholes are dependant on the intersection of favourable cracks and joints and yields are normally of 1 to 3 l/s in strata such as the Swanshaw Sandstone Formation (Robins, 1990). Exceptionally, an overflowing borehole at Culzean [NS 252 119] yielded 11 litres/second (l/s) for an unknown drawdown. Robins (1990) considered that this enhanced yield could be due to secondary permeability by solution and erosion caused by greater throughflow of groundwater from this elevated site towards the nearby coast. Around Dailly, water was obtained from workings in the Limestone Coal Formation. In the Scottish Coal Measures Group, many of the coal mines pumped water for drainage purposes at rates up to 75 l/s; this was because the effective hydraulic contact between the mine workings and the aquifer was large. When considered in terms of a single borehole, yields of up to 5 l/s are more typical of Coal Measures strata (Robins, 1990). Many of these aquifers are likely to be confined by a cover of till. However, there is insufficient data to estimate aquifer yield and water quality. The numerous faults and dykes in the district will also affect groundwater flow.

Information sources

Further geological information held by the British Geological Survey relevant to the Ayr district is listed below. It includes published material in the form of maps, memoirs and reports, and unpublished maps and reports. Also included are other sources of information held by BGS in a number of collections, including borehole records, mine plans, fossils, rock samples, thin sections, hydrogeological data and photographs.

Searches of indexes to some of the collections can be made on the Geoscience Data Index (GDI) in BGS libraries and online at www.bgs.ac.uk.

This website also provides access to the BGS Enquiry Service.

Books

British Regional Geology
The Midland Valley of Scotland, 3rd edition, 1985.
Hydrogeology
Robins, N S. 1990. Hydrogeology of Scotland. (London: HMSO for the British Geological Survey.)
Memoirs
Geology of Central Ayrshire (Sheet 14), 1949
The Limestones of Scotland, 1949
The Limestones of Scotland: chemical analyses and petrography, 1956
The Economic Geology of the Ayrshire Coalfields Area IV, 1932
The Economic Geology of the Ayrshire Coalfields Area III, 1930

Reports

Geochemistry
British Geological Survey. 1993. Regional geochemistry of Southern Scotland and part of northern England. (Keyworth, Nottingham: British Geological Survey.)
Biostratigraphy
Dean, M T. 2002 The Carboniferous macropalaeontology and biostratigraphy of Ayr and Mauchline (Scotland Sheet 14). British Geological Survey Internal Report, IR/02/006.
Hydrogeology
Ball, D F. 1999. An overview of groundwater in Scotland. British Geological Survey Technical Report, WD/99/44.
Geophysics
Kimbell, G S, Carruthers, R M, Walker, A S D, and Williamson, J P. 2006. Regional Geophysics of Southern Scotland and Northen England. Version 1.0 on CD-ROM. (Keyworth, Nottingham, British Geological Survey.)
Geothermal potential
British Geological Survey. 1988. Geothermal energy in the United Kingdom: review of the British Geological Survey's Programme 1984–1987. Investigation of the Geothermal Potential of the UK, British Geological Survey.

Maps

Geology
1:625 000
United Kingdom (North Sheet) Solid Geology, 2007
United Kingdom (North Sheet) Quaternary geology, 1977
1:250 000
Clyde (55N 06W), Solid Geology, 1985
Clyde (55N 06W), Sea Bed Sediments and Quaternary Geology, 1985.
1:50 000 and 1:63 360
Details of the original geological surveys are listed on editions of the 1:63 360 geological sheets. Copies of the fair-drawn maps of these earlier surveys may be consulted at the BGS Library, Edinburgh.
Sheet 14 (Ayr) (Solid & Drift), 1951
Sheet 14W (Ayr) Drift, 1978
Sheet 14W (Ayr) Solid, 1978
Sheet 14W (Ayr) Bedrock, 2008
Sheet 14W (Ayr) Superficial and Simplified Bedrock, 2009
1:10 000 and 1:10 560
The maps at 1:10 000 scale produced following recent (1998–2002) surveys and lying wholly or in part within the 1:50 000 scale sheet 14W are listed below, together with the surveyors' or compilers' initials and the date of release. The surveyors/compilers were: A Monaghan (AM), R A Smith (RAS), S L B Arkley (SLBA), and J D Floyd (JDF). The map of the Craighead area incorporates published work by L R M Cocks, P T Toghill and D A T Harper and unpublished work by J D Floyd. The maps are not published but available for consultation in the BGS Library, Edinburgh, and also at Keyworth and in the London Information Office, in the Natural History Museum, South Kensington. Copies may be purchased from the Sales Desk.

Solid and drift
NS21SE and NS21SW (part) (Croy)	RAS	2000
NS30NW (Kirkmichael)	RAS	2006
NS30NE (Loch Spallander)	RAS	2002
NS30SE (Straiton)	RAS	2002
NS31NE (Martnaham Loch)	RAS	1999
NS31SE (Dalrymple)	RAS	2000
NS40NW (Waterside) (part)	RAS	2000
NS40SW (Big Hill on the Baing) (part)	RAS	2000
NS41NW (Drongan) (part)	RAS	1998
NS41SW (Patna) (part)	RAS	1999
NS42NW (Tarbolton) (part)	RAS	1998
NS42NW (Annbank) (part)	RAS	1998

Solid
NS20NW and NS10NE (part) (Maidens)	AM	2003
NS20NE (Maybole)	AM	2003
NS20SW and NS10SE (part) (Craighead)	JDF, AM	2003
NS20SE (Dailly)	AM	2003
NS21NE and NS21NW (part) (Dunure)	AM	2003
NS30SW (Kilkerran)	SLBA, RAS	2007
NS31NW (Doonfoot)	SLBA, RAS	2007
NS31SW (Minishant)	SLBA, RAS	2007
NS32NW (Prestwick)	AM	2004
NS32NE (Monkton)	AM	2001
NS32SW (Ayr)	AM	2001
NS32SE (Whitletts)	AM	2001

Drift
NS20NW and NS10NE (part) (Maidens)	AM	2003
NS20NE (Maybole)	AM	2003
NS20SW and NS10SE (part) (Craighead)	AM	2003
NS20SE (Dailly)	AM	2003
NS21NE (Dunure)	AM	2003
NS30SW (Kilkerran)	SLBA	2007
NS31NW (Doonfoot)	SLBA	2007
NS31SW (Minishant)	SLBA	2007
NS32SE (Whitletts)	AM	2001
NS32NW (Prestwick)	AM	2001
NS32SW (Ayr)	AM	2001
NS32NE (Monkton)	AM	2001

3D models
1:25 000 NS32
1:25 000 NS42
1:25 000 NS41
1:50 000 Ayrshire coalfield merged model
1:1000 000 Western MVS onshore-offshore regional model

Geophysics

1:1 000 000
British Geological Survey. 1998. 1:1 000 000 UTM Series: Gravity Anomaly Map, Northern Britain 54°N-60°N, 6°W-0°E. Smith, I F, and Edwards, J W F (compilers). In British Geological Survey. 2006. 1:1 000 000 UTM Series: Gravity Anomaly Maps. Chacksfield, B C, and Edwards, J W F, (compilers). (Keyworth, Nottingham: British Geological Survey.)
British Geological Survey. 2004. 1:1 000 000 UTM Series: Total Field Magnetic Anomaly Map, Northern Britain 54°N-60°N, 6°W-0°E. Chacksfield, B C, and Royles, C P, (compilers). In British Geological Survey. 2006. 1:1 000 000 UTM Series: Magnetic Anomaly Maps. Chacksfield, B C, and Edwards, J W F, (compilers). (Keyworth, Nottingham: British Geological Survey.)
1:625 0000
United Kingdom (North Sheet) Magnetic anomaly, 2007. United Kingdom (North Sheet) Gravity anomaly, 2007.
1:250 000
Clyde (Sheet 55N 06N) Aeromagnetic anomaly, 1980, print on demand.
Clyde (Sheet 55N 06N), Bouguer gravity anomaly, 1985, print on demand.
1:50 000
Geophysical information maps; these are plot-on-demand maps which summarise graphically the publicly available geophysical information held for the sheet in the BGS databases. Features include:
Regional gravity data: Bouguer anomaly contours and location of observations
Regional aeromagnetic data: total field anomaly contours and location of digitised data points along flight lines
Gravity and magnetic fields plotted on the same base map at 1:50 000 scale to show correlation between anomalies
Location of local geophysical surveys
Location of public domain seismic reflection and refraction surveys
Location of deep boreholes and those with geophysical logs

Geochemistry

1:250 000
Point-source geochemical data processed to generate a smooth continuous surface presented as an atlas of small-scale colour-classified digital maps.
Southern Scotland and part of northern England, 1993.
Geochemical Survey Programme data are also available in other forms including hard copy and digital data.
Hydrogeology
1:625 000
Sheet 18 (Scotland), 1988
Groundwater vulnerability (Scotland), 2001

Documentary collections and indexes

Borehole records
BGS holds records of boreholes which can be consulted at BGS, Edinburgh, where copies of most records may be purchased. For the Ayr district the collection consists of the sites and logs of about 2938 boreholes. Index information, which includes site references, for these boreholes has been digitised. The logs are either hand-written or typed and many of the older records are drillers' logs.
Site investigation reports
This collection consists of records of site explorations carried out to investigate foundation conditions prior to construction. There is a digitised index and the reports themselves are held on microfiche. For the Ayr district there are presently about 289 reports.
Mine plans
BGS maintains a collection of plans of underground mines for minerals other than coal or oil-shale. There are no such mine plans for this area. Coal abandonment plans are held by the Coal Authority, Mining Records Department, 200 Lichfield Lane, Mansfield, Nottingham NG18 4RG.
Hydrogeological data
Records of approximately 52 water boreholes are held at BGS, Edinburgh. The area is covered by Water Supply Papers of the Geological Survey of Great Britain, 1968: Wells in Scottish One-inch Sheets 7–18.
Geochemical data
Records of stream-sediment and other analyses are held at BGS, Keyworth.
Gravity and magnetic data
Records are held at BGS, Keyworth.

Material collections

BGS photographs

Some 338 photographs illustrating aspects of the geology of the Ayr district are deposited for reference in the libraries at BGS, Edinburgh and BGS, Keyworth; and in the BGS Information Office, London. The photographs were taken at various times over the last century. The photographs depict details of the various rocks and sediments exposed either naturally or in excavations, and also some general views. The images can be supplied in a number of formats at various prices from the Photographic Department, BGS, Edinburgh. Some of the BGS collection can be viewed on our website. Please see http://geoscenic.bgs.uk.

Petrological collections

The petrological collections for the Ayr district consist of about 900 hand specimens and thin sections. Most samples and thin sections are of the igneous rocks of the district with some sandstones and conglomerates. The sedimentary rocks are poorly represented. Further details can be found on Britrocks, the BGS online mineralogy & petrology database: http://www.bgs.ac.uk/data/britrocks.html and on GeoIndex, the BGS online GIS system: http://www.bgs.ac.uk/geoindex/home.html. Information on databases, rock samples, thin sections, and geochemical analyses can be obtained from Dr E R Phillips, Mineralogy and Petrology Division, BGS, Edinburgh or through enquiries@bgs.ac.uk.

Borehole samples

There is a small number of borehole core and samples from the Ayr district. Details can be found on the BGS online onshore borehole material database: http://www.bgs.ac.uk/data/bmd/home.html and on GeoIndex, the BGS online GIS system: http://www.bgs.ac.uk/geoindex/home.html. Further information is available through the BGS Enquiries Service: enquiries@bgs.ac.uk.

Palaeontological collections

The collections of biostratigraphical specimens are taken from surface and temporary exposures, and from boreholes in the Ayr district. Most of the collections are primarily locality based (The Survey Collection) and many of their locations can be seen on GeoIndex, the BGS online GIS system: http://www.bgs.ac.uk/geoindex/home.html. Select Map theme: Collections and Click the 'Fossil localities' box. These specimens are held in Edinburgh.

Some specimens are of taxonomic importance. These are in the 'Type & Stratigraphic' Collection, held largely in Keyworth. Many of these specimens are listed on the BGS online fossil database — PalaeoSaurus: http://www.bgs.ac.uk/palaeosaurus/home.cfm as well as on GeoIndex. Select Map theme: Collections and Click the 'Fossil specimens' box.

Macrofossils in the collection held in either Edinburgh or Keyworth, listed by chronostratigraphical unit, currently total:

Ordovician 486
Silurian 614
Devonian 22
Carboniferous
(not stratigraphically defined) 131
Inverclyde Group 187
Strathclyde Group 81
Clackmannan Group 320
Coal Measures 1495
Quaternary 20

Enquiries concerning all fossil material should be directed to the Curator, Palaeontology Section, BGS, Edinburgh or to enquiries@bgs.ac.uk

Seismic Surveys: Data for the district are held by BGS in Keyworth.

Sites of Special Scientific Interest are the responsibility of Scottish Natural Heritage.

Contact Addresses

British Geological Survey, Murchison House >West Mains Road, Edinburgh EH9 3LA. Telephone: 0131 667 1000. Fax: 0131 668 2683
British Geological Survey (Headquarters), Environmental Science Centre, Keyworth, Nottingham. NG12 5GG Telephone: 0115 936 3100. Fax: 0115 936 3200
London Information Office, Natural History Museum, Cromwell Road, London SW7 5BD. Telephone: 020 7589 409. Fax: 020 7584 8270

Website: www.bgs.ac.uk

References

Most of the references listed below are held in the Library of the British Geological Survey at Edinburgh and Keyworth, Nottingham. Copies of the references can be purchased subject to current copyright legislation.

Aldridge, R J, Siveter, David J, Siveter, Derek J, Lane, P D, Palmer, D, and Woodcock, N H. 2000. British Silurian Stratigraphy. Geological Conservation Review Series, No. 19. (Peterborough: Joint Nature Conservation Committee.)

Allen, J R L, and Williams, B P J. 1981. Beaconites antarcticus: a giant channel-associated trace fossil from the Lower Old Red Sandstone of South Wales and the Welsh Borders. Geological Journal, Vol. 16, 255–269.

Anderson, F W, and Pringle, J. 1946. On a section of the Balclatchie Beds at Craighead Quarry, near Girvan, Ayrshire. Geological Magazine, Vol. 83, 172–176.

Armstrong, D. 1957. Dating some of the minor intrusions of Ayrshire. Nature, London, Vol. 180, 1277.

Armstrong, H A, and Owen, A W. 2000. Age and provence of limestone clasts in Lower Old Red Sandstone conglomerates: implications for the geological history of the Midland Valley Terrane. 459–472 in New Perspectives on the Old Red Sandstone. Friend, P F, and Williams, B P J (editors). Special Publication of the Geological Society of London, No. 180.

Armstrong, H A, Owen, A W, Scrutton, C T, Clarkson, E N K, and Taylor, C M. 1996. Evolution of the Northern Belt, Southern Uplands: implications for the Southern Uplands controversy. Journal of the Geological Society of London, Vol. 153, 197–205.

Atherton, M P, and Ghani A A. 2002. Slab breakoff: a model for Caledonian, Late Granite syn-collisional magmatism in the orthotectonic (metamorphic) zone of Scotland and Donegal, Ireland. Lithos, Vol. 62, 65–85.

Bailey, E B. 1928. 57–58 in Summary of Progress of the Geological Survey of Great Britain for 1927, Part 1 (London: HMSO.)

Bamford, D, Nunn, K, Prodehl, C, and Jacob, B. 1978. LISPB-IV. Crustal structure of northern Britain. Geophysical Journal of the Royal Astronomical Society, Vol. 54, 43–60.

Bassett, D A. 1958. Geological excursion guide to the Glasgow district. Printed for the Geological Society of Glasgow by Titus Wilson and Son, Ltd, Kendal.

Batchelor, R A. 1999. Metabentonites from the Silurian inliers of the southern Midland Valley of Scotland: distribution and geochemistry. Scottish Journal of Geology, Vol. 35, 71–77.

Begg, J L. 1939. Some new species of Proetidae and Otarionidae from the Ashgillian of Girvan. Geological Magazine, Vol. 76, 372–382.

Begg, J L. 1946. Some new fossils from the Girvan district. Transactions of the Geological Society of Glasgow, Vol. 21, 29–45.

Bell, B R, and Williamson, I T. 2002. Tertiary igneous activity. 371–407 in The Geology of Scotland (4th edition). Trewin, N H (editor). (London: The Geological Society.)

Bergström, S M. 1980. Conodonts as palaeotemperature tools in Ordovician rocks of the Caledonides and adjacent areas in Scandinavia and in the British Isles. Geologiska Föreningens i Stockholm Förhandlingar, Vol. 102, 377–392.

Bergström, S M. 1990. Biostratigraphic and biogeographic significance of middle and upper Ordovician conodonts in the Girvan succession, south-west Scotland. Courier Forschungsinstitut Senckenberg, Vol. 118, 1–43.

Bluck, B J. 1978. Sedimentation in a late orogenic basin: the Old Red Sandstone of the Midland Valley of Scotland. 249–278 in Crustal evolution in northwestern Britain and adjacent regions. Bowes, D R, and Leake, B E (editors). Geological Journal Special Issue, No. 10.

Bluck, B J. 1980. Structure, generation and preservation of upward fining, braided stream cycles in the Old Red Sandstone of Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 71, 29–46.

Bluck, B J. 1983. Role of the Midland Valley of Scotland in the Caledonian orogeny. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 74, 119–136.

Bluck, B J. 2000. Old Red Sandstone basins and alluvial systems of Midland Scotland. 417–437 in New Perspectives on the Old Red Sandstone. Friend, P F, and Williams, B P J (editors). Special Publications of the Geological Society of London, No. 180.

Bluck, B J. 2002. The Midland Valley terrane. 149–166 in The Geology of Scotland (4th edition). Trewin, N H (editor). (London: The Geological Society.)

Bluck, B J, and Ingham, J K. 1992. The Girvan-Ballantrae Complex. 301–308 in Geological excursions around Glasgow and Girvan. Lawson, J D, and Weedon, D S (editors). (Glasgow: Geological Society of Glasgow.)

Bos, J A A, Dickson, J H, Coope, G R, and Jardine, W G. 2004. Flora, fauna and climate of Scotland during the Weichselian Middle Pleniglacial — palynological, macrofossil and coleopteran investigations. Palaeogeography, Palaeoclimatology, Palaeoecology, Vol. 204, 65–100.

Bowen, D Q (editor). 1999. A revised correlation of Quaternary deposits in the British Isles. Geological Society of London Special Report, No. 23.

Boyd, W E. 1986a. Late Devensian shoreline position in North Ayrshire. Scottish Journal of Geology, Vol. 22, 412–416.

Boyd, W E. 1986b. Fossil Lithothamnium (calcareous algae) rhodoliths from Late Quaternary raised coastal sediments, Irvine, Ayrshire. Scottish Journal of Geology, Vol. 22, 165–177.

Brade-Birks, S G. 1923. Notes on Myriapoda XXVIII. Kampecaris tuberculata n.sp. from the Old Red Sandstone of Ayrshire. Proceedings of the Royal Physical Society, Edinburgh, Vol. 20, 277–280.

Brand, P J. 1977. The fauna and distribution of the Queenslie Marine Band (Westphalian) in Scotland. Report of the Institute of Geological Sciences, No. 77/18.

Brand, P J. 1983. Stratigraphical palaeontology of the West-phalian of the Ayrshire Coalfield, Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 73, 173–190.

British Geological Survey. 1985a. Clyde. Sheet 55N 06W. Solid geology. 1:250 000. (Southampton: Ordnance Survey for British Geological Survey.)

British Geological Survey. 1985b. Clyde. Sheet 55N 06W. Sea Bed sediments and Quaternary geology. 1:250 000. (Southampton: Ordnance Survey for British Geological Survey.)

British Geological Survey. 1995. Carrick. Scotland Sheet 8W. Solid Geology. 1:50 000. (Keyworth, Nottingham: British Geological Survey.)

British Geological Survey. 2003. Dunure. Sheet NS21NW and part of NS21NW. Solid edition. 1:10 000 (Keyworth, Nottingham: British Geological Survey.)

Brookfield, M E. 1978. Revision of the stratigraphy of Permian and supposed Permian rocks of southern Scotland. Geologische Rundschau, Vol. 67, 110–149.

Browne, M A E. 1980. The Upper Devonian and Lower Carboniferous (Dinantian) of the Firth of Tay, Scotland. Institute of Geological Sciences Report, 80/9.

Browne, M A E, and Mcmillan, A A. 1989. Quaternary geology of the Clyde valley. British Geological Survey Research Report, SA/89/1.

Browne M A E, Dean, M T, Hall, I H S, Mcadam, A D, Monro, S K, and Chisholm, J I. 1999. A lithostratigraphical framework for the Carboniferous rocks of the Midland valley of Scotland. Version 2. British Geological Survey Research Report, RR/99/07.

Browne, M A E, Smith, R A, and Aitken, A M. 2002. Stratigraphical framework for the Devonian (Old Red Sandstone) rocks south of a line from Fort William to Aberdeen. British Geological Survey Research Report, RR/01/04.

Bruce, S A H. 1992. The systematic description and palaeo-environmental significance of Silurian ichnofaunas along the northern margin of the Iapetus Ocean in Ireland and Scotland. Unpublished PhD thesis, National University of Ireland.

Bull, E E, and Loydell, D K. 1995. Uppermost Telychian graptolites from the North Esk Inlier, Pentland Hills, near Edinburgh. Scottish Journal of Geology, Vol. 31, 163–170.

Burns, W. 1888. Notes on the Heads of Ayr. Transactions of the Geological Society of Glasgow, Vol. 8, part 2, 287–291.

Busby, J P, Akhurst, M C, and Walker, A S D. 2009. A new high-resolution aeromagnetic dataset over central Ayrshire: insights into the concealed geology. Scottish Journal of Geology, Vol. 54, 1–13.

Cameron, I B, and Stephenson, D. 1985. British Regional Geology: the Midland Valley of Scotland (3rd edition). (London: HMSO for British Geological Survey.)

Cameron, I B, Stone, P, and Smellie, J. 1986. Geology of the country around Girvan. Memoir of the British Geological Survey, Sheet 7 (Scotland).

Carter, A V F, and Mcadam, A D. 1968. Well Catalogue Series: Records of wells in the areas of Scottish one-inch Geological Sheets Girvan (7), Carrick (8), Maxwelltown (9), Dumfries (10), Langholm (11), Campbeltown (12), South Arran (13), Ayr (14), Sanquhar (15), Moffat (16), Jedburgh (17) and Morebattle (18). Water Supply Papers of the Geological Survey of Great Britain.

Chambers, L M, and Pringle, M S. 2001. Age and duration of activity at the Mull Tertiary igneous centre, Scotland, and confirmation of the existence of subchrons during anomaly 26r. Earth and Planetary Science Letters, Vol. 193, 333–345.

Charlesworth, J K. 1925. The Glacial Geology of the Southern Uplands of Scotland, West of Annandale and Upper Clydesdale. Transactions of the Royal Society of Edinburgh, Vol. 55, 1–25.

Chesher, J A, Deegan, C E, Ardus, D A, Binns, P E, and Fannin, N G T. 1972. IGS marine drilling with M V Whitethorn in Scottish waters 1970–71. Report of the Institute of Geological Sciences, No. 72/10.

Clarkson, E N K. 1985. A brief history of Scottish palaeontology 1834–1984. Scottish Journal of Geology, Vol. 21, 389–406.

Clarkson, E N K, and Howells, Y. 1981. Upper Llandovery trilobites from the Pentland hills, near Edinburgh, Scotland. Palaeontology, Vol. 24, 507–536.

Clarkson, E N K, Eldredge, N, and Henry, J-L. 1977. Some Phacopina (Trilobita) from the Silurian of Scotland. Palaeontology, Vol. 20, 119–142.

Clarkson, E N K, Harper, D A T, and Höey, A N. 1998. Basal Wenlock biofacies from the Girvan district, SW Scotland. Scottish Journal of Geology, Vol. 34, 61–71.

Clayton, G. 1971. Lower Carboniferous miospore assemblage from the Calciferous Sandstone measures of the Cockburnspath region of eastern Scotland. Pollen et Spores, Vol. 12, 577–600.

Cleal, C J, and Thomas, B A. 1995. Palaeozoic palaeobotany of Great Britain. Geological Conservation Review Series, No. 9. (Peterborough: Joint Nature Conservation Committee.)

Cocks, L R M, and Toghill, P. 1973. The biostratigraphy of the Silurian rocks of the Girvan District, Scotland. Journal of the Geological Society of London, Vol. 129, 209–243.

Cocks, L R M, Woodcock, N H, Rickards, R B, Temple, J T, and Lane, P D. 1984. The Llandovery Series of the Type Area. Bulletin of the British Museum (Natural History) (geology), Vol. 38, 131–182.

Craig, G Y (editor). 1991. Geology of Scotland (3rd edition). (London: The Geological Society of London.)

Davies, A. 1972. Carboniferous rocks of the Muirkirk, Gass Water and Glenmuir areas of Ayrshire. Bulletin of the Geological Survey of Great Britain, No. 40, 1–49.

Davies, H C, Dobson, M R, and Whittington, R J. 1984. A revised seismic stratigraphy for Quaternary deposits on the inner continental shelf west of Scotland between 55°30´N and 57°30´N. Boreas, Vol. 13, 49–66.

Dean, M T. 2000. Preliminary report on the conodont bio-stratigraphy of a limestone 'pebble' in a conglomerate at Croy Foreshore, Ayrshire. British Geological Survey Technical Report, WH/00/10R.

Dean, M T. 2002. The Carboniferous macropalaeontology and biostratigraphy of Ayr and Mauchline (Scotland Sheet 14). British Geological Survey Internal Report, IR/02/006.

Deegan, C E, Kirby, R, Rae, I, and Floyd R. 1973. The superficial deposits of the Firth of Clyde and its sea lochs. Report of the Institute of Geological Sciences, No. 73/9.

Dentith, M C, and Hall, J. 1989. MAVIS — an upper crustal seismic refraction experiment in the Midland Valley of Scotland. Geophysical Journal International, Vol. 99, 627–643.

De Souza, H A F. 1979. The geochronology of Scottish Carboniferous volcanism. Unpublished PhD thesis. University of Edinburgh.

De Souza, H A F. 1982. Age data from Scotland and the Carboniferous timescale. 456–465 in Numerical Dating in Stratigraphy. Odin, G S (editor). (Chichester: Wiley.)

Dewey, J F, and Strachan, R A. 2003. Changing Silurian-Devonian relative plate motion in the Caledonides: sinistral transpression to sinistral transtension. Journal of the Geological Society, London, Vol. 160, 219–229.

Dineley, D L, and Metcalf, S J. 1999. Fossil fishes of Great Britain. Geological Conservation Review Series, No. 16. (Peterborough: Joint Nature Conservation Committee.)

Donovan, S K. 1992. A new crinoid from the Ashgill Starfish Bed, Threave Glen. Scottish Journal of Geology, Vol. 28, 123–126.

Donovan, S K, and Clark, N D L. 1992. An unusual crinoid columnal morphospecies from the Llandovery of Scotland and Wales. Palaeontology, Vol. 35, 27–35.

Dorning, K J. 1982. Early Wenlock acritarchs from the Knockgardner and Straiton Grit Formations of Knockgardner, Ayrshire. Scottish Journal of Geology, Vol. 18, 267–273.

Doyle, E N, Höey, A N, and Harper, D A T. 1991. The rynchonellide brachiopod Eocoelia from the upper Llandovery of Ireland and Scotland. Palaeontology, Vol. 34, 439–454.

Durant, G. 1999. Port Schuchan to Dunure Castle; Culzean Harbour; Turnberry Lighthouse to Port Murray. 542–552 in Caledonian Igneous Rocks of Great Britain. Stephenson, D, Bevins, R E, Millward, D, Highton, A J, Parsons, I, Stone, P, and Wadsworth, W J. Geological Conservation Review Series, No. 17. (Joint Nature Conservation Committee, Peterborough.)

Emeleus, C H, and Bell, B R. 2005. British Regional Geology: the Palaeogene volcanic districts of Scotland. (4th edition) (Nottingham: British Geological Survey.)

Eyles, V A, Simpson, J B, and Macgregor, A G. 1929. The igneous geology of Central Ayrshire. Transactions of the Geological Society of Glasgow, Vol. 18, 361–387.

Eyles, V A, Simpson, J B, and Macgregor, A G. 1930. Economic Geology of the Ayrshire Coalfields. Area III. Memoir of the Geological Survey, Scotland.

Eyles, V A, Simpson, J B, and Macgregor, A G. 1949. Geology of Central Ayrshire. (2nd edition). Memoir of the Geological Survey of Great Britain, Sheet 14 (Scotland).

Fallick, A E, Jocelyn, J, Donnelly, T, Guy, M, and Behan, C. 1985. Origin of agates in volcanic rocks from Scotland. Nature, Vol. 313, 672–674.

Fearnhead, F E, and Harper, D A T. 2007. Petalocrinus (Echinodermata, Crinoidea) from the Llandovery (lower Silurian: Rhuddanian) of the Girvan district, SW Scotland. Scottish Journal of Geology, Vol. 43, 69–74.

Finlayson, A, Merritt J W, Browne, M A E, Merritt J E, Mcmillan, A A, and Whitbread, K. 2010. Late Devensian ice sheet build up, dynamics and decay configurations in west central Scotland. Quaternary Science Reviews, Vol. 29, 969–988.

Fisher, R V, and Schmincke, H U. 1984. Pyroclastic Rocks. (Berlin: Springer-Verlag.)

Floyd, J D. 1982. Stratigraphy of a flysch succession: the Ordovician of W Nithsdale, SW Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 73, 1–9.

Floyd, J D. 1996. Lithostratigraphy of the Ordovician rocks in the Southern Uplands: Crawford Group, Mofat Shale Group, Leadhills Supergroup. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 86, 153–165.

Floyd, J D. 1999. Geology of the Carrick–Loch Doon district. Memoir of the British Geological Survey, Sheets 8W and 8E (Scotland).

Floyd, J D, and Williams, M. 2003. A revised correlation of Silurian rocks in the Girvan district, SW Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 93, 383–392.

Floyd, J D, Williams, M, and Rushton, A W A. 1999. Late Ordovician (Ashgill) ostracodes from the Drummuck Group, Craighead Inlier, Girvan district, SW Scotland. Scottish Journal of Geology, Vol. 35, 15–24.

Forsyth, I H, and Chisholm, J I. 1977. The Geology of East Fife. Memoir of the Geological Survey of Great Britain, Sheet 41 and part of sheet 49 (Scotland).

Forsyth, I H, Hall, I H S, and Mcmillan, A A. 1996. Geology of the Airdrie district. Memoir of the British Geological Survey, Sheet 31W (Scotland).

Fortey, R A, Harper, D A T, Ingham, J K, Owen, A W, Parkes, M A, Rushton, A W A, and Woodcock, N H. 2000. A revised Correlation of Ordovician Rocks in the British Isles. Special Report of the Geological Society of London, No. 24.

Freshney, E C. 1959. An extension of the Silurian succession in the Craighead Inlier, Girvan. Transactions of the Geological Society of Glasgow, Vol. 24, 27–31.

Friend, P F, Williams, B P J, Ford, M, and Williams, E A. 2000. Kinematics and dynamics of Old Red Sandstone basins. 29–60 in New Perspectives on the Old Red Sandstone. Friend, P F, and Williams, B P J (editors). Special Publication of the Geological Society of London, No. 180.

Geikie, A. 1869. Part of the coast of Ayrshire. Memoir of the Geological Survey, Sheet 13 (Scotland).

Geikie, A. 1897. The ancient volcanoes of Great Britain. (London: Macmillan.)

Geikie, A, Geikie, J, and Peach, B N. 1869. Ayrshire: Southern district. Memoir of the Geological Survey, Sheet 14 (Scotland).

Geological Survey of Great Britain (Scotland). 1966a. Sheet NS32SE. Solid. 1:10 560. (Chessington, Surrey: Ordnance Survey for the Geological Survey.)

Geological Survey of Great Britain (Scotland). 1966b. Sheet NS32NE. Drift. 1:10 560. (Chessington, Surrey: Ordnance Survey for the Geological Survey.)

Geological Survey of Scotland. 1929. Ayrshire Sheet 50SW. Solid and Drift. 1:10 560. Revised. (Southampton: Ordnance Survey for the Geological Survey of Scotland.)

Gibb, F G F, and Henderson, C M B. 1978. The petrology of the Dippin sill, Isle of Arran. Scottish Journal of Geology, Vol. 14, 1–27.

Goodlet, G A. 1970. Sands and gravels of the southern counties of Scotland. Report of the Institute of Geological Sciences, No. 70/4, 82 pp.

Gradstein, F M, Ogg, J G, and Smith, A G (compilers). 2004. A Geologic Time Scale 2004. (Cambridge: University Press.)

Greig, D C. 1988. Geology of the Eyemouth district. Memoir of the British Geological Survey, Sheet 34 (Scotland).

Guion, P D, Fulton, I M, and Jones, N J. 1995. Sedimentary facies of the coal-bearing Westphalian A and B north of the Wales–Brabant High. 45–78 in European Coal Geology. Whatley, M K, and Spears, D A (editors.). Special Publication of the Geological Society of London, No. 82.

Guppy, E M, and Thomas, H H. 1931. Chemical analyses of igneous rocks, metamorphic rocks and minerals. Memoir of the Geological Survey Great Britain.

Hall, I H S, Browne, M A E, and Forsyth, I H. 1998. Geology of the Glasgow district. Memoir of the British Geological Survey, Sheet 30E (Scotland).

Harper, D A T. 1979. The environmental significance of some faunal changes in the Upper Ardmillan succession (upper Ordovician), Girvan, Scotland. 439–445 in The Caledonides of the British Isles — reviewed. Harris, A L, Holland, C H, and Leake, B E (editors). Special Publication of the Geological Society of London, No. 8.

Harper, D A T. 1981. The stratigraphy and faunas of the Upper Ordovician High Mains Formation of the Girvan district. Scottish Journal of Geology, Vol. 17, 247–255.

Harper, D A T. 1982. The stratigraphy of the Drummuck Group (Ashgill), Girvan. Geological Journal, Vol. 17, 251–277.

Harper, D A T. 1984. Brachiopods from the Upper Ardmillan Succession (Ordovician), of the Girvan District, Scotland. Monograph of the Palaeontographical Society, Vol. 136, Part 1, 1–78.

Harper, D A T, Mitchell, W I, and Rong, J-Y. 1994. New faunal data from the highest Ordovician rocks at Pomeroy, County Tyrone, Northern Ireland. Scottish Journal of Geology, Vol. 30, 187–190.

Haughton, P D W, and Halliday, A N. 1991. Significance of a late Caledonian igneous complex revealed by clasts in Lower Old Red Sandstone conglomerates, central Scotland. Bulletin of the Geological Society of America, Vol. 103, 1476–1492.

Heddle, M F. 1901. The Mineralogy of Scotland. Volume 1. (Edinburgh: David Douglas.)

Higgs, K T, and O'Connor, G. 2005. Stratigraphy and palynology of Westphalian strata of the Leinster Coalfield, Ireland. Irish Journal of Earth Sciences, Vol. 23, 65–84.

Holden, W G. 1977. The glaciation of Central Ayrshire. Unpublished PhD thesis, University of Glasgow.

Howells, Y. 1982. Scottish Silurian trilobites. Monograph of the Palaeontographical Society. London, 1–76, pls 1–15 (publication No. 561, part of Vol. 135 for 1982).

Ince, D. 1984. Sedimentation and tectonism in the Middle Ordovician of the Girvan district, SW Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 75, 225–237.

Ingham, J K. 1978. Geology of a contintental margin 2: middle and late Ordovician transgression, Girvan. 163–176 in Crustal evolution in northwestern Britain and adjacent regions. Bowes, D R, and Leake, B E (editors). Geological Journal Special Issue, No. 10.

Ingham, J K. 1992a. Excursion 29. Upper Stinchar Valley and adjacent areas. 378–395 in Geological excursions around Glasgow and Girvan. Lawson, J D, and Weedon, D S (editors). (Glasgow: Geological Society of Glasgow.)

Ingham, J K. 1992b. Excursion 30. Girvan Foreshore. 396–416 in Geological excursions around Glasgow and Girvan. Lawson, J D, and Weedon, D S (editors). (Glasgow: Geological Society of Glasgow.)

Ingham, J K. 1992c. Excursion 31. The Craighead Inlier. 417–439 in Geological excursions around Glasgow and Girvan. Lawson, J D, and Weedon, D S (editors). (Glasgow: Geological Society of Glasgow.)

Institute of Geological Sciences. 1976. Cumnock. Scotland Sheet 14E. Solid Geology. 1:50 000. (Southampton: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1978a. Ayr. Scotland Sheet 14W Solid Geology. 1:50 000. (Southampton: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1978b. Ayr. Scotland Sheet 14W. Drift Geology. 1:50 000. (Southampton: Ordnance Survey for Institute of Geological Sciences.)

Jolly, R J H, and Sanderson, D H. 1995. Variation in the form and distribution of dykes of the Mull swarm, Scotland. Journal of Structural Geology, Vol. 17, 1543–1557.

Jones, N S. 2001. Sedimentology of the Swanshaw Sandstone Formation in Ayrshire. British Geological Survey Internal Report, IR/01/164.

Jones, O T. 1928. Plectambonites and some allied genera. Memoirs of the Geological Survey of Great Britain. Palaeo-ntology. Vol. 1, 367–527.

Kimbell, G S, Carruthers, R M, Walker, A S D, and Williamson, J P. 2006. Regional Geophysics of Southern Scotland and Northern England. Version 1.0 on CD-ROM. (Keyworth, Nottingham: British Geological Survey.)

Kokelaar, B P. 1982. Fluidization of wet sediments during the emplacement and cooling of various igneous bodies. Journal of the Geological Society of London, Vol. 139, 21–34.

Lamont, A. 1933. Transgression in the Craighead Limestone. Proceedings of the Geologists' Association, Vol. 44, 84–86.

Lamont, A. 1935. The Drummuck Group, Girvan; A Strati-graphical Revision, with descriptions of new fossils from the lower part of the group. Transactions of the Geological Society of Glasgow, Vol. 19, 288–334.

Lamont, A, and Lindstrom, M. 1957. Arenigian and Llandeilian cherts identified in the Southern Uplands by means of conodonts, etc. Transactions of the Edinburgh Geological Society, Vol. 17, 60–70.

Lapworth, C. 1882. The Girvan Succession. Quarterly Journal of the Geological Society of London, Vol. 38, 537–666.

Lawrie, J W, and Anderson, F W. 1953. Undersea borings off Prestwick. Transactions of the Geological Society of Glasgow, Vol. 21, 521.

Leake, R C, Cameron, D G, Bland, B J, Styles, M T, and Fortey, N J. 1997. The potential for gold mineralisation in the British Permian and Triassic red beds and their contacts with underlying rocks. Mineral Reconnaissance Report, British Geological Survey, No. 144.

Lele, K M, and Provan, D. 1962. Note on a Mississippian spore assemblage from Ayrshire, Scotland. Transactions of the Geological Society of Glasgow, Vol. 24, 287–289.

Le Maitre, R W. 1989 (editor). A Classification of Igneous Rocks and Glossary of Terms. Recommendations of the International Union of Geological Sciences Subcommission on the Systematics of Igneous Rocks. (Oxford, London, Edinburgh, Boston, Melbourne: Blackwell Scientific Publications.)

Longman, C D, Bluck, B J, and Van Breeman, O. 1979. Ordovician conglomerates and the evolution of the Midland Valley. Nature, London, Vol. 280, 578–580.

Lowe, J J, Birks, H H, Brooks, S J, Coope, G R, Harkness, D D, Mayle, F E, Sheldrick, C, Turney, C S M, and Walker, M J C. 1999. The chronology of palaeoenvironmental changes during the last Glacial–Holocene transition: towards an event stratigraphy for the Briish Isles. Journal of the Geological Society of London, Vol. 156, 397–410.

Lukševicˇs, E. 2001. Bothriolepid antiarchs (Vertebrata, Placodermi) from the Devonian of the north-western part of the East European Platform. Geodiversitas, Vol. 23, 489–609.

Macdonald, R, and Fettes, D J. 2007. The tectonomagmatic evolution of Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 97, 213–295.

Macdonald, R, Thomas, J E, and Rizello, S A. 1977. Variations in basalt chemistry with time in the Midland Valley province during the Carboniferous and Permian. Scottish Journal of Geology, Vol. 13, 11–22.

Macdonald, R, Wilson, L, Thorpe, R S, and Martin, A. 1988. Emplacement of the Cleveland Dyke: evidence from geo-chemistry, mineralogy and physical modelling. Journal of Petrology, Vol. 29, 559–583.

Macgregor, M, Lee, G W, and Wilson, G V. 1920. Special reports on the Mineral Resources of Great Britain Vol. 11. The Iron Ores of Scotland. Memoirs of the Geological Survey. Scotland.

Macpherson, H G. 1989. Agates. (British Museum (Natural History) and the National Museums of Scotland).

Macpherson, K A T, Smith, R A, and Akhurst, M C. 2000. Geology of the Kilmarnock district. Sheet Description of the British Geological Survey, 1: 50 000 Series Sheet 22E Kilmarnock (Scotland).

Mclean, A C. 1966. A Gravity Survey in Ayrshire and its Geological Interpretation. Transactions of the Royal Society of Edinburgh, Vol. 66, 239–265.

Mclean, A C, and Deegan, C E (editors). 1978. The solid geology of the Clyde Sheet (55N/6W). Report of the Institute of Geological Sciences, No. 78/9.

Mcmillan, A A, Hamblin, R J O, and Merritt, J W. 2005. An overview of the lithostratigraphical framework for the Quaternary and Neogene deposits of Great Britain (Onshore). British Geological Survey Research Report, RR/04/04.

Mcnestry, A. 1993. Palynology Report on 18 samples from 1:50k Sheet 15W New Cumnock. British Geological Survey Technical Report, WH/93/335R.

Miall, A D. 1996. The geology of Fluvial Deposits, Sedimentary Facies, Basin Analysis, and Petroleum Geology. (Berlin: Springer.)

Miles, R S. 1968. The Old Red Sandstone Antiarchs of Scotland: Family Bothriolepididae. Monograph of the Palaeontographical Society (London: The Palaeontological Society.)

Mitchell, W I (editor). 2004. The Geology of Northern Ireland — Our Natural Foundation. (Belfast: Geological Survey of Northern Ireland.)

Molyneux, S G. 1992. A palynological report on samples from the Silurian of the Lesmahagow and Hagshaw Hills inliers (Scottish 1:50 000 Sheet 23). British Geological Survey Technical Report, WH/92/237.

Molyneux, S G. 2001. Palynological investigation of three samples from the Straiton Group (Silurian) of the Girvan district. British Geological Survey Internal Report, IR/01/040.

Molyneux, S G, Barron, H F, and Smith, R A. 2008. Upper Llandovery–Wenlock (Silurian) palynology of the Pentland Hills inliers, Midland Valley of Scotland. Scottish Journal of Geology, Vol. 44, 151–168.

Monaghan, A A. 2001. Geology of the Culzean to Dipple area. British Geological Survey Internal Report, IR/01/165.

Monaghan, A. 2004. Geology of the Dailly area, Ayrshire. British Geological Survey Internal Report, IR/04/035.

Monaghan, A A, and Browne, M A E. 2010. Nine ⁴⁰Ar/³⁹Ar dates from Carboniferous igneous rocks of the Midland Valley of Scotland. British Geological Survey Open Report, OR/10/065.

Monaghan, A A, and Parrish, R R. 2006. Geochronology of Carboniferous–Permian magmatism in the Midland Valley of Scotland: implications for regional tectonomagmatic evolution and the numerical timescale. Journal of the Geological Society of London, Vol. 163, 15–28.

Monaghan, A A, and Pringle, M S. 2004. Ar/Ar geo-chronology of Carboniferous–Permian volcanism in the Midland Valley, Scotland. 219–241 in Permo-Carboniferous Magmatism and Rifting in Europe. Wlison, M, Neumann, E-R, Davies, G R, Timmerman, M J, Heeremans, M, and Larsen, B T (editors). Special Publication of the Geological Society of London, No. 223.

Monro, S K. 1999. Geology of the Irvine district. Memoir of the British Geological Survey, Sheet 22W and part of Sheet 21E (Scotland).

Muir, A, and Hardie, H G M. 1956. The limestones of Scotland: chemical analyses and petrography. Memoirs of the Geological Survey. Special Reports on the Mineral Resources of Great Britain, Vol. 37.

Murchison, R I. 1851. On the Silurian rocks of the South of Scotland. Quarterly Journal of the Geological Society of London, Vol. 7, 137–178.

Mykura, W. 1965. White trap in some Ayrshire Coals. Scottish Journal of Geology, Vol. 1, 176–184.

Mykura, W. 1967. The Upper Carboniferous rocks of south-west Ayrshire. Bulletin of the Geological Survey of Great Britain, No. 26, 23–98.

Neumann, E-R, Wilson, M, Heeremans, M, Spencer, E A, Obst, K, Timmerman, M J, and Kirstein, L. 2004. Carboniferous–Permian rifting and magmatism in southern Scandinavia, the North Sea and northern Germany: a review. 11–40 in Permo–Carboniferous Magmatism and Rifting in Europe. Wilson, M, Neumann, E-R, Davies, G R, Timmerman, M J, Heeremans, M, and Larsen, B T. (editors). Special Publication of the Geological Society of London, No. 223.

Neves, R, Guienn, K J, Clayton, G, Ioannides, N S, Neville, R S W, and Kruszewska, K. 1973. Palynological correlations within the Lower Carboniferous of Scotland and Northern England. Transactions of the Royal Society of Edinburgh, Vol. 69, 4–70.

Oliver, G J H. 2002. Chronology and terrane assembly, new and old controversies. 201–211 in The Geology of Scotland (4th edition). Trewin, N H (editor). (London: The Geological Society.)

Oliver, G J H, Smellie, J L, Thomas, L J, Casey, D M, Kemp, A E S, Evans, L J, Baldwin, J R, and Hepworth, B C. 1984. Early Palaeozoic metamorphic history of the Midland Valley, Southern Uplands–Longford–Down massif and the Lake District, British Isles. Transactions of the Royal Society of Edinburgh, Earth Sciences, Vol. 75, 245–258.

Oliver, G H J, Wilde, S A, and Wan, Y. 2008. Geochronology and geodynamics of Scottish granitoids from the late Neoproter-ozoic break-up of Rodinia to Palaeozoic collision. Journal of the Geological Society of London, Vol. 165, 661–674.

Owen, A W. 1986. The uppermost Ordovician (Hirnantian) trilobites of Girvan, SW Scotland with a review of coeval trilobite faunas. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 77, 231–239.

Paterson, I B, Hall, I H S, and Stephenson, D. 1990. Geology of the Greenock district. Memoir of the British Geological Survey, Sheet 30W and part of Sheet 29E (Scotland).

Paterson, I B, Mcadam, A D, and Macpherson, K A T. 1998. Geology of the Hamilton district. Memoir of the British Geological Survey, Sheet 23W (Scotland).

Peach, B N, and Horne, J. 1899. The Silurian Rocks of Britain, Volume 1. Scotland. Memoir of the Geological Survey of the United Kingdom. (Edinburgh: HMSO.)

Peacock, J D. 1995. Glacial Drift and Marine Shells. 40–35 in John Smith of Dalry, geologist, antiquarian and natural historian. Part 1 — Geology. Ayrshire Monographs, No. 16 (Ayrshire Archaeological and Natural History Society.)

Phillips, E R. 1999. Petrology of the igneous rocks exposed in the Ayr District (Sheet14W) of the Southern Midland Valley, Scotland. British Geological Survey Technical Report, WA/99/18.

Phillips, E R. 2000. Petrology of the Carrick Volcanic Formation and Heads of Ayr vent, Ayr (Sheet 14W), Southern Midland Valley, Scotland. British Geological Survey Technical Report, WA/00/65.

Phillips, E R. 2001. Petrographic descriptions of a number of thin sections of sandstone and igneous rocks from the Ayr district (Sheet 14W), Scotland. British Geological Survey Internal Report, IR/01/168.

Phillips, E R. 2002. Petrology of the sedimentary and igneous rocks exposed in the Ayr District (Sheet 14W), Scotland. British Geological Survey Internal Report, IR/02/076.

Phillips, E R, and Smith, R A. 2008. Petrology and geo-chemistry of the igneous and sedimentary rocks exposed in the Ayr District (Sheet 14W) of the Southern Midland Valley, Scotland. British Geological Survey Open Report, OR/08/058.

Phillips, E R, Smith, R A, and Carroll, S. 1998. Strike-slip, terrane accretion and the pre-Carboniferous evolution of the Midland Valley of Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 89, 209–224.

Pollard, J E. 1995. John Smith's Discoveries of Trace Fossils from Old Red Sandstone and Carboniferous rocks of Southwest Scotland. 30–39 in John Smith of Dalry, geologist, antiquarian and natural historian. Part 1 — Geology. Ayrshire Monographs, No. 16 (Ayrshire Archaeological and Natural History Society.)

Raines, M G, Greenwood, P G, Hallam, J R, Cripps, A C, and Shedlock, S L. 1997. Milton Wellfield Geophysical Survey. British Geological Survey Technical Report, WN/97/16C.

Read, W A. 1988. Controls on Silesian sedimentation in the Midland Valley of Scotland. 222–241 in Sedimentation in a synorogenic basin complex: The Upper Carboniferous of Northwest Europe. Besley, B M, and Kelling, G (editors). (Glasgow: Blackie.)

Read, W A. 1989. The interplay of sedimentation, volcanicity and tectonics in the Passage Group (Arnsbergian, E2 to West-phalian A) in the Midland Valley of Scotland. 143–152 in Arthurton, R S, Gutterridge, P, and Nolan, S C (editors). The Role of Tectonics in Devonian and Carboniferous Sedimentation in the British Isles. Occasional Publications of the Yorkshire Geological Society, No. 6.

Read, W A, Browne, M A E, Stephenson, D, and Upton, B G J. 2002. Carboniferous. 251–299 in Trewin, N H. (editor) The Geology of Scotland (4th edition). (London: The Geological Society.)

Reading, H G. 1996. Sedimentary environments: processes, facies and stratigraphy. (3rd edition). (Oxford: Blackwell Science.)

Reed, F R C. 1917. The Ordovician and Silurian Brachiopoda of the Girvan district. Transactions of the Royal Society of Edinburgh, Vol. 51, 795–998.

Reed, F R C. 1935. Palaeontological evidence of the age of the Craighead Limestone. Transactions of the Geological Society of Glasgow, Vol. 19, 340–372.

Richey, J E, Anderson, E M, and Macgregor, A G. 1930. The geology of north Ayrshire. Memoir of the Geological Survey, Scotland. Sheet 22.

Rippon, J H, Read, W A, and Park, R G. 1996. The Ochil Fault and the Kincardine Basin: key structures in the tectonic evolution of the Midland Valley of Scotland. Journal of the Geological Society of London, Vol. 153, 573–587.

Robertson, D, and Crosskey, H W. 1874. Post Tertiary Beds. Transactions of the Geological Society of Glasgow, Vol. 4, 255–256.

Robins, N S. 1990. Hydrogeology of Scotland. (London: HMSO.)

Rong, J Y, and Harper, D A T. 1988. A global synthesis of the latest Ordovician Hirnantian brachiopod faunas. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 79, 383–402.

Rushton, A W A, and Tripp, R P. 1979. A fossiliferous lower Canadian (Tremadoc) boulder from the Benan Conglomerate of the Girvan district. Scottish Journal of Geology, Vol. 15, 321–327.

Rushton, A W A, Owen, A W, Owens, R M, and Prigmore, J K. (editors). 1999a. Craighead Quarry. 342–344 in British Cambrian to Ordovician Stratigraphy. Geological Conservation Review Series, Vol. 18. (Peterborough: Joint Nature Conservation Committee.)

Rushton, A W A, Owen, A W, Owens, R M, and Prigmore, J K. (editors). 1999b. South Threave. 344–347 in British Cambrian to Ordovician Stratigraphy. Geological Conservation Review Series, Vol. 18. (Peterborough: Joint Nature Conservation Committee.)

Scott, A. 1915. The Crawford John Essexite and associated rocks. Geological Magazine (New series), Vol. 2, 455–461 and 513–519.

Scott, A C, Galtier, J, and Clayton, G. 1984. Distribution of anatomically-preserved floras in the Lower Carboniferous in Western Europe. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 75, 311–340.

Sedgwick, A. 1851. On the geological structures and relations of the frontier chain of Scotland. Notices and abstracts of miscellaneous communications to the sections. 103–107 in Report of the twentieth meeting of the British Association for the Advancement of Science held in Edinburgh in July and August 1850. (London: John Murray.)

Simpson J B, and Macgregor, A G. 1932. Economic Geology of the Ayrshire Coalfields, Area IV. Memoir of the Geological Survey, Scotland. (Edinburgh.)

Smith, D E. 1997. Sea-level change in Scotland during the Devensian and Holocene. 136–151 in Reflections on the Ice Age in Scotland — an update on Quaternary studies. Gordon, J E (editor). (Glasgow: Scottish Natural Heritage, Scottish Association of Geography Teachers.)

Smith, J. 1895 (for 1892). From the Doon to the Girvan Water, along the Carrick Shore. Transactions of the Geological Society of Glasgow, Vol. 10, 1–12.

Smith, J. 1898a. On a globular structure in a patch of shale enclosed in the Deil's Dyke near Greenan Castle, Ayrshire. Transactions of the Geological Society of Glasgow, Vol. 11, 128–129.

Smith, J. 1898b. The drift or glacial deposits of Ayrshire. Transactions of the Geological Society of Glasgow, Vol. 11 (supplement), 134 pp.

Smith, J. 1909. Upland fauna of the Old Red Sandstone Formation of Carrick, Ayrshire. (Kilwinning: A W Cross.)

Smith, J. 1910. Semi-precious Stones of Carrick. (Kilwinning: A W Cross.)

Smith, R A. 1995. The Siluro-Devonian evolution of the southern Midland Valley of Scotland. Geological Magazine, Vol. 132, 503–513.

Smith, R A. 1999a. Geology of the New Cumnock district. Sheet Explanation of the British Geological Survey. Sheet 15W (Scotland).

Smith, R A. 1999b. Geology of the Martnaham Loch area. British Geological Survey Technical Report, WA/99/66.

Smith, R A. 2000a. Geology of the Croy area. British Geological Survey Technical Report, WA/00/13.

Smith, R A. 2000b. Geology of the Dalrymple area. British Geological Survey Technical Report, WA/00/74.

Smith, R A. 2001. The geology of the Straiton–Spallander area, Ayrshire. British Geological Survey Internal Report, IR/01/148.

Smith, R A. 2003. The geology of the Kirkmichael area, Ayrshire. British Geological Survey Internal Report, IR/03/078.

Smith, R A, Jones, N S, Monaghan, A A, and Arkley, S. 2006. Fluvial and aeolian deposition in the Siluro–Devonian Swanshaw Sandstone Formation, SW Scotland. Scottish Journal of Geology, Vol. 42, 161–177.

Sowerbutts, A. 1999. Geology of the Dunure, Carrick Hills and Heads of Ayr area. British Geological Survey Technical Report, WA/99/102.

Sowerbutts, A. 2000. Geology of the south-west Maybole to Craigdow area. British Geological Survey Technical Report, WA/00/26.

Sowerbutts A, and Smith, R A. 1999. Geology of the Ayr and Prestwick area. British Geological Survey Technical Report, WA/99/30.

Speight, J M, Skelhorn, R R, Sloan, T, and Knaap, R J. 1982. The dyke swarms of Scotland. 449–459 in Igneous Rocks of the British Isles. Sutherland, D S (editor). (Chichester: Wiley.)

Stephenson, D, and Gould, D. 1995. British Regional Geology: the Grampian Highlands. Fourth edition. (London: HMSO for British Geological Survey.)

Stephenson, D, Loughlin, S C, Millward, D, Waters, C N, and Williamson, I T. 2003. Carboniferous and Permian Igneous Rocks of Great Britain North of the Variscan Front. Geological Conservation Review Series, No. 27 (Peterborough: Joint Nature Conservation Committee.)

Stephenson, M. 1999a. Report on palynology of samples from Abbothill, Friarland and Springwater Fishery, Ayr, Scotland. British Geological Survey Technical Report, WH/99/18R.

Stephenson, M. 1999b. Report on palynology of samples from Bracken Bay and East Heads of Ayr, Ayrshire. British Geological Survey Technical Report, WH/99/118R.

Stephenson, M. 1999c. Report on palynology of samples from Croy Foreshore, Ayrshire. British Geological Survey Technical Report, WH/99/128R.

Stephenson, M. 2000a. Report on palynology of samples from the Ballagan Formation, west of Heads of Ayr, Ayrshire. British Geological Survey Technical Report, WH/00/13R.

Stephenson, M. 2000b. Report on palynology of samples from the Ballagan Formation Doonfoot, Ayrshire. British Geological Survey Technical Report, WH/00/14R.

Stephenson, M. 2000c. Report on palynology of samples from the Ballagan Formation Dalrymple, Ayrshire. British Geological Survey Technical Report, WH/00/15R.

Stephenson, M. 2001a. Report on the palynology of samples from the Ballagan Formation north of Straiton, Ayrshire. British Geological Survey Technical Report, IR/01/11.

Stephenson, M. 2001b. Report on the palynology of samples from the Tournaisian of Ladyburn and Heads of Ayr, Ayrshire. British Geological Survey Technical Report, IR/01/45.

Stephenson, M. 2001c. Report on the palynology of samples from the Westphalian, Visean, and Tournaisian of Ayrshire. British Geological Survey Technical Report, IR/01/23.

Stephenson, M H. 2002. Algal palynomorphs of the Ballagan Formation and their palaeoenvironmental significance. British Geological Survey Internal Report, IR/02/182.

Stephenson, M, Williams, M, Monaghan, A, Arkley, S, and Smith, R. 2002. Biostratigraphy and palaeoenvironments of the Ballagan Formation (lower Carboniferous) in Ayrshire. Scottish Journal of Geology, Vol. 38, 93–111.

Stone, P. 1988. The Permian successions at Ballantrae and Loch Ryan, south-west Scotland. Report of the British Geological Survey, Vol. 19, No. 2, 13–18.

Stone, P, and Merriman, R J. 2004. Basin thermal history favours an accretionary origin for the Southern Uplands terrane, Scottish Caledonides. Journal of the Geological Society, London, Vol. 161, 829–836.

Stubblefield, C J. 1939. Some aspects of the distribution and migration of Trilobites in the British Lower Palaeozoic Faunas. Geological Magazine, Vol. 76, 49–72.

Sullivan, H J. 1968. A Tournaisian spore flora from the Cementstone Group of Ayrshire, Scotland. Palaeontology, Vol. 11, 116–131.

Sutherland, D G. 1999. Scotland. 99–114 in A revised correlation of Quaternary deposits in the British Isles. Bowen, D Q (editor). Special Report of the Geological Society of London, No. 23.

Syba, E. 1989. The sedimentation and provenance of the Lower Old Red Sandstone Greywacke Conglomerate, southern Midland Valley, Scotland. Unpublished PhD thesis, University of Glasgow.

The new statistical account of Scotland. 1845. Vol. 5. Ayr-Bute. (Edinburgh and London: William Blackwood and Sons.)

Thirlwall, M F. 1979. The petrochemistry of the British Old Red Sandstone volcanic province. Unpublished PhD thesis, University of Edinburgh.

Thirlwall, M F. 1981. Implications for Caledonian plate tectonic models of chemical data from volcanic rocks of the British Old Red Sandstone. Journal of the Geological Society of London, Vol. 138, 123–138.

Thirlwall, M F. 1982. Systematic variation in chemistry and Nd-Sr isotopes across a Caledonian calc-alkaline volcanic arc: implications for source materials. Earth and Planetary Science Letters, Vol. 58, 27–50.

Thirlwall, M F. 1986. Lead isotope evidence for the nature of the mantle beneath Caledonain Scotland. Earth and Planetary Science Letters, Vol. 80, 55–70.

Thirlwall, M F. 1988. Geochronology of Late Caledonian Magmatism in northern Britain. Journal of the Geological Society of London, Vol. 145, 951–967.

Traquair, R H. 1902. 81–83 in Hinxman, L W, and Grant-Wilson, J S. The geology of Lower Strathspey. Memoir of the Geological Survey, Sheet 85 (Scotland).

Trewin, N H, and Thirlwall, M F. 2002. The Old Red Sand-stone. 213–249 in: The Geology of Scotland (4th edition). Trewin, N H (editor). (London: The Geological Society.)

Tripp, R P. 1954. Caradocian Trilobites from Mudstones at Craighead Quarry, near Girvan, Ayrshire. Transactions of the Royal Society of Edinburgh, Vol. 62, 655–694.

Tripp, R P. 1980. Trilobites from the Ordovician Ardwell Group of the Craighead Inlier, Girvan district, Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 71, 147–157.

Turner, N. 1994. Palynology of Devonian and Carboniferous Rocks of the Cumnock and Doon Valley District, Strathclyde. British Geological Survey Technical Report, WH/94/294R.

Turner, N. 1995. Palynology of Tournaisian and Visean Rocks of the Cumnock and Doon Valley District, Strathclyde. British Geological Survey Technical Report, WH/95/238R.

Tyrrell, G W. 1912. The Late Palaeozoic Igneous Rocks of the West of Scotland. Geological Magazine, Vol. 9, 69–8 and 120–131.

Tyrrell, G W. 1914. A petrographical sketch of the Carrick Hills, Ayrshire. Transactions of the Geological Society of Glasgow, Vol. 15, 64–83.

Tyrrell, G W. 1917. Some Tertiary dykes of the Clyde area. Geological Magazine, Vol. 4, 305–315 and 350–356.

Tyrrell, G W. 1920. The igneous geology of the Ayrshire Coast from Doonfoot to the Heads of Ayr. Transactions of the Geological Society of Glasgow, Vol. 16, 339–363.

Tyrrell, G W. 1928a. A further contribution to the petrography of the Late Palaeozoic igneous suite of the west of Scotland. Transactions of the Geological Society of Glasgow, Vol. 18, 259–294.

Tyrrell, G W. 1928b. On some Dolerite-sills containing Analcite-Syenite in Central Ayrshire. Quarterly Journal of the Geological Society, London, Vol. 84, 540–569.

Tyrrell, G W. 1933. Maidens, Dunure, and the Heads of Ayr. Summer Field Meeting, 1932. Proceedings of the Geologist' Association, Vol. 44, 72–74.

Underhill, J R, Monaghan, A A, and Browne, M A E. 2008. Controls on structural styles, basin development and petroleum prospectivity in the Midland Valley of Scotland. Marine and Petroleum Geology, Vol. 25, 1000–1022.

Upton, B G J, Stephenson, D, Smedley, P M, Wallis, S M, and Fitton, J G. 2004. Carboniferous and Permian magmatism in Scotland. 195–218 in Permo-Carboniferous Magmatism and Rifting in Europe. Wilson, M, Neumann, E-R, Davies, G R, Timmerman, M J, Heeremans, M, and Larsen, B T (editors). Special Publication of the Geological Society of London, No. 223.

Vann, I R. 1978. The siting of Tertiary vulcanicity. 393–414 in Crustal evolution in northwestern Britain and adjacent regions. Bowes, D R, and Leake, B E (editors). Geological Journal Special Issue, No. 10.

Wagner, R H. 1966. On the presence of probable Upper Stephanian beds in Ayrshire, Scotland. Scottish Journal of Geology, Vol. 2, 122–123.

Wagner, R H. 1983. A lower Rotliegend flora from Ayrshire. Scottish Journal of Geology, Vol.19, 135–155.

Walker, E F. 1985. Arthropod ichnofauna of the Old Red Sandstone at Dunure and Montrose, Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 76, 287–297.

Wallis, S M. 1989. Petrology and geochemistry of Upper Carboniferous — Lower Permian Volcanic rocks in Scotland. Unpublished PhD thesis, University of Edinburgh.

Ward, P M. 1989. A palaeoenvironmental interpretation of the Craighead Inlier and Woodland Point, (Lower Silurian), Girvan, Southwest Scotland. Unpublished PhD thesis, University of Edinburgh.

Wellman, C H, and Richardson, J B. 1993. Terrestrial plant microfossils from Silurian inliers of the Midland Valley of Scotland. Palaeontology, Vol. 36, 155–193.

Whyte, F. 1963. The Heads of Ayr Vent. Transactions of the Geological Society of Glasgow, Vol. 25, 72–97.

Whyte, F. 1992. Excursion 24. Heads of Ayr. 295–300 in Geological excursions around Glasgow and Girvan. Lawson, J D, and Weedon, D S (editors). (Glasgow: Geological Society of Glasgow.)

Williams, A. 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan district, south-west Ayrshire, with descriptions of the Brachiopoda. Memoir of the Geological Society of London, No. 3.

Williams, D M, and Harper, D A T. 1988. A basin model for the Silurian of the Midland Valley of Scotland and Ireland. Journal of the Geological Society of London, Vol. 145, 741–748.

Williams, M. 2001. The stratigraphical palaeontology of the Craighead Inlier, Girvan District, south-west Scotland. British Geological Survey Internal Report, IR/01/029.

Williams, M, and Floyd, J D. 2000. Mid-Caradoc (Ordovician) ostracodes from the Craighead Limestone Formation, Girvan district, SW Scotland. Scottish Journal of Geology, Vol. 36, 51–60.

Williams, M, and Taylor, P. 2001. Ostracods from the Ballagan Formation (Tournaisian; lower Carboniferous), Ayrshire, south-west Scotland. British Geological Survey Technical Report, IR/01/31.

Williams, M, Stephenson, M H, Wilkinson, I P, Leng, M J, and Miller, C G. 2005. Early Carboniferous (late Tournaisian-early Visean) ostracods from the Ballagan Formation, central Scotland, UK. Journal of Micropalaeontology, Vol. 24, 77–94.

Williams, M, Leng, M J, Stephenson, M H, Andrews, J E, Wilkinson, I P, Siveter, D J, Horne, D J, and Vannier, J M C. 2006. Evidence that Early Carboniferous ostracods colonised coastal flood plain brackish water environments. Palaeogeography, Palaeoclimatology, Palaeoecology, Vol. 230, 299–318.

Williamson, I T. 2003a. East Fife Coast, Fife. 155–165 in Carboniferous and Permian Igneous Rocks of Great Britain north of the Variscan Front. Stephenson, D, Louglin, S C, Millward, D, Waters, C N, and Williamson I T (editors). Geological Conservation Review Series, No. 27. (Peterborough: Joint Nature Conservation Committee.)

Williamson, I T. 2003b. Howford Bridge, East Ayrshire. 166–169 in Carboniferous and Permian Igneous Rocks of Great Britain north of the Variscan Front. Stephenson, D, Louglin, S C, Millward, D, Waters, C N, and Williamson I T (editors). Geological Conservation Review Series, No. 27. (Peterborough: Joint Nature Conservation Committee.)

Williamson, I T, and Monaghan, A A. 2003. Heads of Ayr, South Ayrshire. 104–108 in Carboniferous and Permian Igneous Rocks of Great Britain north of the Variscan Front. Stephenson, D, Louglin, S C, Millward, D, Waters, C N, and Williamson I T (editors). Geological Conservation Review Series, No. 27. (Peterborough: Joint Conservation Committee.)

Wilson, H M, and Anderson, L I. 2004. Morphology and taxonomy of Palaeozoic millipedes (Diplopoda: Chilognatha: Archipolypoda) from Scotland. Journal of Paleontology, Vol. 78, 169–184.

Wilson, R B. 1967. A study of some Namurian marine faunas of central Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 66, 445–490.

Wilson, R B. 1983. Note on the correlation of the Upper Limestone Group (Namurian) in south Ayrshire. Scottish Journal of Geology, Vol. 19, 183–188.

Wilson, R B. 1989. A study of the Dinantian marine macro-fossils of central Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 80, 91–126.

Winchester, J A, and Floyd, P A. 1977. Geochemical discrimination of different magma series and their differentiation products using immobile elements. Chemical Geology, Vol. 20, 325–343.

Wintersgill, M T. 2004. The John Smith Agate Collection and its relevance to current theories of agate formation. Senior Thesis Internship Option Submitted to the Faculty of Baylor University in partial fulfilment of the requirements for the Degree of Batchelor of Science. (Texas, USA).

Wood, A. 1933. On three trilobites from Girvan. Appendix A. Summer Field Meeting, 1932; Girvan–Ballantrae. Proceedings of the Geologists' Association, Vol. 44, 81–84.

Figures, plates and tables

Figures

(Figure 1) Simplified map of bedrock geology for the Ayr district.

(Figure 2) Plan of Craighead Quarry showing distribution of members in the Craighead Limestone Formation and its relation to the Craighead Volcanic Formation and the Ardwell Farm Formation. Modified after Rushton et al., 1999a.

(Figure 3) Map of the distribution of Ordovician and Silurian rocks in the Craighead Inlier (adapted from the 1:50 000 Bedrock map).

(Figure 4) Simplified map of bedrock geology showing the distribution of the Lanark Group (Swanshaw Sandstone and Carrick Volcanic formations) near the south-west Ayrshire coast.

(Figure 5) Palaeocurrent rose diagrams for the Swanshaw Sandstone Formation at well-exposed localities. Most data derived from measurements of sets of fluvial channel cross-bedding with a few measurements (less than 5 per cent) from primary current lineation or cross-lamination. Data from aeolian dunes, crevasse splays and wave ripples are annotated. After Smith et al., 2006.

(Figure 6) Palaeogeographical diagrams illustrating successive depositional phases A to G (youngest) during the accumulation of the Swanshaw Sandstone Formation. After Smith et al., 2006.

(Figure 7) Geochemical discrimination plots for Siluro-Devonian lavas (top left) Zr/TiO₂-Nb/Y classification diagram (Winchester and Floyd, 1977); (top right) SiO₂-Zr/TiO₂ classification diagram (Winchester and Floyd, 1977); and (below) K₂O-SiO₂ classification diagram (Le Maitre, 1989).

(Figure 8) Summary geological map of the coastal section from Bracken Bay to Longhill Point with palynology and ostracod sample localities (modified after Stephenson et al., 2002).

(Figure 9) Graphical section of the upper Ballagan Formation including the Lindsayston Burn Member, from the Lindsayston Burn, Dailly [NS 2821 0082] to [NS 2737 0129].

(Figure 10a) Graphical section of the junction of the Kinnesswood and Ballagan formations exposed in March Burn at [NS 376 070], with the horizons of palynology samples (MPA numbers) indicated. From Smith (2001, fig 3).

(Figure 10b) Graphical section within the upper Ballagan Formation from the Troquhain area [NS 378 093] with horizons of palynology samples (MPA numbers) indicated. From Smith (2001, fig 4).

(Figure 11) Superimposed channel and scour features in the Clyde Sandstone Formation, looking east at [NS 2894 0433], in Glenton Glen. From Monaghan (2004, fig 6).

(Figure 12) Summary of geological structure of the Dailly area, with trends of interpreted syndepositional thickening for the Lawmuir (LWM) and Limestone Coal (LSC) formations. Map, sketch and strain ellipse to right of figure shows structures acting under dextral transtension from the Mid Mississipian to Pennsylvanian. From Monaghan (2004, fig 1).

(Figure 13) Summary lithostratigraphy of Carboniferous to Permian strata including coals and thicknesses in the Ayr–Prestwick area. From Sowerbutts and Smith (1999, fig 2).

(Figure 14) Graphical log towards the top of the Middle Coal Measures Formation, above the Ayr Ell Coal, taken from Brockle Wood cliffs, near Auchencruive [NS 3976 2362]. From Sowerbutts and Smith (1999, fig 5).

(Figure 15) Distribution of the Permo-Triassic rocks in south-west Scotland and offshore westwards to Arran and Northern Ireland.

(Figure 16) Detailed plan of Heads of Ayr Volcanic Vent showing its relationship to the country rock.

(Figure 17) Structures, including major fold and fault traces, within the district.

(Figure 18) Bouger gravity anomaly for the Ayr district. Contour interval = 2.5 mGal. White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS.

(Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS." data-name="images/P1000716.jpg">(Figure 19) Aeromagnetic anomaly map of the Ayr district compiled from three generations of survey dated 1959 to 2004 (see (Table 5)). White line marks the coastline. Heavy black rectangle marks the limit of the Ayr district, which lies within grid square NS.

(Figure 20a) Subsurface contour map of the Ayr Hard Coal Seam in the Ayr district. From a 3D digital model based on borehole, mining and outcrop data.

(Figure 20b) Subsurface contour map of the base of the Lower Limestone Formation in the Ayr district. From a 3D digital model based on borehole, mining and outcrop data.

(Figure 21) 3D digital block model image of the Ayrshire Coalfield and Mauchline Basin.

(Figure 22a) Simplified Quaternary map of the Ayr district showing lithodemic units. (b) NEXTMap BritainTM image for the Ayr district. The topography is dominated by an easterly trending pattern of drumlins and ice-moulded rock-knolls.

(Figure 22b) NEXTMap BritainTM for the Ayr district. The topography is dominated by an easterly trending pattern of drumlins and ice-moulded rock-knolls.

(Figure 23) Distribution and thickness of Quaternary sediment offshore with vertical cross-section through offshore geology (adapted from 1:250k Clyde Sheet 55N 06W Sea Bed Sediments and Quaternary map).

(Figure 24) Map showing the approximate extent of undermining, and location of collieries mentioned in the text, around Ayr and Prestwick. Based on Sowerbutts and Smith (1999, fig 1).

Plates

(Plate 1) Plantinhead Member (Ardwell Farm Formation) with calcareous sandstone forming a waterfall above graptolitic black mudstones, in Plantinhead Burn, Craighead, near Girvan [NS 2338 0157]. P507302.

(Plate 2) Fossil trilobite fossil Lonchodomas drummuckensis from South Threave, Craighead Inlier [NS 2500 0380], probably from the Farden Member of the South Threave Formation. P064573.

(Plate 3) Horizontally-stratified conglomerate interbedded with planar laminated sheet sandstone, of the Channel Facies Association A. Isle Port Conglomerate Member, Swanshaw Sandstone Formation at Isle Port [NS 2453 1297]. Tape measure extended to 1.5 m. P818147.

(Plate 4) Trough cross-bedded conglomerate of the Channel Facies Association A. Isle Port Conglomerate Member, Swanshaw Sandstone Formation at Isle Port [NS 2453 1297]. This face illustrates the distinctly troughed form of these sets. The set arrowed is approximately 1 m high. P818148.

(Plate 5) Cross-bedded sandstone in the Channel Facies Association A of the Swanshaw Sandstone Formation at Dead Knowe [NS 2466 1208]. Note the low-angle and flaggy nature of many of the sets, particularly at the top of the face (arrowed), indicating that they formed in conditions close to upper flow regime. Face is approximately 8 m high. P818149.

(Plate 6) The Isle Port Conglomerate Member, Swanshaw Sandstone Formation illustrating the multistorey nature of the channel facies in the Isle Port area [NS 2460 1302]. The bounding surface, at the channel base marked by the arrows, has significant erosive relief. The face is approximately 7 m high. P818150.

(Plate 7) Aeolian dune cross-bedded sandstone of the Aeolian Facies Association B, Chapelton Burn Member, Swanshaw Sandstone Formation in Chapelton Burn gorge section [NS 2286 0489]. Note coarser grainflow foreset lamination (arrowed) reaching 3 cm in thickness. P818151.

(Plate 8) 'Pinstripe' laminated sandstone in the Aeolian Facies Association B. Chapelton Burn Member, Swanshaw Sandstone Formation. Chapelton Burn gorge section [NS 2284 0489]. Note the thinly bedded nature of the sandstone and the internal pinstripe lamination. Deposition on a dry (aeolian) sandsheet is indicated for this facies, with the pinstripe lamination formed by the migration of wind ripples. Pencil is 130 mm long. P818152.

(Plate 9) Wavy laminated damp sandsheet facies in the Aeolian Facies Association B. Chapelton Burn Member, Swanshaw Sandstone Formation. Chapelton Burn gorge section [NS 2284 0489]. Note the thinly bedded nature of the sandstone and the faint wavy to convoluted nature of the internal laminae. P818153.

(Plate 10) Floodplain Facies Association C, Chapelton Burn Member, Swanshaw Sandstone Formation at Chapelton Burn [NS 2284 0489]: sheet flood sandstones with sharp bed bases (arrowed) and wave reworked parts (bracketed). Black scale bar is 50 mm in length. P818154.

(Plate 11) Mudflat Facies Association D interbedded with sheetflood sandstone (Facies C2), Crane Dyke Member, Swanshaw Sandstone Formation, at Isle Port [NS 2453 1372]. The sandstone has a wave rippled top with symmetrical straight crested ripples. The pale mudstone clasts in the bed below form part of a mudstone breccia. P818155.

(Plate 12) Mudflat Facies Association D1, Crane Dyke Member, Swanshaw Sandstone Formation at Isle Port foreshore [NS 2453 1368] with horizontal U-shaped Beaconites burrows with retrusive backfilled structures. P818156.

(Plate 13a) Photo-micrograph of pilotaxitic basaltic andesite with sieve texture in zoned plagioclase microphenocrysts. Carrick Volcanic Formation Sample N2158 from Brown Carrick Hill [NS 2865 1588] taken in plane polarised light, objective lens X2.5. P516445.

(Plate 13b) Photo-micrograph of pilotaxitic basaltic andesite with sieve texture in zoned plagioclase microphenocrysts. Carrick Volcanic Formation Sample N2158 from Brown Carrick Hill [NS 2865 1588] taken with crossed polars, objective lens X2.5. P516446.

(Plate 14a) Photo-micrograph of plagioclase-clinopyroxene-orthopyroxene macrophyric basaltic andesite with glomerophyric sieve textured plagioclase and pyroxene. Carrick Volcanic Formation Sample N2162 from Brown Carrick Hill [NS 2875 1662] taken in plane polarised light, objective lens X2.5. P516451.

(Plate 14b) Photo-micrograph of plagioclase-clinopyroxene-orthopyroxene macrophyric basaltic andesite with glomerophyric sieve textured plagioclase and pyroxene. Carrick Volcanic Formation Sample N2162 from Brown Carrick Hill [NS 2875 1662] taken with crossed polars, objective lens X2.5. P546452.

(Plate 15) Intercalated pink-brown sandstone within basalt and basaltic andesites of the Carrick Volcanic Formation, with a low dip to the north-west. The igneous sheet has disrupted the sedimentary rocks at its base. Looking south to Turnberry Lighthouse and beyond to Ailsa Craig from [NS 2000 0750]. P571636.

(Plate 16a) Photo-micrograph of medium-grained calcareous sandstone from the Stratheden Group Sample N2057 from east Carlandcheek [NS 2775 1808] taken in plane polarised light; objective lens X2.5. Quartz grains clear; mudstone fragment dark brown. P516461.

(Plate 16b) Photo-micrograph of medium-grained calcareous sandstone from the Stratheden Group Sample N2057 from east Carlandcheek [NS 2775 1808] taken with crossed polars; objective lens X2.5. P516462.

(Plate 17) Ballagan Formation (Tournaisian) dark grey silty mudstone gently dipping north-east below thin-bedded rippled sandstones on the foreshore at [NS 3062 1925] about 50 m west of Deil's Dyke. P703202.

(Plate 18) Bedded grey coarse-grained volcaniclastic sandstone and conglomerate of the Greenan Castle Pyroclastic Member, Clyde Plateau Volcanic Formation dipping east-north-east on the shoreline at [NS 3118 1933] below Greenan Castle. Camera case 100 mm long. P703198.

(Plate 19) Heads of Ayr Vent (latest Tournaisian to early Asbian) forming twin headlands, taken from [NS 3070 1931] near the Deil's Dyke, south-west of Ayr. P703204.

(Plate 20) Photo-micrograph of slightly reworked, altered lithic-rich lapilli-tuff with quartz (clear). Sample N2058 from the eastern end of Heads of Ayr Vent [NS 2930 1878] taken in plane polarised light, objective lens X2.5. P516453.

(Plate 21) Vertical and near horizontal jointing in Deil's Dyke, a prominent Palaeogene dolerite dyke on the foreshore at [NS 3070 1931] west of Longhill Point. P703200.

(Plate 22) Open syncline in the Ballagan Formation (Tournaisian) picked out by the more prominent sandstone beds, plunging westwards out to sea, viewed from [NS 3060 1922] about 100 m west of Deil's Dyke, south-west of Ayr. P703203.

(Plate 23) Holocene raised beach between Greenan Castle and Deil's Dyke, photograph taken from Greenan Castle [NS 312 193], with Heads of Ayr in the distance. P601190.

(Plate 24) Former Auchalton Quarry, 800 m south-east of Auchalton [NS 335 036] in soft argillaceous sandstones with horizons of pedogenic carbonate nodules (middle of section) over a worked seam of pedogenic nodular limestone (cornstone) at base. Quarry and mine already disused when the photo was taken in 1926 by W Manson, Geological Survey. P216842.

(Plate 25a) Onyx agate amygdale, consisting of chalcedony in the form of parallel-banded onyx and concentrically-banded agate from John Smith's collection obtained during excavations in the Carrick Volcanic Formation for the Dunure Railway. The pink, white and grey bands represent different layers laid down at different times by fluids percolating through the lava and precipitating layers in the open cavities or vesicles, originally formed as gas bubbles. P521520.

(Plate 25b) Onyx agate with an outer rim of fine banded agate and onyx with quartz crystals projecting into an open 'drusy' cavity or geode. Collected during excavations for the Dunure Railway from the Carrick Volcanic Formation. P521517.

(Front cover) Culzean Castle (National Trust for Scotland) partly constructed of local Siluro-Devonian Swanshaw Sandstone Formation.(P601213.

(Rear cover)

Tables

(Table 1) Stratigraphical column for the Lower Palaeozoic successions in the Craighead and Girvan Main inliers of the Ayr district. Not to scale. pt = part. (adapted from Fortey et al., 2000, Williams, 2001 and Floyd and Williams, 2003).

(Table 2) Lithostratigraphical position of the Benan Conglomerate Formation in the Ayr district relative to the Craighead inlier and the North Stinchar area (Carrick district). Fm=Formation.

(Table 3) Lithostratigraphy of the Siluro-Devonian Lanark Group in the south-west Midland Valley with depositional settings (dashed lines indicate gaps in exposure).

(Table 4) Depositional environments of observed rock types in Lower and Middle Coal measures of the Ayr district

(Table 5) Aeromagnetic survey data collected over the Ayr district

(Table 6) Quaternary onshore lithostratigraphy in the Ayr district and correlation with adjacent areas

Tables

(Table 4) Depositional environments of observed rock types in Lower and Middle Coal Measures of the Ayr district

Rock type	Depositional environment
Black fissile mudstone with Lingula sp.	Marine incursion
Uniform grey mudstones and siltstones with mussels, plant debris, thin coals	Lacustrine or overbank
Silty seatrock with rootlets, thin coals and abundant plant debris	Palaeosol
Coal, extensive with abundant plant remains	Mire
Finely interbedded siltstones and fine-grained sandstones with rippled/trough/erosional lamination, abundant plant fragments	Distal lacustrine delta
Sharp-based, decimetre-scale, fine-grained sandstones, in places upwards coarsening, micaceous	Lacustrine delta or crevasse splay
1 to 5 m thick erosive-based sandstones with cross-bedding, trough-bedding and superposed metre-scale channels	Minor channel
Nodular layers of limestone within siltstones and mudstones	Diagenetic precipitation or calcrete formation in overbank deposits

(Table 5) Aeromagnetic survey data collected over the Ayr district

	Flight line orientation	Flight line spacing (m)	Tie line spacing (m)	Nominal flying height (m)
BGS CA59	N–S	2000	10 000	305
JEBCO/ARK	N–S	250	2500	80/170¹
BGS HIRES Ayrshire	E–W	200	—	56/250²
1 Offshore/onshore 2 Rural/urban