• British Regional Geology
• The Midland Valley of Scotland, 3rd edition, 1985
• Memoirs
• The economic geology of the Central Coalfield of Scotland, Area I, 1937, 1920, Area V, 1926 and Area VII, 1920
• Geology of the Greenock district, 1990 Geology of the Airdrie district, 1996
• Geology of the Ben Lomond district, in press Geology of the Hamilton district, 1997
• BGS Report
• Lithostratigraphy of the late Devonian and early Carboniferous rocks in the Midland Valley of Scotland. Vol. 18, No. 3, 1986

• 1:625 000
• United Kingdom (North Sheet) Solid geology, 1979
• Quaternary geology, 1977 Aeromagnetic anomaly, 1972 Bouguer anomaly, 1981
• Regional gravity, 1981
• 1:250 000
• Argyll (Sheet 56N 06W) Solid geology, 1987 Aeromagnetic anomaly, 1981, Bouguer anomaly, 1979
• Clyde (Sheet 55N 06W) Solid geology, 1985  Aeromagnetic anomaly, 1980 Bouguer anomaly, 1985
• Tay-Forth (Sheet 56N 04W) Solid geology, 1986 Aeromagnetic anomaly, 1981 Bouguer anomaly, 1979
• Borders (Sheet 55N 04W) Solid geology, 1986 Aeromagnetic anomaly, 1980 Bouguer anomaly, 1981
• 1:50 000
• Sheet 30E (Glasgow) Solid, 1994; Drift, 1995
• Sheet 22 (Kilmarnock) Solid, 1928; Drift, 1928
• Sheet 23 (Hamilton) Solid, 1929; Drift, 1929
• Sheet 23W (Hamilton) Drift, 1995; Solid, 1995
• Sheet 31W (Airdrie) Solid, 1992; Drift, 1992
• Sheet 38W (Ben Lomond) Solid, 1987
• Sheet 30W (Greenock) Solid, 1990; Drift, 1989

It is generally considered that deposition of the Lower Devonian sequence took place in an elongate, NE-trending basin, centred on the Midland Valley and bounded on the north-west by the Highlands (Bluck, 1978; Morton, 1979; Armstrong et al., 1985). Sediment derived from the Highlands was laid down along the northern basin margin in alluvial fans and subsequently partly dispersed south-westwards along the basin by axial drainage from the north-east. The fluvial sandstones and siltstones of the Teith Formation were deposited during one of the last phases known in the long and complex history of the Silurian-Devonian basin when the depositional environment was predominantly one of meandering streams on a broad floodplain.

After a lengthy period of widespread earth movements accompanied by uplift and erosion, the regional Lower Devonian palaeoslope was reversed and Upper Devonian sedimentary rocks were laid down in a major ENE-trending depositional basin dominated by eastward-flowing rivers (Figure 3) . During deposition of the Stratheden Group strata, the rainfall was probably moderate and somewhat seasonal. The presence of aeolian sandstones throughout the upper part of the Stockiemuir Sandstone Formation suggests that the rainfall decreased during this period. However, there would also be a predisposition to the development of aeolian sandstones since land vegetation was not as yet well developed. The water-laid sandstones in the western part of the district were probably laid down largely in channels on a extensive flood-plain by meandering rivers. However, the finer-grained sandstones seen near Croftamie may represent the deposits of local short-lived flood basins. A return to exclusively fluvial sedimentation is seen in the overlying Kinnesswood Formation.

Classification of all except the highest part of the Lower Carboniferous has been discussed, type sections erected and formal names defined by Paterson and Hall (1986) and Chisholm et al. (1989). This part of the succession includes all the strata up to base of the Hurlet Limestone, traditionally taken as the top of the former Calciferous Sandstone Measures. The overlying dominantly argillaceous cyclic sequences with sandstones, generally thin limestones and a few coals, which occur between the base of the Hurlet Limestone and the top of the Top Hosie Limestone, has been renamed the Lower Limestone Formation and assigned to the Clackmannan Group (Forsyth et al., 1996).

The Upper Carboniferous starts with the Limestone Coal Formation. The lower part of this formation comprises a sequence of dominantly argillaceous cyclothems with sandstones, mudstones, seatrocks, coals and ironstones. The upper part of the formation consists of dominantly arenaceous cyclothems. Two marine bands and several Lingula bands are present but there are few marine limestones. Stratigraphically above, is a succession of dominantly arenaceous cyclothems with mudstones, marine limestones and thin coals forming the Upper Limestone Formation. The succeeding strata, which are mainly fluvial in origin with thin marine bands, comprise the Passage Formation. The overlying fluviodeltaic coal-bearing sequence is divided into Lower, Middle and Upper Coal Measures at the Vanderbeckei (Queenslie) and Aegiranum (Skipsey's) marine bands respectively. The Lower Limestone, Limestone Coal, Upper Limestone and Passage formations have been assigned to the Clackmannan Group. The Lower, Middle and Upper Coal Measures are considered to be informal units with formation status and are placed in the Coal Measures, which has group status (Forsyth et al., 1996).

The Carboniferous is divided into two subsystems, the Dinantian (Lower Carboniferous) and Silesian (Upper Carboniferous). The base of the Dinantian is internationally defined by the presence of particular marine fossils (Paproth et al., 1991). These fossils have not been found in Scotland and consequently the base cannot be precisely defined. Miospores are the only fossils of any biostratigraphical significance which occur in the lowest Carboniferous in the district. The Ballagan Formation is placed in the CM miospore zone (Neves et al., 1972), the oldest that can be recognised in Scotland. At least two lower zones are recognised elsewhere and these are thought to be present in the underlying unfossiliferous Kinnesswood Formation. Hence, at least part of this formation is considered to be of early Carboniferous age.

The base of the Silesian is taken about 1 m below the Top Hosie Limestone at the lowest recorded occurrence of the goniatite Cravenoceras. The Silesian is divided into the Namurian and Westphalian series which in turn are subdivided into stages, though only some of the diagnostic goniatites are present in Scotland. The base of the Westphalian is defined by the presence of Gastrioceras subcrenatum. This goniatite has not been found in Scotland and the base of the Lower Coal Measures has been taken at a suitable arbitrary position. Within the Glasgow district this is the Lowstone Marine Band (Table 1) . Miospore evidence (Neves et al., 1965) suggests, however, that one of the marine bands in No. 6 Marine Band group (not known in the district) in the Passage Formation is the equivalent of the C. subcrenatum Marine Band and consequently, that the base of the Westphalian may lie within the upper part of this formation.

The strata of this formation comprise commonly variegated, purple-red, yellow, white and grey-purple sandstones with nodules and thin beds of concretionary carbonate (cornstone) and a few thin beds of red mudstone. No fossils have been found in these strata within this district but to the west, at Fairy Knowe Quarry [NS 369 789] near Dumbarton, disarticulated scales of Bothriolepis and Holoptychus of Famennian age have been found in the basal conglomerate (Aspen, 1974). However, general considerations suggest that the Kinnesswood Formation is mostly of early Carboniferous age (Paterson and Hall, 1986). These rocks were laid down on a broad alluvial plain by a generally eastward-flowing river system which occupied a large part of the Midland Valley at this time. The cross-bedded sandstones were deposited in the river channels and the cornstones formed as carbonate nodules in soil profiles that developed on the associated floodplains under the influence of a fluctuating water table in a semi-arid climate. Since argillaceous overbank deposits are rarely present in this district, the cornstones occur in sandstones at the top of channel-fill sequences and formed after the channels had been abandoned. The cornstones range from immature, in which the sandstones have a partly carbonate matrix with ill-defined isolated concretions, to mature, in which well-defined nodules (glaebules) are elongated in a vertical sense and are overlain by laminar and pisolitic structures (Plate 2) . The laminar structures, which develop subparallel to the bedding, may be brecciated and the carbonate replaced by chert. Where silicification occurs, the fabric of the cornstone is preserved and the mineralogical change appears to be an integral part of the pedogenic process. Analyses of cornstones from just to the west of the district ( (Table 2) , (Figure 5) ) suggest that the carbonate is largely dolomite, probably secondary, though limestone is also known to occur. Since there is very little sediment deposition during the formation of cornstone profiles, it is unlikely that there was sufficient meteoric water to supply enough magnesium for the formation of primary dolomite. There is no evidence of a potential source in the underlying strata; the only other source is seawater. It also seems unlikely that the argillaceous nature of the overlying Ballagan Formation would allow sufficient groundwater to percolate downwards, particularly after lithification, to effect the secondary dolomitisation of the cornstones. However, mature cornstones develop over a lengthy period (up to 1 million years, Machette, 1985) during which they are exposed at the surface under a regime of low enough rainfall to allow mostly calcium carbonate to be deposited. A high rate of evaporation may have caused lateral movement of magnesium-rich ground-water, derived either directly from saline water in marginal marine areas where early Ballagan Formation strata were being deposited or indirectly from unlithified sediments during their early diagenesis, to effect dolomitisation of the cornstone profiles. A good example of a mature cornstone profile 2 m thick and traceable for 0.5 km occurs at Overton Burn [NS 437 777] .

Outcrops of the formation only occur north of the River Clyde and are largely confined to two areas. One is an area of high moorland north of the Kilpatrick Hills and the other lies on the eastern side of Strath Blane extending north from Blanefield to the northern margin of the district. Faulting makes accurate assessment of thickness of the formation across the district difficult but estimated values are shown in (Figure 4) . In general they appear to be greater than the 170 m which was encountered in the Kipperoch Borehole [NS 3727 7742] to the west of the district, where a complete sequence was drilled.

On Dumbarton Muir, north of the Kilpartick Hills, good exposures of Kinnesswood Formation occur, mostly in stream sections. In this area the formation reaches more than 250 m but is variable in thickness and character. The section in and adjacent to Finland Burn [NS 450 818] - [NS 432 800] is comprised of poorly bedded, fine- to medium-grained grey- and red-purple sandstones interbedded with paler, coarser-grained cross-bedded sandstones. Many of the finer-grained sandstones contain carbonate concretions that may pass up into cornstones, though the latter are less common in the upper part of the sequence. The coarser-grained sandstones, which are interpreted as channel-fill deposits, have sharp or erosional bases and are frequently cross bedded. In the lower part of the sequence these sandstones may have cornstone clasts near their bases, whereas in the upper part they frequently have quartz pebbles up 10 cm in diameter. The finer-grained sandstones may be overbank deposits or channel deposits which have lost their original structures when soil profiles developed in the fill of abandoned channels. The cornstones are 0.3-1 m thick and some show partial silicification. Several of these were quarried and burnt locally. Millstones were quarried from one of the coarser sandstones on Dumbarton Muir [NS 444 804] . In contrast, less than 2 km to the east, in the Gallangad [NS 455 814] - [NS 450 783] and Knockupple [NS 452 802] - [NS 458 786] burns, cornstones are rare in the lower part of the sequence but well developed in the upper part. Quartz pebbles also occur near the bases of the coarser-grained sandstones throughout the sequence. Farther east on Stockiemuir [NS 48 81] and on the west side of Strath Blane, cornstones are not seen and fine-grained sandstones, where present, show few concretionary textures.

On the eastern side of Strath Blane, good sections of the Kinnesswood Formation are exposed in several streams draining off the Campsie Fells. In this area the formation is about 170 m thick near Blanefield and thickens northward to 400 m near Finny at the western (lid of the Gargunnock Hills (Francis et al., 1970). At the northern end of this area on Ballikinrain Muir [NS 56 86] , the formation is in two generally upward-lining sequences, The lower cycle, which is about 280 m thick, comprises coarse-grained, pink and purple-red, pebbly sandstones interbedded with grey- and red-purple line-grained sandstones and cornstones. The coarser-grained sandstones become less frequent and the cornstones better developed higher in the sequence. Here, the pebbles are of quartz, quartzite and schist. The upper cycle, which is 50 m thick, consists of white, coarse-grained siliceous sandstone with lenses of small quartz pebbles, overlain by pale grey, line-grained sandstones interbedded with subordinate greenish siltstones and cornstones. Several sandstones were formerly quarried for millstones in this area. Both cycles change in character when traced southward towards Blanefield. The coarse pebbly facies at the base of the lower cycle is replaced by about 60 m of pale purplish or greenish grey, medium-grained, massive, cross-bedded or flaggy sandstone. This is overlain by alternations of red-purple coarse-grained sandstones with quartz pebbles and finer-grained sandstones with cornstones, the latter becoming more frequent near the top of the cycle. The upper cycle commences with 10 m of white coarse gritty, siliceous sandstone with frequent quartz pebbles up to 15 cm in diameter and is followed by 20 m of fine- to medium-grained white sandstone, commonly calcareous, with mature cornstones and a few interbeds of greenish grey mudstone.

South of the River Clyde, a complete sequence was drilled in the BGS Glenburn Borehole [NS 4783 6066] where the thickness was found to be 73 m. Here, mature cornstones are not well developed, though some small carbonate concretions are present at intervals throughout the lower 10 m. Mudstone clasts and small quartz pebbles are present throughout the sequence, usually being more abundant at or near the base of sandstone units. Many of the medium and coarse-granted sandstones are cross-laminated, fine upwards and have erosive bases, features typical of channel-fill deposits.

The succeeding Ballagan Formation, named after the type section in Ballagan Glen [NS 510 800] (Plate 3) , is well exposed at the western end of the Kilpatrick Hills and along the western and northern flanks of the Campsie Fells. Between these two areas the formation is largely cut out by faulting. The sequence in the two areas is very similar. Typically, the strata consist of grey, poorly laminated silty mudstones with frequent thin nodular beds of dolomitic limestone (cementstone) and a few thin, fine-grained sandstones occurring at intervals throughout the sequence. At the base of the formation, more frequent sandstone beds are intercalated among the mudstones in a zone tip to 20 m thick, which is transitional with the underlying Kinnesswood Formation. Reddening, which is normally associated with the sandstones, can also affect the mudstones and cementstones locally.

South of the River Clyde, the Glenburn Borehole [NS 1783 6066] proved the Clyde Plateau Volcanic Formation lying directly on the Kinnesswood Formation (Figure 4) , strata of the Ballagan Formation and the overlying Clyde Sandstone Formation having been removed by erosion during the mid-Dinantian (Forsyth et al., 1996).

The cementstones occur as thin beds of nodular carbonate or as layers of discrete ellipsoidal nodules elongated parallel to the bedding. Though rarely exceeding 0.2 m in thickness, individual beds may persist laterally for distances of over 2 km. Most have no apparent internal structure but some are brecciated, and a few are laminated. Septarian cracks commonly occur and are mostly infilled with crystalline dolomite which rarely is accompanied by copper and lead minerals. Cavities are common and may be empty or lined with gypsum. In one case the gypsum was encrusted with copper minerals. All the cementstones analysed from this district were found to be dolomitic ( (Table 2) ; (Figure 5) ), as was the case in adjacent districts (Paterson et al., 1990; Francis et al., 1970). The occurrence of evaporitic minerals (see below), and the lack of other potential sources, points to the magnesium in the dolomite being derived from seawater which has been concentrated by evaporation in restricted saline lakes. It has been suggested that the laminated cementstones were deposited as carbonate muds (Belt et al., 1967) and that the nodular beds are diagenetic in origin, probably formed early and at shallow depth (Andrews et al., 1991).

The mudstones are normally grey, poorly laminated and may have desiccation cracks. They are patchily calcareous in places and, where seen in borehole core, have frequent gypsum nodules and are cut by numerous veins of fibrous gypsum up to 8 cm thick. It is noticeable that much more gypsum is present in strata drilled in boreholes (e.g. Barnhill and Loch Humphrey) than in nearby exposures. This is thought to be due to gypsum having been dissolved out by recent surface weathering. The abundance of gypsum veins may be due to the conversion of anhydrite to gypsum since, as suggested by Shearman et al. (1972), this change involves an increase in volume of 63 per cent in the solid phase and gypsum veins occurring in association with gypsum nodules represent, at least in part, the additional volume created by this conversion. The veins were considered to be emplaced by hydraulic fracture. Also seen in the cores are brecciated zones which may have been caused by collapse after the solution of thin salt deposits (? halite and anhydrite). Halite pseudomorphs do occur at intervals in the mud-stones and are also present in two of the fine-grained sandstones. Scott (1986) recorded textural and mineralogical features from the Ballagan Formation of Berwickshire which he interpreted as evidence for a former sulphate evaporite facies. He also pointed out that the deposition of gypsum would reduce the amount of dissolved calcium thus increasing the relative amount of magnesium in the brine and promoting the requisite Mg/Ca ratio for the formation of dolomite. A little pyrite and plant debris are also present in the mudstones.

Thin fine-grained sandstones which may show ripple cross-lamination or cross-bedding, occur at intervals. A few are very micaceous and poorly cemented but the majority are more massive and have a carbonate cement. The cross-laminated sandstones are small channel deposits, the remainder may be sheet-flood deposits.

The only fossils which have been found in the Ballagan _. Formation in the district are scattered occurrences of 'Estheria' sp. and ostracods which were recorded in BGS Barnhill [NS 4269 7571] and Loch Humphrey [NS 4592 7555] boreholes, and Spirorbis sp. and ostracods in surface exposures.

The dolomitic cementstones are thought to have been deposited in coastal sabkhas, restricted saline lakes or lagoons frequently replenished by seawater. The frequent repetition of cementstones interbedded with mud-stones, siltstones and sandstones, deposited as terrigenous sediments on a mature coastal alluvial plain dominated by mudflats, forms the characteristic element of a sequence of strata unique in the Scottish Carboniferous.

A return to a fluviatile environment is seen in the Clyde Sandstone Formation, the youngest division of the Inverclyde Group. In the Glasgow district, outcrops of this formation occur in geographically isolated areas north of the River Clyde. Good exposures occur at the western end of the Kilpatrick Hills and at intervals flanking the Campsie Fells with only a few fault-bounded outcrops between. South of the River Clyde, in the Gleniffer Braes area, the formation has been removed by mid-Dinantian uplift and erosion (Forsyth et al., 1996).

The sandstones are coarser grained than those of the Kinnesswood Formation and though commonly calcareous and concretionary, only locally develop into cornstones. Analyses ( (Table 2) ; Francis et al., 1970) and field examination suggest that all the carbonate is calcite rather than the dolomite which characterised the underlying formations. In the more northerly areas the sandstones are conglomeratic with quartz and Highland pebbles in places. Elsewhere, the pebble content is dominated by intraformational clasts of mudstone or limestone. Farther south there are no clasts present. Thus the strata of this formation represent the deposits of a wide variety of fluviatile environments, ranging from braided stream to floodplain with well-developed overbank.

West of the Kilpatrick Hills the formation is represented by the Overtoun Sandstone Member. The complete sequence was penetrated by the Barnhill Borehole where it was 57 m thick. The lower part is dominated by grey, fine-grained sandstones which are thin bedded or partly concretionary. There are also some grey mudstones which contain bands or concretions of limestone, and erosive-based coarser-grained sandstones with carbonate clasts. The upper part consists largely of sandstones with erosional bases and containing carbonate clasts; concretionary sandstones and mudstones are uncommon. Exposures occur in the Overtoun Burn [NS 423 760] .

Between this area and Strath Blane, faulted outcrops occur on Auchineden Hill [NS 485 802] , [NS 497 809] and Dumgoyach Brae [NS 525 806] . On Auchineden Hill, about 6 m of white gritty sandstone are exposed, containing abundant quartz pebbles up to 6 cm in diameter. On Dumgoyach Brae there is a 12 m sequence of brownish trough cross-bedded sandstone with bands of quartz pebbles; some of the sandstones are concretionary.

Exposures of the formation also occur along the northern and western flanks of the Campsie Fells. However, on the south side, in Campsie Glen [NS 610 800] and Fin Glen [NS 600 799] , lavas of the overlying Clyde Plateau Volcanic Formation lie directly on the Ballagan Formation, strata of the Clyde Sandstone Formation having been removed during the mid-Dinantian (Forsyth et al., 1996). In Ballagan Glen [NS 572 801] , to the west of these localities, the formation is about 12 m thick but the section is now largely inaccessible. It consists of white siliceous sandstones interbedded with siltstones and mudstones. A seatearth with numerous coal streaks was formerly seen (Macgregor et al., 1925) near the top of the section. Clasts appear to be absent in this area. Along the east side of Strath Blane, the formation is poorly exposed but farther north there is a good section in Little Conic [NS 576 844] where the sequence is about 45 m thick. The lower part consists of coarse-grained pebbly sandstone passing up into commonly ripple-laminated, micaceous, finer-grained sandstones, grey siltstones and mudstones with desiccation cracks. Some of the sandstones are concretionary with a calcareous matrix. The bulk of the clasts are quartz with some quartzite. This is in marked contrast to the sandstones to the north-east in the Stirling district where the clasts are entirely of limestone (Read and Johnson, 1967).

The distribution and types of clast suggest a southerly or south-easterly flow direction. This is in keeping with the south-south-easterly current direction found by Read and Johnson (1967) for the equivalent strata in the Stirling district.

Rocks of the Clyde Plateau Volcanic Formation occur north of Glasgow in a zone about 6 km wide running north-east from Dumbarton forming the high ground of the Kilpatrick Hills and the Campsie Fells. South of the River Clyde they occur in a small area at Old Kilpatrick, in the Beith–Barrhead Hills, and in the Cathkin Braes (Figure 1) . Each of these areas forms a discrete, largely fault-bounded lava block with its own stratigraphy and petrographical characteristics (Macdonald, 1975). The lava sequences are all thought to be of similar age but correlation between the blocks is limited. Along the northern and western margins of the Campsie Fells the basal pyroclastic rocks abruptly overlie the Clyde Sandstone Formation but along the south margin the underlying sandstone is progressively cut out in an easterly direction between Spout of Ballagan [NS 572 802] and Campsie Glen [NS 610 800] . In the Kilpatrick Hills the base is transitional in the south-west [NS 426 750] and in the Loch Humphrey Borehole [NS 458 755] where thin sandstones occur in the lowest part of the formation, but farther north the base appears to be disconformable. In the Gleniffer Braes the basal volcaniclastic rocks lie directly on the Kinnesswood Formation, with the Clyde Sandstone and Ballagan formations both absent.

The top of the formation is seen along the eastern flanks of the Kilpatrick Hills where it is overlain by volcanic detritus of the Kirkwood Formation in places. Elsewhere it is overlain by sandstone or conglomerate of the Lawmuir Formation. There is considerable variation in preserved thicknesses in the different areas though the sequences are thought to be penecontemporaneous. The small area of lavas at. Old Kilpatrick, which is the top part of the Renfrewshire Hills sequence, is thought to be underlain by up to 1000 m of lavas (Paterson et al., 1990). The sequence in the Campsie Fells, where the top is not preserved, is more than 500 m thick, that in the Kilpatrick Hills greater than 400 m and that in the Beith–Barrhead area less than 300 m. The various sequences are thought to result from one major volcanic episode which took place within a relatively short space of time at the beginning of the Viséan. The rocks of the Clyde Plateau Volcanic Formation are mostly lava with varying proportions of pyroclastic deposits and volcaniclastic rocks in different areas.

Interflow volcaniclastic material forms a very small proportion of the lava pile in distal facies sequences but the proportion increases rapidly around active vents where tephra cores are developed comprising tuffs, agglomerates and breccias, with interbedded highly vescicular lava tongues. In distal facies sequences, the lava sheets, which are mostly massive in character, are separated by weathered zones. Typically, these weathered zones have a crudely stratified layer of scoriaceous volcanic debris overlain by fine-grained red-brown bole-like material. In most cases, this bole-like material has ghost stratification and is thought to be tuff rather than weathered lava top. In proximal facies sequences, fragmental rock with blocks, scoria bombs, spatter and stratified tuffaceous material form a high proportion of the sequence. The lava flows in this facies are highly vesicular and commonly compound, comprising two or more flow units. The texture of the lavas is coarser than that found in distal facies lavas.

A large number of associated vents, plugs and minor intrusions occur in the Campsie Fells, the Kilpatrick Hills and also cutting the older strata lying to the north. Many of these volcanic centres were active contemporaneously and supplied most, if not all, of the lavas in these areas and also perhaps, the topmost part of the Renfrewshire Hills sequence seen at Old Kilpatrick. The source is not known of the lavas in the Beith–Barrhead Hills and in the Cathkin Braes.

The Kirkwood Formation consists of sandstone, siltstone, mudstone and some conglomerate composed of weathered detritus derived from the upstanding volcanic terrain formed by the Clyde Plateau Volcanic Formation. The strata are generally red, brown or purple in colour. In the Glasgow district, the formation mainly occurs at the surface south of the River Clyde, in the ground south of Johnstone and Paisley. Its thickness is variable and largely uncertain, but it may reach 30 m in this area.

North of the river it crops out around Douglas Muir [NS 525 745] . To the west of Douglas Muir [NS 508 748] , the formation, which is about 15 m thick, overlies the Clyde Plateau Volcanic Formation and is in turn overlain by the Douglas Muir Quartz-Conglomerate Member. To the east of Douglas Muir [NS 532 745] , the relationship with the overlying conglomerate is the same but here, the Kirkwood Formation is underlain by fine-grained sandstone interbedded with seatearth of the Lawmuir Formation. Farther east and to the north on Craigmaddie Muir [NS 565 780] , the formation is absent and conglomerate lies directly on lavas. To the south, volcaniclastic rocks of the formation have been found in a few boreholes such as Lawmuir [NS 5183 7310] , where they are 15 m thick.

Generally, the Kirkwood Formation immediately overlies the lavas of the Clyde Plateau Volcanic Formation, from which it is clearly derived, and filled hollows in the top surface of the eroded and subsiding lava terrain. However, upstanding lava terrain occurred at intervals to supply volcaniclastic detritus, at least locally, after the onset of fluvial and fluviodeltaic sedimentation which laid down the Lawmuir Formation.

In most of the district, the Lawmuir Formation overlies the Kirkwood Formation. However, north of the River Clyde, the lower strata of the formation are interbedded or contemporaneous with the Kirkwood Formation in areas where volcaniclastic detritus was not being supplied to the basin.

The Lawmuir Formation belongs mainly to the Brigantian Stage of the Dinantian but some of the lowest beds may be Asbian in age. The top of the formation is now drawn at the base of the Hurler Limestone, not the Hurlet Coal. The formation was previously known as the Upper Sedimentary Group of the Calciferous Sandstone Series or Measures. It is present over most of the central and southern parts of the district, but is little known except at or near its outcrops (Figure 9) . The larger of these extends in a WSW–ENE belt across the centre of the district from Bishopton through Clydebank and Milngavie; the other occupies most of the Paisley area. At its type locality, the Lawmuir Borehole [NS 5183 7310] , the Lawmuir Formation is fully 250 m thick. This appears to be a fairly typical thickness, but at Paisley the formation reaches 330 m before thinning rapidly to 130 m at Johnstone and only 12 m at Howwood, just beyond the western margin of the district.

The Lower Limestone Formation forms the highest part of the Brigantian sequence and belongs to the VF Miospore Zone. It is present over most of the southern part of the district. The formation thickens markedly eastwards from 60 m near the western margin of the district to a reported maximum of 179 m in an underground borehole [NS 5694 6264] at Titwood No. 2 Pit. It also thickens southwards from 110 m in Balmore No. 2 Borehole [NS 6013 7468] to 165 m in Milngavie No. 5 Borehole [NS 5735 7285] . A detailed description of the stratigraphy is given in Forsyth (1993).

The Limestone Coal Formation is the lowest part of the Namurian sequence and belongs entirely to the Pendleian (E ₁ ) Stage. It occurs in most of the south-east part of the district and also on the north-west side of the Paisley Ruck from Johnstone to Renfrew. The formation can be divided into two parts at the top of the Black Metals, a development of mudstones about 20 m thick.

The Upper Limestone Formation crops out in a sinuous belt from Barrhead in the south-west through central Glasgow northwards to Balmore and thence south-west to north-west Glasgow, and also in a few outliers about which little is known. It belongs partly to the Pendleian (E ₁ ) Stage of the Namurian and partly (from the Orchard Limestone upwards) to the Arnsbergian (E ₂ ) Stage. These two parts (Forsyth, 1982) show contrasting successions (Figure 12) .

The Passage Formation where fully developed is over 300 m thick and contains both Namurian and Westphalian A strata, but in this district it is thin and largely unfossiliferous. The formation occurs in south and east Glasgow, where some of the sandstones are exposed on Necropolis Hill [NS 65 60] , and near Cadder [NS 58 73] north of Glasgow, where only the lowest beds are preserved unexposed beneath thick Quaternary deposits. Their presence near Cadder is known only from Cadder No. 6 Borehole, which cut more than 40 m of sandstone and fireclay, plus a limestone, 0.35 m thick With 1.5 m of mudstone above it, that may be the Roman Cement Limestone.

Neither the Castlecary Limestone at the top of the underlying Upper Limestone Formation, nor the Low-stone Marine Band at the base of the overlying Lower Coal Measures, has been recognised in south and east Glasgow. The thickness of the Passage Formation therefore is uncertain but appears to be about 85 m, for instance in a borehole [NS 6039 6240] in Govan No. 5 Pit (Forsyth, 1961, p.229). It consists mainly of sandstones, some of which are coarse grained with quartz pebbles up to 2.5 cm across.

Unbedded mudstones, usually mottled red, purple, green and yellow, also occur; some are seatclays and others may merit being called fireclays. Bedded mudstones and coals are rare and thin. The only record of a marine band is in Hampden Park No. 9 Borehole [NS 5889 6152] in which a limestone, 0.25 m thick, and the overlying mudstone yielded a sparse fauna, including productoid brachiopods, which suggests that it belongs to the Passage Formation, probably to No. 0 or No. 1 Marine Band.

The Passage Formation, therefore, is a dominantly fluvial sequence, full of minor unconformities. The depositional basin probably contracted at this time, almost completely excluding the sea, and subsidence was reduced so that the basin was repeatedly filled with fluvial sand, thus minimising the opportunities for coal-forming plant debris to escape destruction below the water table. The basal unconformity cut down into the Upper Limestone Formation to about Plean No. 2 Limestone and repeatedly river channels were cut across the area as sand was alternately deposited and eroded. Away from these channels mud locally and intermittently accumulated on the floodplains.

The Lower Coal Measures belong to the Langsettian (Westphalian A), the lowest part of which is either absent or undefinable in the Passage Formation. They occur unexposed in south and east Glasgow, cut off to the southwest by the Dechmont Fault. Dips are generally to the south-east. Only a few borehole records are available; none covers the whole sequence. The stratigraphy (Forsyth, 1979) is still imperfectly known (Figure 13) . The thickness is about 100 m. The top is well defined by the base of the Vanderbeckei (Queenslie) Marine Band, but the Lowstone Marine Band at the base has not been found.

The Middle Coal Measures comprise the whole of the Duckmantian (Westphalian B). They occur in a limited area in south and east Glasgow, bounded to the southwest by the Dechmont Fault. The strata mostly dip to the south-east but in the east there is a reversal to WNW. Surface exposures are few. The stratigraphy is derived from numerous boreholes, mostly put down for urban development, plus a few sections from shafts put down to exploit the thick coals, mining of which has long since ceased. The sequence is 160 m thick (Forsyth and Brand, 1986), which is less than is found in most Scottish coalfields. It contains an unusually high proportion of mud and silt, and rich nonmarine faunas; both features indicate a tendency to return to the quiet backwater conditions of late Dinantian and early Namurian times, in a basin subsiding at a moderate rate. Fluvial sandstones are rare. After the initial marine transgression, represented by the Vanderbeckei (Queenslie) Marine Band (Brand, 1977), there were only two brief recurrences of quasi-marine conditions. The stratigraphy up to the Glasgow Upper Coal is shown in (Figure 14) . Most of the coals are thinner than they are in the Airdrie district to the east (Forsyth et al., 1996), but the Virgin and Glasgow Upper coals are notable exceptions. The Coatbridge Mussel band is the most poorly developed of the nonmarine bivalve hands. The Cambuslang Marble lies unusually close to the Glasgow Main Coal but is well developed, with its unusual combination of Euestheria and nonmarine bivalves, the latter being notably different from those in the lower bands. The strata above the Glasgow Upper Coal are rather different and less well known. Faunas are sparse, channel sandstones occur and coals are thin; only one very small working is known. The Euestheria band which characteristically occurs above the Glasgow Upper Coal in the Airdrie district is poorly developed, especially in Govanhill, where there is a mudstone 10-12 m thick, near the top of which sparse occurrences of Lingula mytilloides represent the Glasgow Upper Marine Band. The Drumpark Marine Band is known only as a Lingula band in Hamilton Road Route Borehole No. 20 [NS 6093 6287] . It is the only other marine horizon present. The Shettleston Sandstone is over 20 m thick in the Prospecthill Borehole [NS 5999 6184] , where it cuts down to 2 m above the Glasgow Upper Coal.

The Upper Coal Measures occur in a very limited area in south-east Glasgow. They lie in a broad syncline and are cut off to the south by the Dechmont and Rutherglen faults. Only the lower part (about 100 m thick) is present; it belongs entirely to the Bolsovian (Westphalian C). There are scarcely any exposures and the only borehole of consequence is Prospecthill [NS 5999 6184] , where the sequence consists of 3 m of mudstone at the base, containing the Aegiranum (Skipsey's) Marine Band; the Westmuir Sandstone (60 m thick and mainly fine to medium grained but with medium-to coarse-grained bands); and, at the top, 23 m of mainly argillaceous rocks, including a thin coal and containing the Bothwell Bridge Marine Band. Some of the strata are red or purple stained, but the staining is not intense and some beds retain their original pale grey colour. The fauna of the Aegiranum Marine Band at Prospecthill is typical of the western end of the Central Coalfield (Forsyth and Brand, 1986).

Outwash deposits related to the advance of the Dimlington Stadial Ice Sheet occur on the south side of the Kelvin valley in the Cadder area (BGS, 1994, section C). Here sand and gravel of the Cadder Formation rest on the Baillieston Till Formation and are in turn overlain by the Wilderness Till Formation. The poorly known subsurface sands and gravels and interbedded diamicton beds under the floodplain of the River Kelvin are also assigned to the Cadder Formation. Within the area of the Kelvin's bedrock depression the formation may be at least 44 m thick as illustrated by the Cadder No. 16 Borehole [NS 6002 7326] .

In the Cadder area, although exposures were formerly plentiful in workings for sand and gravel, there are now only two sections remaining where the Cadder Formation can be seen. The typical Ethological assemblages consist of mainly framework-supported bouldery gravel and sand, and coarse- to fine-grained, sometimes pebbly, sand with silt. Locally a thin clay and silt bed of possible lacustrine origin (? Broomhill Formation) rests between the sand and gravel and the overlying Wilderness Till Formation. Based on observations from sections in the top 30 m of the formation, much of the gravel is thickly bedded and displays trough cross-bedding in sets up to 3 m thick. These interfinger with sands which are trough cross-bedded, ripple laminated and horizontally laminated.

A recent temporary section in the Wilderness Gravel Pit, Bishopbriggs [NS 595 725] , is illustrated in (Figure 19) , together with an earlier record by Rolfe (1966, fig. 2) and the record of a recent borehole. The base of the Wilderness Till Formation is complex with shearing and incorporation of the underlying gravel. Locally there is a merging basal contact, sometimes with a basal deposit consisting of a matrix-supported sandy pebble gravel. Elsewhere, planar or even stepped basal contacts are developed. This diamicton is a deformation and lodgement till.

Faulting, folding and shearing of the Cadder Formation sediments below may be due in part to glacial overriding by the ice sheet that deposited the overlying Wilderness Till Formation. The faulting may also partly reflect the presence of buried masses of ice which have subsequently melted. Clay-filled fissures penetrating the top of the formation in the Wilderness Pit were observed by Rolfe (1966). These features were said to be common in this area and were regarded as ice-wedge casts which formed under periglacial conditions.

The precise status of the reddish brown and brown diamictons assigned to the Buchley Till Member in the Wilderness Pit (Figure 19) is uncertain. The base of this till is sharp and undulating (10 cm relief) and its top is erosive and planar. Possibly it formed during the ice advance which was responsible for the deposition of both the Cadder and Wilderness Till formations. The diamicton is thought to be glacial in origin and marks a possible local advance–retreat cycle within the earliest late Devensian. It may be much older (Baillieston Till Formation?). The underlying Cadder Formation sediments are folded, one 10 cm S-fold exhibiting a sheared core. All the beds between 6.50 and 8.40 m depth are folded into an east-trending anticline, the beds dipping to the north at 20-40°, and to the south at 10-20°.

From the Wilderness Pit [NS 6006 7230] , Rolfe (1966, pp.253-258) described the finding of bones of the woolly rhinoceros. One of the bones collected provided a radiocarbon date of 27 550 BP. This date is compatible with recently reported dates from a stratigraphically similar position from Sourlie [NS 338 414] , near Irvine (Jardine et al., 1988).

The Cadder Formation was formerly seen in a sandpit south-west of Bearsden Station [NS 5390 7148] , in a gravel pit east of Bearsden Station [NS 5441 7183] and in a gravel pit at Ferguston Hill [NS 5530 7223] . In each case the Wilderness Till overlies the granular deposits, being 3 to 4 m thick at the first site but only 1 m at the other two. Between 9 and 15 m of mainly large-scale trough cross-bedded sand with highly contorted clay bands were to be seen in the pit south-west of Bearsden Station. There the base of the overlying till consists of cemented gravel and is rich in boulders. Only 1 or 2 m of cross-bedded sand and gravel were seen at the other two localities.

The sedimentological and faunal evidence both point to deposition of the Cadder Formation in a periglacial environment with the possibility that some of the sediments are ice-contact deposits. The sediments appear to be fluvideltaic in origin, based on the scale of the trough cross-bedded units. Overall the deposits appear to form a major outwash system laid down in front of the advancing Dimlington Stadial Ice Sheet about 27 000 years ago. The out-wash deposits were subsequently overridden by the ice which deposited the overlying Wilderness Till Formation.

Pre-Dimlington Stadial lacustrine deposits are found in various areas around Glasgow. Subsurface data and temporary sections reveal that laminated muds rest discontinuously upon the Baillieston Till. Accorded the name Broomhill Formation, these sediments were proved in the BGS Erskine Bridge Borehole (Figure 20) , which is taken as the standard section for this formation (from 16.31 to 32.75 m depth). Its stratigraphical position between the two tills is clearly demonstrated here. The typical lithology is thinly bedded silty clay with wisps, laminae and bands of silt and sometimes sand. The deposit is reddish brown in colour with grey or buff silt and sand. Isolated clasts, up to 14 cm, are considered to be dropstones. Thin diamicton bands and patches occur locally, with sonic graded bedding (turbidity flows) and micro-faulting. Some polished surfaces and steeply dipping microfaults may indicate glacial disturbance by overriding ice. The laminations are probably varves, but with multiple layers, and may collectively represent 600 to 1000 years of sedimentation.

In Broomhill Park No. 2 Borehole [NS 5985 6650] the formation (12.3-21.0 m) consists of dark brownish grey clays with reddish brown laminae and bands. Dips of greater than 30° are recorded in places and wisps and bands of diamicton (laminated in part) are common. The varved couplets are estimated to represent 1000 to 1500 years of sedimentation. Polished surfaces, joints and faults, together with folded bedding, all suggest glacial overriding.

Two sections in the New City Road area of Glasgow [NS 5799 6743] to [NS 5809 6744] ; [NS 5781 6747] to [NS 5790 6753] were recorded by E Hull in 1869. His notes describe up to 4 m of 'upper' boulder clay on what he said was 'in all probability the ordinary brickclay' (that is the Paisley Formation). The notes indicate that the upper till was indistinguishable from the ordinary older boulder clay (that is the Wilderness Till Formation) and the brick clay elsewhere rests on it. In the first section crumpling of the lamination of the clay was recorded at the west-end, with 'boulder clay inextricably confused'. It seems likely these structures are glacitectonic in origin and that the clays are part of the Broomhill Formation locally preserved below the Wilderness Till (that is the 'upper boulder clay'). The lower boulder clay is therefore part of the Baillieston Till Formation.

The evidence from sections and boreholes in the Broomhill Formation points to the deposition of the laminated muds in the form of varves representing seasonally controlled sedimentation in a temporary lake occupying part of the Clyde valley. Lake shorelines are estimated to have been in excess of 70 m above OD. Isolated clasts, interpreted as dropstones, and rafts of diamicton testify to the proximity of ice and it appears likely that the lake developed at some stage during the build-up of the Dimlington Stadial Ice Sheet. An estimate of between 600 and 1500 years has been made for the duration of lake sedimentation. The Broomhill Formation, as previously noted, may also be present in the Kelvin valley but the data are equivocal.

Evidence from the Cadder area indicates that the build-up of the Dimlington Stadial Ice Sheet began at about 27 500 BP. At its maximum, around 18 000 BP, the ice sheet was probably over 1 km thick in central Scotland. The direction of ice movement is recorded by moulded land-forms such as drumlins, some of which are rock-cored, and by crag-and-tail features, and by striae. Movement is generally from west to east with local variations from a northwesterly direction (Figure 18) . In the Glasgow district the easterly movement of ice is confirmed by the distribution cone of erratic blocks of the distinctive Lennoxtown essexite (nepheline-monzogabbro) (Peach, 1909; Shakesby, 1976). The exposures of the essexite are located just to the east of Clachan of Campsie in the Airdrie district.

The principal deposit of the Dimlington Stadial Ice Sheet is the Wilderness Till Formation. Following the classification of Rose (1981) and Rose et al. (1988), the formation is named after the Wilderness Plantation north of Bishopbriggs [NS 605 720] . In this area the formation comprises a hard, massive, reddish brown, sandy, silty diamicton which rests with stepped but low-angle disconformity on the cross-bedded sands of the Cadder Formation. At Drumry Wood [NS 5093 7162] the contact was seen striking 350° and dipping 30°E. The diamicton contains clasts, the more elongate of which are preferentially oriented, and the deposit usually has systematic sets of joints. The colour of the matrix varies, depending upon the nature of the local bedrock. Minor features noted in the Wilderness Till include pockets and beds of medium-grained sand and thin beds of laminated clay. In the Cadder area, temporary sections in the Wilderness Till revealed that the diamicton is locally graded, displaying both upward and downward coarsening of the dispersed pebble- to boulder-sized clasts. Also recognised were zones within the formation in which the clasts become far less common. Locally the diamicton also appeared to have a 'structurally stratified' top with discontinuous partings sometimes of sand but elsewhere with no obvious lithological association. At Barhill Plantation [NS 4620 7171] a horn of red deer was found in a bed of sand in stiff red till during the 19th century. Wright (1896) recorded the occurrence of marine calcareous microfaunas from the formation at a number of places, including Huntershill Quarry [NS 607 695] and Hamilton Hill Brickworks [NS 584 675] .

The effects of the Dimlington Stadial Ice Sheet overriding pre-existing glacial, fluvial and lacustrine sediments may be seen in sections at Cadder and Bearsden. In bore-holes at Broomhill and Erskine Bridge cores exhibited steeply dipping faults and polished surfaces in geotechnically overconsolidated clay, silt and diamicton.

At Cadder, glaciotectonic disturbance in the Cadder Formation, which underlies the Wilderness Till Formation (Plate 10) , is restricted to minor faulting and small-scale folding. In a former sandpit south of Bearsden Station [NS 5390 7148] , the undulating nature of the base of the Wilderness Till on the underlying Cadder Formation is thought to be partly glaciotectonic in origin. Neilson (1896) described 'seams' of coal (detritus?) in till in the Alexandria Parade area [NS 616 654] , just west of the district boundary. He noted that a seam was seen to 'rise at about an angle of 45° and turning across a sharp anticline descends at a similar angle'. He also recorded a 'seam' which curved, the upper end of which was nearly perpendicular. These structures may be glaciotectonic in origin, but also reflect the origin of the local diamicton as a deformation till.

The initial stages of deglaciation were marked by ice wastage on the high ground of the Campsie Fells and the Kilpatrick and Renfrew hills. At first, summits of hills reappeared from beneath the ice. Continuing wastage in high valleys resulted in the deposition of 'morainic drift deposits', ill-sorted sediments comprising boulders and gravel in a sandy clay matrix. Hummocky morainic land-forms are only developed in the Campsie Fells.

In time ice became restricted to the principal valleys and meltwaters issuing from glaciers transported and deposited large volumes of sediment. In west-central Scotland the direction of retreat was north-westwards towards the main Highland ice-source. As a result, in the Clyde valley ice receded progressively down-valley and seaward towards the inner Firth of Clyde and the fjords of the western Highlands. Likewise in the Kelvin valley, west of Kelvinhead [NS 757 785] at the watershed between the westerly flowing River Kelvin and the easterly draining Bonny Water, the valley glacier receded westwards and down-valley. Nonmarine sand and gravel transported and deposited by meltwaters in both valleys are referred to the Broomhouse (glacial) and Ross (fluviodeltaic) formations.

The surface morphologies associated with the deposits of this formation are those of ice-contact. There are mounds and isolated flat-topped kames but no closed hollows (kettleholes). The lithological associations are rare matrix-supported bouldery gravel with sand, common framework-supported bouldery gravel with sand, and pebbly coarse- to fine-grained sand and silt. Overall, the most abundant deposit is sand, except where buried esker ridges are present. The deposits are up to 25 m thick. Much of the gravel is massive to crudely bedded but planar beds and trough cross-bedded units are also present. The sands are planar bedded, trough cross-bedded, ripple laminated and horizontally laminated. Deformed bedding, including reverse faults and folds, probably marks the former contacts between dead-ice masses and the sediments. Deformation is most common in the esker ridge deposits. The fine-grained sands and silts exhibit minor load casts and flame structures are present locally.

The Broomhouse Formation was formerly seen in excavations at the Erskine Bridge [NS 4703 7286] where at least 2 m of glaciofluvial sand and gravel was overlain by 2.8 m of beach sand and gravel (Killearn Formation). This section and the associated ground ice (periglacial) structures were described and illustrated by Rose (1975, fig. 2). The formation was also seen in a sandpit adjacent to Clydebank Cemetery [NS 4755 7285] by Carruthers in 1904. Beneath 1 m or so of red sand and gravel, now interpreted as a beach deposit (Killearn Formation), were several metres of slightly cross-bedded, red silty sand with thin clay bands. The whole succession was locally faulted and the bedding highly contorted.

A buried esker ridge (BGS, 1994, section B) was met with in the excavation of the Clyde Tunnel [NS 542 662] . The esker is at least 15 m high and in cross-section 27 m wide at the top and 48 m at the base. The trend of the esker is approximately WNW with the southern and northern sides in contact with encasing sediments at dips of 57° and 45° respectively.

The sedimentological features of the Broomhouse Formation point to the operation of fluvial and deltaic processes. The coarsest gravel deposits are associated with esker ridges. They may have been deposited by glacial meltwaters flowing through tunnels in the wasting ice sheet. Massive deposits may have been laid down in the full-pipe sliding bed phase described by Saunderson (1977) but most show cross-bedding features consistent with fluviodeltaic processes and transport eastwards (data from the Airdrie district). The presence of sandy clay diamicton units may also point to the presence of decaying glacial ice but these might also have formed as debris flows in a fluvial context.

In the Clyde valley (Airdrie and Hamilton districts), upstream of Blantyreferme [NS 673 595] , a north-westward direction of deglaciation has been associated with the concept of an ice-dammed lake Clydesdale' (Bell, 1871). Some authors, however, (e.g. Sissons, 1976, p.128) have preferred an up-valley, south-eastward direction of ice retreat in the Clyde valley. Browne and McMillan (1989a) and Forsyth et al. (1996) have identified proglacial lake deposits (Figure 21) , the Bellshill Formation, the distribution of which appears to confirm the existence of such a lake (or series of lakes). The typical lithological association is of silty clay with laminae and beds of silt and sometimes sand. The deposit is dark brownish grey to brownish grey in colour. In the Glasgow district, the formation is known only from the Williamwood area [NS 57 58] at an elevation of 46-53 m above OD. However, these deposits appear to have been laterally ice-wedged as they occur west of the position of the terminal moraine.

Flat-topped and moundy deposits of sand and gravel, assigned to the Ross and Broomhouse formations, occupy mainly the northern slopes of the Kelvin and Glazert valleys and the south-east end of Strath Blane. Farther east in the Airdrie district, marginal drainage channels cut into some of these sands and gravels, and especially into the Wilderness Till and rock at topographically higher levels, indicating easterly drainage of meltwaters. These deposits form an integral part of the complex series of sediments which comprise the glaciofluvial and glaciomarine delta complex of the River Carron in the Forth valley (Browne et al., 1984).

South-west of Kilsyth (largely in the Airdrie district), a complex sequence of sediments flank the valley sides and infill the bedrock depression of the Kelvin. On the valley sides the sediments are disposed principally in the form of dissected terrace surfaces which are predominantly composed of sand with finely laminated silt beds and locally with gravel. The deposits which were once extensively worked at Torrance, Birdston and Cadder (Robertson and Haldane, 1937; Cameron et al., 1977) were considered to represent the margin of a glacial lake formed during a late phase of deglaciation as the valley glacier continued to retreat westwards (Clough et al., 1911). Recent borehole data tends to confirm the presence of both deltaic and lake bottom sediments which can be referred to the Ross and Bellshill formations respectively (BGS, 1994, section C). Locally ice-contact deposits of sand and gravel of the Broomhouse Formation are probably also present. These were formerly seen in a gravel pit at Langbank [NS 5718 7313] , where about 3 m of gravel and fine-grained sand displayed faulting and marked folding with subvertical downturns.

A succession through the glaciolacustrine sediments can be illustrated from Cadder No. 16 Borehole [NS 6002 7326] where the Bellshill Formation is 17 m thick. It is overlain by 18 m of sand, part of which probably belongs to the Ross Formation. Little is known of the sedimentology hereabouts of the Bellshill Formation.

The presence of deltas building out into 'Lake Kelvin' is indicated by the sand and gravel bodies which have been assigned to the Ross Formation. Sediments deposited at the west end of the proglacial lake are interpreted as a complex of ice-contact and deltaic sands and gravels, assigned respectively to the Broomhouse and Ross formations. The transport direction of the Ross Formation was clearly displayed in temporary sections exposed at the Bishopbriggs No. 2 Sand Pit [NS 625 734] which is just on the western edge of the Airdrie district. Here, large-scale trough cross-bedding indicated easterly current directions. Over 12 m of sand were formerly seen at this locality with beds of fine gravel at the top and silt layers near the base (? top of the underlying Bellshill Formation). Ice-contact association is indicated by the presence locally in these sediments of high-angle reverse and normal faults. The surface distribution of the Ross Formation at heights generally between 30 m and 55 m above OD is an indication of the lake level at the time this and the Bellshill Formation were deposited.

At about the time when Highland glacier ice in the Firth of Clyde had wasted sufficiently to allow 'Lake Clydesdale' to drain, the sea gained access to the Clyde estuary (Figure 22) . As the ice front continued to retreat northwards as independent glaciers in the valleys of the southwestern Highlands, any remaining ice in the Glasgow area and the Kelvin valley is considered to have become detached. Relative sea level in the inner Firth of Clyde was probably in excess of 35 m above OD at this time, some 13 500 BP (Browne et al., 1983b; Browne and McMillan, 1989a, pp.13-14). The Loch Lomond basin (Sheet 38W) and the 100 m-deep basin in Strath Blane (Doody et al., 1993) would also have been cleared of ice by northerly retreat. This would have happened whilst sea level was at about its highest, allowing the late-Devensian sea to gain access to both these areas not much later than in the Clyde and Kelvin valleys.

If observations and deductions about the altitudinal limits of marine (>35 m OD) and lacustrine (>70 m OD) deposits are reasonably accurate for the lower Clyde valley either side of Glasgow, failure of the ice dam at the Blantyreferme terminal moraine ( (Figure 22) ; Browne and McMillan, 1989a) would have released lakewaters under a considerable hydraulic head. These waters would have had great potential for erosion of pre-existing sediments and for deposition farther downstream. In particular, large volumes of sand and gravel would be available for erosion from the Ross and Broomhouse formations in eastern Glasgow. Redeposition of these is believed to have produced the Bridgeton Formation.

The Bridgeton Borehole [NS 6120 6367] contains the standard but incomplete section for this formation between 29.31 m and the base of the borehole at 40.38 m (Figure 20) . The two typical lithologies recognised are beds of very fine- to medium-grained (sometimes coarse-grained) sand, and fine to coarse gravel and boulders with a sandy matrix. Some of the sand was seen to be flat bedded but no other sedimentological features were identifiable. A nearby site investigation borehole recorded the base of the formation at 50.29 m, resting on the Wilderness Till. The formation is mainly unfossiliferous.

Sections in the formation at Shieldhall [NS 536 664] have been recorded by McMillan and Browne (1989; (Figure 23) ) who commented on evidence for former buried ice masses, and noted sedimentological features reminiscent of those described by Rust and Romanelli (1975) and Cheel and Rust (1982) for subaqueous outwash fans forming at an ice-front. A major difference between the Bridgeton Formation and the Broomhouse and Ross formations is that the deposits at Shieldhall were transported by westward flowing currents. This indicates that the associated ice upstream in the Glasgow area was probably detached from the active glacier ice which was retreating towards the western Highlands. Whether dead ice in eastern Glasgow could allow the transmission of such volumes of sediment and meltwater is open to debate but a source for the deposit is available in the sediments of the Broomhouse and Ross formations farther upstream. A source of meltwater could be the rapid draining of 'Lake Clydesdale' in which the Bellshill and Ross formations had been laid down. McMillan and Browne (1989) interpreted the Bridgeton Formation on this basis, as an outwash deposit formed in a submarine environment. The slump and load structures described by Jardine (1965, figs 1 and 2) in fine-grained sand at St Vincent Street [NS 5811 6567] in central Glasgow could be easily matched with similar ones seen at Shieldhall in the Bridgeton Formation. This suggests that Jardine's section was cut in Bridgeton and Paisley formation deposits.

Dead ice must also have remained to cap some of the remnant mounds of sand and gravel of the Broomhouse Formation. Locally the mounds would have been situated at elevations low enough to allow burial by the marine clay and silt of the Paisley Formation. As at Shieldhall, the Greenoakhill sections in the Airdrie district reveal evidence for burial of ice which survived perhaps for many hundreds of years in or under deposits of sand and gravel sealed under a thick cap of marine clay and silt of the Paisley Formation. The bedding of the overlying muds is disrupted by normal and reverse high-angle faults, and by low-angle thrusts with folding. These structures developed when the ice, buried in the underlying Broomhouse and Ross formations, melted.

Shieldhall and Greenoakhill are not the only localities where evidence for the survival of buried ice beneath marine sediments may exist. Bennie (1866) observed sections (Figure 24) at Windmillcroft Dock [NS 583 647] where the base of a marine clay (Paisley Formation) always rests irregularly in large hollows on the underlying fine-grained, white sand (Bridgeton Formation). The sand contains rare isolated boulders some of which were polished and striated. The lamination in the marine clay was observed to be inclined at high angles and to be cut by sand dykes. Wallace (1905) at the Trongate [NS 596 649] recorded dips of over 30° in finely laminated clay with sand partings (basal Paisley or top of Bridgeton Formation). Robertson and Crosskey (1874, p.246) reported a major excavation (Figure 25) in the Stobcross railway cutting [NS 5639 6574] to [NS 5718 6554] in which 'violent contortions' in the clay (Paisley Formation) were seen bent into deep hollows or troughs into the underlying sand (Bridgeton Formation).

Peacock (1971) suggested that the route of entry of the late-Devensian sea to the Paisley–Glasgow basin was not via the Clyde estuary but by a more southerly route, through the Lochwinnoch Gap between Johnstone [NS 43 63] and Dalry [NS 29 49] (Sheet 22E). There is no stratigraphical evidence to support this suggestion despite two attempts to find the linking marine deposits by drilling: at North Kerse [NS 339 556] Dickson et al. (1976) and at Lochwinnoch [NS 352 581] (Browne and McMillan, 1989a, pp.44, 61). In the absence of either stratigraphical or geomorphological evidence to the contrary, it is thought that dead ice blocked the gap until after deglaciation of the Glasgow area.

Once the water levels of 'Lake Clydesdale', 'Lake Kelvin' and the sea had reached equilibrium in the Clyde valley area, high energy environments were restricted to the beach zone and to the mouths of tributary valleys where lateral deltas built out. Initially, meltwater plumes dominated and the Paisley Formation reflects this influence. The formation was proved in the BGS Bridgeton [NS 6120 6387] , Killearn [NS 5100 84671 and Linwood [NS 4459 6588] boreholes. The Linwood Borehole (Figure 20) contains the standard section where the formation (26.5-31.1 m) consists of four lithofacies. The uppermost consists of clay and silt (26.5-28.14 m), finely colour banded in shades of brownish grey, greyish brown and grey, that contain a sparse calcareous marine microfauna and macrofauna. The second is silt with clay hands (28.14-29.20 m), that is grey, well bedded, and contains a sparse calcareous marine microfauna. The third is clay with silt layers (29.20-30.23 m), very finely colour laminated in shades of reddish brown, dusky red, orange and grey. The lowermost consists of grey, well-bedded silt, with some sand and clay beds (30.23-31.10 m). This sequence is typical of the Clyde valley west of Govan and the Loch Lomond area.

The succession is a lot thicker to the east of Govan, and the formation in the Bridgeton Borehole (Figure 20) consists of three lithofacies. The uppermost is silty clay, with many silt laminae and thin beds (6.9-16.23 m), which is colour banded in shades of brownish grey and sometimes reddish brown. Overfolds, normal and low-angle reverse faults, primary current lineation and burrows are present. The middle is silty clay with many silt laminae and thin beds (16.23-21.77 m), which is colour banded in shades of brownish grey and reddish brown with beds of finely colour laminated clay in shades of reddish brown, brown and grey. The lowermost is silt, with clayey silt and clay bands (21.77-29.31 m), which is very thinly laminated to thinly bedded, brownish grey, reddish brown and grey. Slumped bedding including overfolds is developed. All three lithofacies are rarely fossiliferous.

Other sections in the thin (western) variant of the Paisley Formation include the temporary sections at Shieldhall. Here the lithofacies sequence resembles that of the Linwood Borehole (Browne and McMillan, 1989a, p.24). Sedimentological features recorded included pipe-like burrows, sweep and burst laminations (Mackiewicz et al., 1984), graded units, normal and reverse microfaults, small rip-up clasts (less than 1 cm) and diamicton wisps up to 5 cm long by 5 mm thick. In the Killearn Borehole [NS 5100 8467] simple burrows, microfaults, primary current lineation and, as in most sections, dropstones were seen. In Mansionhouse Road, Paisley [NS 492 642] a temporary section revealed 45 cm of Paisley Formation, the highly laminated clay and silt resting on the eroded top of the Wilderness Till Formation and dipping to the south at an angle of 16°. The upper surface of the laminated deposit is also eroded with isolated parts of laminae preserved at the contact with a ripple-laminated silt (15 cm thick) forming the base of the overlying shelly Linwood Formation (P Aspen, written communication).

Some of the sedimentological features of the Paisley Formation suggest deposition by traction currents (turbiditic density underflow); these include graded bedding, primary current lineation, erosion scours and rip-up clasts. A turbidite origin is more likely to explain the localised ponded distribution of these deposits commented on by McGown and Miller (1984) and interpreted by them as evidence for deposition in isolated lakes. Piper et al. (1983, fig. 11) illustrate acoustic facies in fjord basins where ponded distribution of sediments may be related to a turbidity-flow origin.

The most striking feature, however, is the fine-scale layering of some of the clays. This has been attributed to seasonal banding (varves) in proglacial lakes, a view most recently stated by McGown and Miller (1984), but Jardine's review (1986) of the general stratigraphy of the Clyde estuary states the now widely held belief that they are glaciomarine in origin. This view is compatible with the faunal evidence, even if the scarce marine assemblages of Glasgow and Paisley areas were derived by erosion from older sediments, for it is unlikely that the marine assemblages at the Hawthornhill site by Dumbarton [NS 375 760] and at Ardyne [NS 098 683] in the Firth of Clyde (Peacock et al., 1978) are anything other than contemporaneous with the sediments in which they are found. Such fine-scale layering has been reported from glaciomarine environments close to fjord glaciers. Mackiewicz et al. (1984) believed that the origin of the layering in their interlaminated facies was due to flocculation from overflow or interflow meltwater plumes interacting with a sea-water wedge that fluctuated in response to the tidal cycle. A similar origin for the Paisley Formation is possible, in effect an ice-clearance facies deposited in a salinity-stratified estuarine environment. Rapid sedimentation is likely because the evidence for bioturbation of sediments is slight, with only a few examples of trails and burrows recorded, except in the uppermost part of the succession in the Linwood Borehole and Shieldhall sections.

There is a major disparity in thickness between the Linwood and Bridgeton developments of the formation and, although the lithofacies associations are generally similar, they are present in different proportions. The difference is believed to be due to a much higher rate of sediment supply to the depositional basin east of central Glasgow.

The marine fauna is substantial in the more seaward parts of the Clyde region (Peacock et al., 1978; Browne et al., 1983h; Browne and McMillan, 1989a, pp.23-26) but sparse farther inland, in the Glasgow–Hamilton area, where sediment supply was greater. The maximum altitude to which clays assigned to the Paisley Formation are known to occur is generally over 40 m above OD in eastern Glasgow. Possible equivalent but unfossiliferous deposits at Dougalston Loch [NS 561 738] , to the north-west of the Kelvin valley, attain an elevation of 41 m above OD. However, sediments even tentatively identified as belonging to the Paisley Formation have yet to be located farther east in the Kelvin valley.

In the Glasgow and Paisley areas the Paisley Formation is commonly overlain by thickly bedded muds of the Linwood Formation (Browne and McMillan, 1989a, pp.26-27). This formation was proved by the Linwood and Killearn boreholes. The Linwood Formation is sulphide rich, and contains debris rafted by ice and some wisps of diamicton. Air-dried samples of the mud suggest the presence of vague and discontinuous silty or micaceous wisps which may have been formed in the manner described as 'sweep and burst texture' by Mackiewicz et al. (1984). Elsewhere there is evidence of strong reworking by burrowing.

The Linwood Borehole contains the standard section for this formation (2.64-26.5 m) and shows three lithofacies. The uppermost consists of silt and clay, medium grey and micaceous (2.64-6.65 m). Sand occurs at and near the top, in layers up to 3 cm thick. Seaweed and other plant remains are present, with eroded and stained foraminifera, and traces of rooty horizons near the top. The middle is of clayey silt and silty clay (6.65-18.85 m) which are medium to dark grey becoming brownish grey at depth. Silt and sand layers (and beds) and black sulphide beds and mottling are developed. Seaweed and marine shells are present. The lowermost is clayey silt and silty clay (18.85-26.5 m), brownish grey in colour but sometimes medium grey, thickly bedded to massive, with silt dustings and wisps or indistinct partings. Many marine shells are present and there is some sulphide mottling. The beds are heavily burrowed. It is possible that the middle unit of this borehole is the same deposit as the dark grey clay recorded by Robertson and Crosskey (1874) at Fairfield Dock [NS 546 662] . In the Killearn Borehole (34.8-45.41 m) only the lowest unit of the Linwood Formation is present. The deposit contains rare beds of diamicton usually less than 5 cm but in one case up to 36 cm thick. Some sand layers and sharp-based beds up to 18 cm thick (? storm beds) are present.

The Linwood Formation was deposited as sea-bottom muds in moderate to shallow water depths in a well-mixed estuarine environment. The presence of isolated gravel to pebble-sized clasts interpreted as ice-rafted dropstones indicates the presence of shore ice in winter as it is believed the whole area was free of glacier ice at the time of deposition. Overall the succession coarsens upwards as the marine embayment became shallower and more estuarine in character. The presence of black sulphide beds and mottlings indicates anoxic sea-bottom conditions. This probably reflects the infilling of the marine embayment coupled with lowering of the relative sea level and the existence of a fjordic sill at the mouth of the basin around the narrows of the Erskine Bridge area. The heterolithic upper unit of the Linwood Borehole, with evidence of rooty horizons and eroded and stained foraminiferan tests, is probably intertidal in origin. The roots may he associated with the Clippens Peat Formation above.

At a number of localities, the Linwood Formation is sandy or contains beds of gravel rich in shells. An example of the former is the section on the M8 Motorway east of the Renfrew Road Bridge described by Aspen and Jardine (1968, fig. 1). Here at least 3 m of shelly silty sand (Bed 2) were seen at their localities A [NS 4937 6571] and B [NS 4967 6570] . Examples of gravel rich in shells were described by Browne et al. (1977) at Inchinnan (Figure 26) . At one locality [NS 4766 6940] the shelly gravel is fine to coarse (0.075 m maximum size), angular to sub-rounded, with a sandy and clayey matrix. A valve of Arctica islandica was dated at 13 142 BP whereas a valve of Modiolus modiolus gave a younger date of 12 011 BP. The ground level is 11.0 m above OD, the base of the gravel at 6.1 m above OD, and the top at 7.22 m above OD. The gravel at lnchinnan is overlain by fine to medium-grained, yellow brown sand, with thin brownish grey silty clay laminae, and rare isolated subangular stones up to 0.05 m. This deposit is flat bedded with a low-angle dip towards the centre of the local valley. Vegetable matter is present near the base. On top of the sand is a brownish grey or greyish brown clay or clayey silt, with many thin hands of yellow-brown fine-grained sand. Flame textures and load casts are present at the tops and bases of sand beds. Isolated subangular stones up to 0.05 m are scattered in the clay. Both the clay and the underlying sand may be a mix of soliflucted and alluvially deposited sediments.

The beds of gravel or gravelly clay in the Inchinnan and other sections appear in marginal marine, sublittoral situations in which higher energy conditions prevented the deposition of fine deposits or gently reworked them (Peacock, 1989). The Inchinnan localities are all geographically close to conspicuous shoreline features of late-Devensian age at and below 15 m above OD which are less than 12 000 years old.

Numerous radiocarbon dates from in-situ and derived shell material originating in the most fossiliferous basal part of the Linwood Formation yield a spectrum of ages between 13 150 and 11 300 BP. Sedimentation was intermittent in the Linwood Formation and, as with the present sea floor, deposition of mud, sand or gravel, or growth of shell beds, was taking place in some places while elsewhere there was erosion or non-deposition.

There is some evidence for locally variable contacts and relationships between the Linwood, Paisley and Wilderness Till formations and bedrock. Anderson in 1938 observed sewer trench and tunnel sections at the Cart Harbour [NS 4844 6526] to [NS 4855 6556] and noted that over a short distance the Linwood Formation rests directly on both of the other formations and on bedrock.

The fauna of the Paisley and Linwood formations, Appendix 6, includes cold-water indicators such as Chlamys islandica, Macoma calcarea, Nuculana permula and Elphidium clavatum, reflecting a harsh but not fully arctic glaciomarine environment. The greater abundances and diversity of the fauna in the Linwood Formation suggests that conditions were a little more favourable and more fully marine than those prevailing during the deposition of the Paisley Formation.

The Killearn Borehole contains the standard section in the Killearn Formation (15.61-34.8 m). The chief Lithology is fine- to medium-grained sand. Units of fine-grained sand with silt or clay layers also occur, especially near the top and base, and beds of fine gravel or sand with gravel (3 cm maximum clast size) are present in the upper part. The beds of the formation are reddish brown or orange in colour. The deposit forms an upward-coarsening deltaic unit from the base at 34.8 m up to 17.95 m and then fines up to the top of the formation at 15.61 m. Unfortunately the basal contact with the Linwood Formation was not seen due to core loss.

Similar deposits are present locally in the Clyde valley in central and eastern Glasgow, as proved in commercial boreholes, for example near the Bridgeton Borehole. They may include deposits of beaches and river terraces as well as deltas. The stratigraphical relationships of this formation are not well known, but are partly illustrated by sections at Scotstoun House [NS 525 680] described by Jardine and Moisley (1967, fig. 1), by Rose (1975) at Erskine Bridge [NS 4703 7286] and also in the temporary section by Clydebank Cemetery [NS 4755 7285] seen by Carruthers in 1904. In the last two sections the beach deposits are well-bedded sand and gravel.

The sea-level curve for the Clyde area following deglaciation is poorly known compared with that for the eastern coast firths (Sissons, 1976, fig. 9d; Armstrong et al., 1985, figs 13, 15). In the Rutherglen [NS 610 618] and Carmyle [NS 640 620] areas deltaic sand and gravel deposits of the Killearn Formation occur. Elsewhere in Glasgow, the Firth of Clyde and Strath Blane, raised shorelines, raised beach and delta deposits are present but no clear assessment of relative sea-level changes can be reconstructed in the absence of systematic levelling. It is likely, as in the Forth valley, that, during the Windermere Inter-stadial and by the time of the Loch Lomond Stadial, the relative sea level fell to OD or below, due to the local dominance of isostatic uplift of the land over the generally rising world sea level.

The only lacustrine deposits of equivalent age to the Linwood Formation which have been studied in any detail are to be found at Garscadden Mains [NS 5327 7059] . Here Mitchell (1952) described 2.15 m of muds with sand and clay and proposed a stratigraphy for this classic site. Anderson and Simpson added an addendum on the marine origin of the varved clay now assigned to the Paisley Formation. Jardine (1969) described further sections near Garscadden but these did not encounter the lake muds.

On land, the contemporary pioneering vegetation appears to have been dominated by grasses, sedges and ferns. This heathland, mossy in part, contained patches of hare ground or disturbed soil as indicated by pollen of Rumex and Plantago. These two forms may indicate near periglacial conditions but limited snow cover.

Movements of relative sea level during the stadial are not well known. At or just before the time of the maximum extent of the ice, sea level may have been as high as 11 m OD, based upon the presence of former intertidal platforms and clifflines at about this height in the Firth of Clyde and around the shores of Loch Lomond (Rose in Jardine, 1980 pp.29-31). The view that such features were formed by severe shore erosion in the periglacial environment of the Loch Lomond Stadial (references in Jardine, 1986) must be questioned in the light of the discovery of till on the platform at Cardross in the Greenock district. The evidence presented by Browne and McMillan (1984) implies reoccupation of a pre-existing marine feature here, and by implication elsewhere, during the Loch Lomond Stadial. In Danes Drive, Scotstoun [NS 532 679] and at Shiels, near Renfrew [NS 523 665] drumlins, planated by the sea in Loch Lomond Stadial times, now form parts of the former intertidal platform as developed in the Clyde valley. Former excavations in the Glasgow Road, Clydebank [NS 4990 6987] to [NS 5055 6925] showed potholes up to 7 m deep eroded into the Wilderness Till which here forms the platform. Carruthers' notes describe the infill as mud and water. The presence of potholes might not be expected if the platform was entirely cut in periglacial conditions.

On the east coast of Scotland sea level may have fallen below that of the present at some period during the Loch Lomond Stadial (Browne, 1985 in discussion of Sutherland, 1984; Armstrong et al., 1985, pp.87-89). The implications are that on the west coast sea level was also lower than OD. This is a possibility in the Clyde as implied by McGown and Miller (1984, fig. 7) for central Glasgow, where they recognise a channel incised to over 40 m below OD, aligned close to the course of the River Clyde and back-filled with river sands and gravels. However, dating of the supposed incision is open to doubt because of the problem of correlating largely unfossiliferous superimposed successions of sands and gravels which could include the late Devensian Broomhouse, Killearn and Bridgeton formations (Figure 20) overlain by similar Flandrian formations. The incision may therefore, as previously suggested, be related to much earlier erosional events.

Landslips are common on the upper slopes of the Campsie and Kilpatrick hills (Plate 12) . They were probably initiated during the Loch Lomond Stadial. They consist principally of rotated blocks of basaltic lavas and tuffs (Plate 13) . Adjacent areas of thin geliflucted drift (head) are also present. The largest landslips which cover an area of up to 1 km ² occur beneath basalt lava scarps near Lang Craigs [NS 430 765] , Tomibeg [NS 510 773] and at the corries of Balglass [NS 580 855] to [NS 590 855] . Other extensive areas are found north of Blanefield [NS 55 85] and on the Kilpatrick Braes [NS 474 790] .

In Glasgow unusual subsoil profiles observed on the flanks of drumlins may also be the product of large-scale solifluction (Dickson et al., 1976). Both at Robroyston [NS 633 676] in the Airdrie district and at Springburn [NS 609 678] (Forsyth et al., 1996, fig. 24) a grey unweathered till overlies peat, which is radiometrically dated at between 11 650 and 11 100 BP. The peat contains a flora indicative of tundra conditions. At Springburn the till is overlain by a younger peat dated at 5994 BP. Because these sites lie well beyond the limit of the Loch Lomond Stadial ice, the tills seem most likely to be the products of the Dimlington Stadial glaciation. The available evidence therefore suggests that till has moved over the organic materials probably by solifluction processes which were a major feature of periglacial activity in west-central Scotland during the Loch Lomond Stadial. At least two units of soliflucted sediments were recorded in the section described by Mitchell (1952) at Garscadden Mains [NS 5341 7124] ; the lower was of Loch Lomond Stadial age.

Soliflucted deposits are probably present above the marine succession in the Inchinnan area. A section seen in a sewer trench [NS 4764 6935] showed 2.7 m of thinly bedded, brownish grey clayey silt, with dark grey laminae and sand layers. Layers, 2 to 4 cm-thick with plant debris, were present at the base and 14 cm above. This silt rested on a brownish grey, clayey in part, gravelly, sandy, diamicton 0.9 m thick. The gravel portion was subangular to angular in the lower part, and subrounded to rounded in the upper part, the clast size averaging 2 to 4 cm. The abundant matrix was massive and contained marine shells. Underlying the diamicton is shelly, brownish grey silty clay that is also gravelly with angular to subrounded clasts up to 5 cm. This deposit is part of the Linwood Formation. As the locality is well outside the limits of the Loch Lomond Readvance Ice and shells from underlying sediments nearby are under 13 000 years old, the diamicton in this section is unlikely to be glacial in origin. The most likely explanation for its origin is that it is a marginal marine gelifluction deposit of Loch Lomond Stadial age. The overlying silt is interpreted as an alluvial sediment interlayered with soliflucted material partly in-filling the floor of the minor valley where the section is located.

At Portnauld Farm railway cutting [NS 4915 6855] to [NS 4940 6857] a possibly similar gelifluction deposit was seen by Jack. Described in his notebook of 1870 as 'unmistakenly boulder clay', the diamicton had the appearance of a coarse gravel especially near the top. The stones were enclosed in a stiff unstratified paste, were lying at all angles, and many were distinctly striated. The bed was 0 to 0.9 m thick. Beneath, was a stony, sometimes clayey sand. Shells occurred in groups, with balanids adhering to the stones and on some shells. Paired valves were common. Some of the stories were large and considered to be washed out of the underlying till. The bed had an undulating base and was up to 0.3 m thick.

Further occurences of soliflucted diamictons may also explain the presence of apparently typical Linwood Formation shelly deposits below 'boulder clay'. Unlike Inchinnan and Portnauld which are both lower than 12 m above OD, these localities (seen by Jack) are higher, approaching the late-Devensian marine limit as recognised in the Linwood–Paisley embayment. In the first section [NS 4112 6482] at Tweeniehills railway cutting (SL c.36-37 m above OD), the surface deposit is a diamicton composed mostly of shale fragments. Some of the small stones are distinctly striated but there is no sign of stratification. This deposit was said to be similar to that at the top of a section at Windyhill. It may be 3.4 m thick, resting on a dark blue, finely laminated clay that is rather sandy in places. Shells were noted only in the sandy part of the succession. The clay is 1.4 m thick but dies out eastwards within 6 m. Underneath is another diamicton at least 1.7 m thick and believed to be a glacial till. In the eastern section [NS 4118 6478] in the cutting (SL c.34 m above OD), a blue clay up to 0.6 m thick with a thin peat layer is underlain by a very stiff, laminated, blue clay with very fragmentary mussel and other shells in a layer 1.2 m from the surface. This clay was about 1.2 m thick and rested on a diamicton 1.4 m thick. No trace of the soliflucted diamicton was recorded.

In the Windyhill railway cutting [NS 4220 6455] (SL c.27 m above OD), an unstratified diamicton 1.2 m thick, composed of debris of shale, contains a few fragments large enough to show that they were glacially striated. Underneath is a stratified blue clay, 0.45 m thick, with occasional layers almost wholly of Carboniferous shale fragments. This deposit rested on another blue clay (1.2 m thick) that included a shell bed (0.30 m thick). At the base of the section was another glacial diamicton 1.1 m thick.

A diamicton at surface at Drumcross Farm [NS 4493 7120] may also be a solifluction deposit. The locality is at about 27 m above OD. The diamicton (described as boulderclay) had stories of all sizes up to 36 cm lying in all directions. The stones were not rounded or striated. The bed lacked stratification and was up to 1.22 m thick. Beneath was a stiff clay with shells, much 'interrupted' by stones of the same shape as in the diamicton above. There was a very large stone at the top with balanids adhering above and below. The clay had sandy pockets and became very sandy and even gravelly towards the base. Specimens of Mya truncata were seen by Bennie and Jack. This clay bed (up to 36 cm thick) is believed to be lenticular and very localised as the adjacent section, reportedly recorded by Fraser, was of a 15 cm-thick, lenticular, shelly gravel.

One further locality at which soliflucted diamicton may be present is on the Bargaran and Shilton March Burn [NS 4578 7089] . The surface level is around 26 m above OD. Here a diamicton (described as boulder clay), of unknown thickness, rests on a strong, dark blue clay containing driftwood (Scots pine), fragmentary decayed shells and a beetle.

The Mains of Kilmaronock Borehole [NS 4483 8829] provides evidence for three possible marine incursions into the Strath Blane basin from Loch Lomond during the early to middle Flandrian. The earliest incursion is represented by the seasonally banded muds of the Buchanan Formation (Browne and McMillan, 1989a) and is recognised on the basis of the occurrence of dinoflagellate cysts and foraminifera. This formation rests almost directly on the Gartocharn Till Formation, here consisting of glacially reworked marine muds from the Linwood Formation. As a result, there is a possibility of faunal reworking from the older formation. The Buchanan Formation is over 9000 years old, as the pollen analysis indicates the presence of pioneer heath vegetation. Sea water could have entered Loch Lomond from the Vale of Leven at times of high sea level, as was the case at that time in the Forth (Sissons, 1976).

The typical lithology of the Buchanan Formation is a thinly bedded silty clay with many laminae and thin bands of silt and sometimes of sand. The deposits are brown, brownish grey and reddish brown, and are conspicuously colour banded, the banding being emphasised by the grey coloured units of clay. Isolated clasts up to 4 cm are present (possibly dropstones). Overall the dark and light colour banding of the deposit suggests seasonally controlled deposition (varves); the thickness of the couplets decreases upwards from as much as 15 cm to less than 1 cm.

The Buchanan Formation is overlain by the lacustrine muds of the Kilmaronock Formation. The Mains of Kilmaronock Borehole contains the standard section (13.60-19.95 m). The typical lithology is a very thinly bedded silt with many clayey silt and silty clay layers. Sand layers are also common. The deposits are brown, dark brown, brownish grey and grey in colour, with plant remains and locally with dark organic-rich bands. Grains and flecks of vivianite are present throughout. The lithology of the Kilmaronock Formation is consistent with deposition of lake bottom muds in the Loch Lomond basin and perhaps the Strath Blane basin. The alternation of organic-rich and organic-poor horizons suggests seasonal controls on deposition. The increasing abundance of Corylus pollen in the upper half of the formation would suggest an age of not more than 9000 years (Price, 1983).

The Kilmaronock Formation is overlain by the predominantly upward-coarsening sandy Endrick Formation of deltaic and fluvial origin (5.99-13.60 m in the Mains of Kilmaronock Borehole). The typical lithology is loose, fine- to medium-grained sand with some silt layers. The deposit is reddish brown and contains abundant plant remains and some dark organic clay bands.

A second marine incursion possibly took place over 7000 years ago (pre-Alnus rise). It is represented by a bed of grey silt with dinoflagellate cysts in the Endrick Formation at a depth of 9 m in the Mains of Kilmaronock Borehole. At the top of the formation a mixed forest flora is indicated by the pollen and, following Price (1983), this would mean that the strata at about 6.6 m depth in the borehole are at least 6500 years old.

The stratigraphical record of the Mains of Kilmaronock Borehole is completed by a third marine incursion represented by the Erskine and Gourock formations. Not only are dinoflagellate cysts present in the Erskine Formation but there are at least two layers composed of partly rotted shells of the bivalve Mytilus edulis (at depths of 5.15 m and 5.46 m respectively). As there is no indication of the Ulmus decline in the pollen record this unit is at least 5000 years old. Based upon the report of Dickson et al. (1978) on almost certainly the same marine incursion found in offshore cores, the transgression lasted from around 6900 to 5400 years ago.

In the Linwood and Paisley embayment of the lower Clyde valley, two views of the extent of the Flandrian marine incursions have been proposed. Browne (in Jardine, 1980) suggested that the large peat mosses may have excluded the sea from most of this area and that extensive flat tracts at about 6 to 9 m above OD are, for the most part, remnants of a late-Devensian surface produced by erosion of slightly older late-Devensian estuarine sediments. The surface concerned may have been produced at the time of the 'Main Late-glacial Shoreline'. Boyd (1982, fig. 8:4) suggested a more extensive area of Flandrian submergence, on the basis of a detailed study of the Linwood Moss area. He identified a mid-Flandrian shoreline position, with associated deposits of silt up to 6 m thick and extending up to 8 m above OD. Boyd also identified an 'old' and a 'young' peat. The former is present in his Linwood Moss Wood Borehole [NS 4395 6605] and in the BGS Linwood Borehole, the latter in his Moss Cottage Borehole [NS 4440 6610] . There is doubt about the apparently young age of 9290 RP (Boyd, 1986) for the base of the 'old' peat obtained in the Linwood Moss Wood Borehole. In the Linwood Borehole the base was dated at 9540 BP. The base of the 'young' peat was dated at 3650 BP in the Moss Cottage Borehole. These dates are consistent with previous ones reported from this area (Boyd, 1982; 1986). From the radiocarbon dates alone, it is obvious that the history of Linwood Moss is complicated. No research has been carried out on the histories of the nearby Barochan, Fulwood and Paisley mosses or on remnant patches of peat in the lower Clyde valley east of Renfrew.

The only stratigraphical evidence for Mondrian marine transgression in the Linwood–Paisley area is a 5 cm thick band of grey mud within the Clippens Peat Formation in the Linwood Borehole. This is unfossiliferous apart from a specimen of Onoba semicostata, a marine gastropod which may have been derived from late-Devensian sediments. However, taken with pollen evidence of a hydroseral influence in the peat below and above the mud, an estuarine or near-estuarine origin for the deposit is acceptable. The mud unit is probably younger than the 8000 BP that Boyd (1982) envisaged, for a radiocarbon assay of the base of the overlying peat gave an age of 7110 BP. However, this marine event, which ended before the Alnus rise, is older than the middle Flandrian marine transgression (6900-5400 BP) that is represented by the Gourock and Erskine formations in the Loch Lomond basin. However, the 7110-year date and Corylus-dominated pollen record do suggest a correlation of the Onoba-bearing mud with the dinoflagellate cyst-bearing grey silt at 9.4 m in the Mains of Kilmaronock Borehole. Although there are reservations about identifying these two deposits as marine, on the basis of a single shell and a few dinoflagellate cysts, it is considered that both relate to the same rise of relative sea level, on the evidence of their similar ages.

Evidence for changes in Flandrian sea level is seen in the presence of three series of Mondrian to Recent levels in the lower Clyde valley (Figure 28) . The upper limits of these surfaces are at about 12 m above OD, less than 7.5 m above OD and less than 5 m above OD (based on a review of Ordnance Survey spot heights). However, in the Linwood–Paisley embayment only the lowest of these levels can be recognised. Between 5 and 12 m above OD there are generally gently sloping surfaces not divisible into discrete terraces, and these are overlain by remnants of previously more extensive peat mosses. In both areas there is generally a marked, low cliffline at about 12 m above OD. This feature is partly concealed by peat to the west of Barochan Moss, north of Fulwood Moss and south-west of Linwood Moss. This fact suggests that the cliffline could be an inherited feature of late-Devensian age, certainly in those areas where the mosses are likely to have accumulated continuously from the beginning of the Flandrian. From Boyd's work (1982) it is also likely that, wherever the surface level of the moss approaches 15 m above OD, peat accumulation has been continuous from the beginning of the Flandrian, and the surface and sediments below are of late-Devensian age. Indeed, the late-Devensian Linwood Formation is at surface over extensive parts of the flat ground between 8 and 12 m above OD in the Linwood–Paisley area.

However, it appears that the Flandrian marine incursion flooded all that part of the Linwood–Paisley area below 7.5 m above OD, the general upper limit of the middle series of surfaces in the Glasgow district. Above this limit, the only evidence for Flandrian transgression is the mud unit with the Onoba shell in the Clippens Peat Formation in the Linwood Borehole at about 8.4 m above OD. Boyd (1982, fig. 8.4) constructed his conjectural configuration based on an altitudinal limit of 8.7 m above OD outside the mosses and a surface level of 11 m above OD inside. The reconstruction shown in (Figure 28) is based on similar premises.

The deposition of the 'young' peat marks the final regression of the Flandrian sea in the district. The Moss Cottage radiocarbon date of 3650 BP for the base of the peat places sea level well below 7.8 m above OD by this time. In fact, it is likely that sea level was below 5 m above OD, this level being the lowest recorded on the outer edge of the middle series of estuarine flats in the valley.

The oldest Flandrian estuarine deposit is the Longhaugh Formation. It consists mainly of grey, loose, fine- to medium-grained, sometimes silty sand containing sonic comminuted shell debris. It forms the infill of an erosive slot cut into older estuarine and glacial sediments in the lower Clyde valley. The floor of this channel appears to be 20 m or more below OD in central Glasgow (McGown and Miller, 1984, fig. 7). Gravel with shells was recorded in the Longhaugh No. 20 Borehole [NS 4291 7319] , below sand, between levels of 19.90 and 21.26 m below OD. The Longhaugh Formation was probably deposited in estuarine channel environments.

The Erskine Formation (Figure 27) comprises subtidal to intertidal silty clays with sand and silt laminae and beds. Organic debris and isolated pebbles are present. The deposit is generally brownish grey or greyish brownbut black where it contains much iron sulphide. The sand and silt is ripple laminated in places.

The Gourock Formation consists predominantly of sand, commonly with gravel. It was proved in the top part of the Bridgeton Borehole where the uppermost metre consisted of brown silt and clay with thin beds and wisps of peat and sand. The Bridgeton sequence is typical of this part of Glasgow as the fieldnotes of Jack and Bennie describe several sewer trenches displaying similar sediments [NS 6080 6375] , [NS 6078 6411] , [NS 6083 6357] . Elsewhere in the district, sections indicate that the direction of transport is consistently towards the west. Dark grey beds rich in organic detritus may be present as in the Shieldhall sections [NS 535 663] . They are also present in the Dalmarnock area as in Dale Street [NS 6083 6357] where, under 1 m of clay, Jack and Bennie recorded 2.1 m of dark river silt with oak driftwood, leaf beds and vegetable debris, these sediments resting on at least 1.3 m of sand and gravel. In John Street [NS 6048 6376] a similar thickness of silt was recorded as containing grasses, sedges and acorns; the silt also had diatoms in it at Muslim Street [NS 6093 6373] . In Hutchesontown [NS 5987 6290] they also noted the presence of an angular block (1.3 m) of laminated clay (i.e. the Paisley Formation) largely encased in sand and gravel but covered by yellow loam (i.e. the clay of the above Dalmarnock localities).

At Windmillcroft Dock [NS 583 647] sandy gravel contains boulders of glacial till, pebbles of brickclay and peat, artefacts and periostraca of the freshwater mussel Unio margaritifera. Plant remains are common including a large trunk of an oak but no beech. The formation has a markedly erosive base, and locally is up to 5.2 m thick where it infilled a channel cut into the underlying units.

The lithological association of the Gourock Formation is compatible with an estuarine origin for the sediments. However, the estuarine environments represented range from one dominated by fluvial processes in the east, in Glasgow, to an almost fully marine environment in the west, at Gourock. The fluvially dominated sediments were deposited in very shallow channels probably linked in a branching anastomosing pattern, isolating low-lying islands in the estuary. In the marine-dominated part of the estuary a pattern of shallow channels forming a complex with extensive intertidal flats is likely.

Peat, radiocarbon-dated as pre-Flandrian in age, has been found in a few places in the Glasgow district. As mentioned above (Landslips and soliflucted deposits), examples are known from Springburn and from Robroyston in the Airdrie district (Dickson et al., 1976). At Springburn [NS 609 678] geliflucted till overlies peat dated at 11 140 BP and is overlain by a thin Flandrian peat dated at 5994 BP.

The principal deposits of peat found in the district are of Flandrian age. During the early Flandrian, as the Scottish climate became warmer and wetter, Sphagnum peat developed in hollows on former lake floors and on poorly drained flat areas. Blanket bogs also formed on sloping ground in the Campsie Fells and Kilpatrick Hills in response to exceptionally high rainfall. Peat, which in the Glasgow area has been assigned to the Clippens Peat Formation (Browne and McMillan, 1989a, pp.35-36), formed in low lying ground particularly to the north of Linwood. Here extensive deposits have accumulated as raised mosses.

The Linwood Borehole contains the type section of the Clippens Peat Formation. Below 45 cm of very peaty stony soil, 1.3 m of peat rests on 5 cm of grey rooty clay, followed by a further 84 cm of peat to the base of the formation at 2.64 m depth. The surface level of the borehole was 10.18 m above OD; thus the base of the upper bed of peat was at 8.43 m above OD and the base of the lower at 7.54 m above OD. Samples representing the basal 3 cm of each peat bed have been radiocarbon dated. The ages are 7110 BP for the base of the upper peat and 9540 BP for the base of the lower bed. Boyd (1982; 1986) recognised that Calluna and Ericales have peaks in their pollen curves (16 per cent and 18 per cent respectively) on either side of the 5 cm clay band at 1.75 to 1.80 m. These peaks may reflect more open conditions just before and after the clay was deposited. As noted above, a single specimen of the marine gastropod Onoba semicostata was found in the clay, tentatively indicating a marine origin for this bed, the transgression ending earlier than 7100 years ago. Reliable indicators of saltmarsh vegetation are not found in the peat record. However, the ericaceous pollen taxa may indicate the development of a coastal heath. The pollen spectrum, as a whole, reflects an open environment which becomes progressively wetter and with an increasing number of trees at higher levels.

Coarse-grained Flandrian sediments deposited in fluvial environments including present-day river courses were assigned by Browne and McMillan (1989a, pp.39-40) to the Law Formation (BGS, 1994, sections C and D). Older Flandrian sand and gravel which may have been partly associated with deltaic and lacustrine environments were referred to the Endrick Formation (Browne and McMillan, 1989a, pp.36-37).

The Law Formation includes much of the Recent sand and gravel alluvium of the River Clyde together with the Endrick Water, the Blane Water, the Black and White Cart waters, the River Gryfe, the River Kelvin and the Glazert Water. The lithology is of loose, grey, fine- to coarse-grained sand with some silt and beds of fine to coarse gravel. Plant remains are present and dark sulphide patches occur. The deposit may be partly massive, sometimes flat-laminated but probably mainly cross-bedded.

In the valley of the Glazert Water up to 10 m of coarse gravel, boulders and sand are present. The deposits are likely to have been derived from Strath Blane to the north-west via the wide 'dry' valley of the Pow Burn. The whole sequence has been assigned to the Law Formation but the basal part could have been deposited by glacial meltwaters during either of the late-Devensian glaciations. Deposits of the Endrick Formation may be present under 'alluvial' terraces of the River Kelvin lying above the general level of the floodplain at elevations of 35 to 50 m above OD.

The Kelvin Formation includes the Recent fine-grained sediments of the valley floor of the River Kelvin and other rivers and streams. Lake alluvium of Flandrian age, present north of Glasgow and west of Bishopbriggs at elevations of about 40 m above OD, is also referred to this formation. The sediments are silty clays and silts, often with organic remains, interbedded with peat, generally formed either on the floodplain or in lacustrine environments such as inter-drumlin hollows. Clay and silt of the Kelvin Formation forms much of the valley floor of the River Kelvin (BGS, 1994, section C). Borehole records show the thickness of these sediments to range from about 2 to 9 m.

In the Kelvin valley lacustrine clay and silt deposits concealed by younger Flandrian sediments are locally present. These deposits probably belong to the Kilmaronock Formation (Browne and McMillan, 1989a, pp.35-36). Typically, the sediments comprise thinly bedded silt with clay laminae and some sand layers. Plant remains are common.

Extensive areas of man-made deposits are present in and around the urban areas of the Glasgow district. Wherever construction for houses, factories and roads has taken place there are likely to be areas where the natural ground surface has been covered by redistributed, geotechnically variable, natural and man-made materials. Typically made ground of this nature is only a few metres thick but in areas close to former excavations thicker deposits may be expected. Examples of major ground disturbance include the M8 motorway and interchanges such as that at Paisley [NS 465 655] . Former pit bings and slag tips also constitute made ground as does the spoil from the dredging of the bed of the Clyde and the excavation of the Clyde Road Tunnel (Nicoll, 1990).

The district has been extensively quarried for minerals including sand and gravel, brick clay, hardrock aggregates, fireclay, common shale, quartz-conglomerate, sandstone and coal. Many former excavations are infilled or partially infilled with variably compacted materials and these deposits, classified as Fill on the Glasgow Drift Sheet, are generally thicker than other made ground and may be in the order of tens of metres.

The Limestone Coal Formation is the principal source of coal in the district. At least 13 seams have been mined (Figure 11) , some of them extensively, such as the Knightswood Gas and Possil Main coals. All but the Kilsyth Coking Coal are in the upper part of the formation. The Middle Coal Measures are rich in thick coal seams, six of which have been mined (Figure 14) , but they only occur in a limited area in this district. The Lawmuir Formation has also produced a great deal of coal, partly from the Quarrelton Thick Coal at Johnstone and its component seams in and around Paisley. The Hurlet Coal has also been worked extensively both south and north of the River Clyde. The Lillie's Shale Coal is the only seam which has been exploited in the Lower Limestone Formation, mainly in the Linwood–Johnstone area. In the Darnley area and in north-west Glasgow, several coals in the Upper Limestone Formation have been mined but to limited extents. The Kiltongue and Airdrie Virtuewell coals in the Lower Coal Measures have been extracted in southern Glasgow to quite appreciable extents. Coal mining began over 300 years ago in Paisley and the Govanhill area of Glasgow. It ceased with the closure of Garscube Colliery [NS 573 698] which finally became uneconomic in 1966.

Both blackband and clayband ironstones have been much exploited in the Glasgow district. The Limestone Coal Formation is the main repository for both types with at least six of the former and three seams of the latter known to have been worked. The Johnstone and Garibaldi clayband ironstones and the Lower and Upper Garscadden, Jordanhill and Lower and Upper Possil blackband ironstones were the most extensively extracted. The Johnstone Clayband Ironstone was mined almost entirely south of the Clyde, the Lower Garscadden (or Linwood) Ironstone extensively both south and north of the river and the others mainly north of it. Mining of the blackband ironstones began about 1830 and ended at Cadder No. 17 Pit [NS 5949 7165] in 1923. Clayband ironstone working may have begun even earlier and ended at Victoria Pit [NS 521 506] in 1921. The only other group to contain workable ironstones is the Lower Limestone Formation, which has the Lower and Upper Househill clayband ironstones in south-west Glasgow and the latter's equivalents, the Campsie Clayband Ironstones, north of the city. At least one attempt was made to work an ironstone of unknown type, possibly above the Castlehead Lower Coal, in the Lawmuir Formation south of Paisley but it was quickly abandoned over 100 years ago. Mining essentially became uneconomic when thick, low-grade deposits, such as the Mesozoic ironstones of England, became available.

Several limestones in the Lawmuir Formation and in the Lower and Upper Limestone formations have been quarried and some have been mined as well. Most of the workings have been abandoned for over 100 years. The stratigraphically lowest is the Hollybush Limestone, which was quarried at Hollybush [NS 495 612] , Brownside [NS 489 606] , Brediland [NS 461 629] , Elderslie [NS 447 631] and Garnieland [NS 481 696] , where it also appears to have been mined. The Blackbyre Limestone was quarried at Nethercraigs [NS 467 610] . The Baldernock Limestone was quarried at Blackhall [NS 497 628] and mined at the Linn of Baldernock [NS 591 757] , where the stoop-and-room workings are still visible. The Hurlet Limestone was quarried and mined in several places, for instance at Hurlet [NS 513 611] , Nethercraigs, Duntocher [NS 422 728] , Windyhill [NS 523 766] and on the South Brae of Campsie around Newlands [NS 608 766] , mostly as a combined seam with the Hurlet Coal and the Alum Shale, though not as extensively as the coal alone. In the Upper Limestone Formation, the Lyoncross Limestone was mined to a small extent at Waulkmill Glen [NS 521 581] . The Orchard Limestone was quarried and to a small extent mined as a cementstone at its type locality [NS 562 590] . The Calmy (or Arden) Limestone was extensively quarried and latterly mined at Darnley [NS 525 589] , where it is unusually thick (3-4 m). Operations did not cease until 1960 when the more accessible resources were worked out, making this much the most recent limestone working in the district.

Several beds of cornstone in the Kinnesswood Formation have been quarried on a small scale and burnt locally, the most extensive workings being on Dumbarton Muir [NS 438 807] .

Sandstones of Lower and Upper Devonian age have been quarried for building stone in the northern part of the district. The former was worked in several quarries near Gartocharn [NS 43 86] where purple-red sandstone was extracted. The latter was worked at Croftamie [NS 478 857] and from several quarries on Blairquhomrie Muir [NS 430 825] exploiting bright red, easily dressed sandstone. Auchincarroch Quarry [NS 423 819] , just to the west of the district, had a working face about 50 m high.

Sandstones from the Lawmuir, Limestone Coal and Upper Limestone formations and the Upper Coal Measures [NS 606 612] have also been quarried, in many cases for local use only. The Nitshill quarries, in two sandstones in the Limestone Coal Formation (Nitshill and a lower one) were quite extensive. However, the two most important sandstones in the district, the Bishopbriggs and the Giffnock, are both in the Upper Limestone Formation. Both were very extensively quarried and subsequently mined for building stone at their type localities, Bishopbriggs [NS 609 695] and Giffnock [NS 569 592] . They were abandoned as uneconomic early in the 20th -century, leaving galleries 15 m high.

The Douglas Muir Quartz-Conglomerate Member, which occurs in the lower part of the Lawmuir Formation, crops out north and north-west of Glasgow. Currently, the deposit is being worked at Douglas Muir [NS 526 748] and Strathblane [NS 574 748] . The pebble content of the conglomerate is almost entirely vein quartz and, when disaggregated, produces a very low shrinkage aggregate and good clean sand suitable for glass and foundry work (Cameron et al., 1977).

Mudstones in the Lower Limestone Formation have been extensively quarried for brick clay around Blairskaith [NS 595 755] and at Fluchter [NS 588 747] . Several seams of fireclay in the Lawmuir Formation were mined at Paisley from the Ferguslie [NS 4606 6346] and Caledonia [NS 4751 4695] pits. Working did not finally cease at the former until 1949. The Darnley Fireclay, a short distance below the Calmy Limestone, was worked locally near Darnley. The fireclay was used to make high-grade sanitary ware. The Alum Shale occurring between the Hurlet Coal and the Hurlet Limestone was mined around Hurlet [NS 513 611] , from the Victoria Pit [NS 521 506] and extensively under the South Brae of Campsie as a source of alum. The Lillie's Shale Coal near the top of the Lower Limestone Formation is a combined seam of coal and oil shale that was mined quite extensively in the Linwood area in the 19th century.

Two quartz-dolerite dykes were formerly quarried at Rashielee [NS 465 709] for roadstone. Alkali dolerite sills have been quarried at a number of places in the Paisley–Johnstone area including Barr Hill, Kilbarchan, but the only quarry still in operation is High Craig [NS 427 613] (Merritt and Elliot, 1984), where the products are used for roadstone and concrete aggregate. The reserves are considerable.

Lavas and associated vents and intrusions of the Clyde Plateau Volcanic Formation have been worked in several areas e.g. Craigengaun [NS 525 769] , Craigend [NS 555 776] near Milngavie and Boylestone [NS 492 598] at Barrhead. Currently, only Milton Hill [NS 435 746] , working a plug-like intrusion, and Dumbuckhill [NS 420 747] , working vent agglomerate and a dyke-like intrusion, are operating.

A summary of clays of superficial origins in Central Scotland that are suitable for the manufacture of bricks is given by Elliot (1985). These resources are also suitable for pipe and tile manufacture. The clays in the Glasgow district are of several types including till, and late-Devensian glaciomarine and glaciolacustrine deposits. Flandrian estuarine, lacustrine and fluvial clays also are present but are of insufficient volumes to be significant resources.

Till is by far the most widepsread Quaternary deposit in the district, both at surface and underlying younger deposits. Where till was worked for brickmaking, or canal bank puddle-clay, the stones and boulders were usually handpicked. In brickmaking, bedrock mudstone was sometimes added. Till of the Wilderness Till Formation was quarried at Hamilton Hill [NS 584 675] and Springburn Brickworks [NS 605 672] .

Late-Devensian glaciomarine clays of the Linwood and Paisley formations have been extensively worked for brick, tile, pottery and pipe manufacture in the Glasgow, Paisley and Rutherglen areas. The history of quarrying extends back several centuries with the last operation at Polmadie [NS 604 624] only closing in the 1960s. When used for bricks, the clay was sometimes mixed with sand and ash. The clay in the Paisley area is known to make a good red facing-brick. It was formerly used at the Belvidere works in east Glasgow to manufacture strong sewer pipes. There are extensive resources remaining in the Linwood area which are still in greenfield locations. Both here and within the urbanised lower slopes of the lower Clyde valley, these clays are generally 15-30 m thick.

Late-Devensian glaciolacustrine clay of the Blane Water Formation has locally been worked for brick, tile and pipe manufacture just to the north of the district. These deposits extend southwards into Strath Blane, and are up to 20 m thick. However, they are located in an area of scenic value. In the south of Glasgow, at Williamwood [NS 571 581] , an older clay of the Bellshill Formation has been worked on a small scale.

A summary of resources of sand and gravel in the district is given in Cameron et al. (1977) and Browne (1977). The principal resources are confined to the valley floors and sides of the rivers Clyde and Kelvin, and the White Cart, Glazert, Endrick and Blane waters.

Moundy glacial sand and gravel is of little significance in the district and has only been exploited locally above the water table. The most extensive occurrences are near Strathblane [NS 55 79] in the Broomhouse Formation, and near Gartness [NS 49 86] in the Drumbeg Formation. The former are of Dimlington Stadial ice-contact glaciofluvial/glaciodeltaic origins, the latter are part of the local Loch Lomond Stadial terminal moraine and are of glaciolacustrine deltaic origin.

The most extensive sand and gravel deposits are mainly of Flandrian age and form the floodplains and terraces of the main rivers and streams. In the Clyde valley under Renfrew and Glasgow, the Quaternary succession, although dominated by glaciomarine clay, is commonly capped by about 6 m of sand, or sand and gravel. The deposit is commonly rich in coal debris and plant remains and is therefore of inferior quality. Historically, it has been exploited on a minor scale, including possibly for glass manufacture. The deposits of the White Cart, Endrick and Glazert waters may he generally coarser than those in the Clyde valley. They are also largely below the water table.

The most exploited deposits of sand and gravel in the district are those of the glaciofluvial/deltaic Cadder Formation. This formation is associated with the 'buried channel' of the River Kelvin. In general, the deposit is overlain by a variable thickness of till overburden (Wilderness Till Formation), probably the main reason why there are currently no active workings. The Cadder Formation is known to extend from Clydebank [NS 50 71] to Cadder [NS 61 72] and eastward into the Airdrie district. Now largely landfill, the main workings were at the Wilderness [NS 600 725] and Cadder [NS 606 721] . The former settling ponds at the Wilderness are currently exploited to provide 'fines' for cell construction at nearby active landfill sites.

Peat has been worked on a small scale on Cameron Muir as a domestic fuel and in historic times to burn concretionary limestone (cornstone) to produce lime. Although there are extensive deposits of hill peat in the Kilpatrick Hills and Campsie Fells, they are generally thin and also remote of access. The remnants of raised mosses in the Linwood and Paisley areas are as much as 6 m thick locally, but both Linwood [NS 44 66] and Fulwood [NS 445 690] mosses are partly covered by landfill. Together with Barochan Moss [NS 430 685] , these are now recognised as scarce ecological habitats.

Mean average annual rainfall is less than 1000 mm in Rutherglen and 1200 mm at Barrhead and Milngavie, but over 1800 mm on top of the Campsie Fells. Potential evaporation and transpiration by plants may amount to 350 to 400 mm in low-lying areas, so that 600 to 650 mm and 800 to 850 mm of effective rainfall is available at Rutherglen and Barrhead respectively. The greater part of the effective rainfall supports run-off and stream flow, but up to a third may become groundwater by infiltrating the soil and drift profile to reach the water table. This amount is equivalent to 200 to 280 mm/a day or enough to sustain a springflow of 9 1/s for each square kilometre of catchment. The widespread distribution of till within the district may, however, inhibit recharge in certain areas.

Groundwater has never been widely used in the district despite its favourable recharge potential. The reasons for this are the abundance of surface water supplies on the outcrop of the Clyde Plateau Volcanic Formation, the availability of piped water from Loch Katrine to Glasgow from 1859 onwards and the generally poor quality of groundwater within the Carboniferous sedimentary rocks.

Glasgow is the focal point for much of the groundwater discharge from the Central Coalfield (Robins, 1990) and therefore standing groundwater levels are near the surface throughout much of the city. The prevailing groundwater flow paths are from the east, northeast and south-east. The hydraulic conductivity of the Carboniferous rocks is of the order of 1 m/d in the more arenaceous horizons but may be only 10 ^-2 m/d or less in the mudstones and coals. The groundwater flow potential of these strata is enhanced by the presence of faulting, fractures and abandoned mine workings.

Lavas and tuffs of the Clyde Plateau Volcanic Formation crop out to the north and south of Glasgow. These rocks form the Kilpatrick Hills and Campsie Fells to the north of the city, beyond which the Devonian sandstones straddle the valleys of the Endrick and Blane waters. Neither the lavas nor the sandstones offer any significant groundwater potential, although there are some isolated spring and borehole sources. Groundwater flow is predominantly local, from beneath high ground towards lower-lying areas, where discharge may occur as baseflow to the rivers.

In the district, there are no records of water boreholes which penetrate rocks of Lower Devonian age. There is, however, a group of nine exploratory water boreholes in these rocks some 3 km farther to the north-west near Balmaha (Bird et al., 1992). They all encountered Garvock Group channel-phase sandstones with some overbank mudstones. There are three groups of three boreholes: Balmaha car park [NS 421 910] , Milton of Buchanan [NS 448 905] and Auchingyle Farm [NS 430 907] . The transmissivity of the aquifer is 2 to 4 m ² /d and the storativity is 10 ^-5 . Borehole specific capacity is, for the most part, less than 0.1 l/s/m and maximum sustainable yield from any one borehole is 31/s. Geophysical fluid logging identified fissure inflow to the boreholes and water quality which differed from one fissure to another. At shallow depths the water is Ca-HCO ₃ type and deeper water is NaC1-SO ₄ (C1) type. Sulphate levels are elevated due to the local presence of barite and the deeper more-saline waters may reflect the presence of a small residual element of Quaternary sea water.

The Upper Devonian and oldest Lower Carboniferous rocks are equally disappointing. In Fife, the Stratheden

Group contains up to 170 m of aeolian sandstone, the Knox Pulpit Formation, in which boreholes sustain abstraction of up to 40 1/s. In the west, the equivalent aeolian sandstones are interbedded with fluvial ones (Stockiemuir Sandstone Formation) and the prospects for groundwater development are far more modest. One successful borehole is recorded in the Upper Devonian in the district, at Glengoyne Distillery [NS 526 825] . However, the sustainable yield is only 1 1/s and the borehole specific capacity is 0.2 l/s/m.

Some groundwater is present in cracks and joints in the Clyde Plateau Volcanic Formation. It is not generally feasible to drill boreholes to intersect the water-bearing fissures, although boreholes with sufficient capacity to support domestic premises do exist in parts of Renfrewshire. Wherever water-bearing fissures intersect the ground surface there may be spring flow; there are many springs and seepages. Water quality may be iron and manganese rich where oxygen has been depleted in older circulating waters. Equally, it may be weakly mineralised young water suitable for bottling, as is the case at Lennoxtown (a spring-field around 645 789) just to the east of the district.

Groundwater potential is poor in the overlying younger Carboniferous sedimentary rocks around Glasgow. At the base of the sequence, the sandstones of the Strathclyde and Inverclyde groups formerly sustained 2 and 6 l/s from adjacent boreholes in Paisley [NS 457 628] . However, these yields are exceptional and are caused by the proximity of heavily faulted rock in the Paisley Ruck. Yields elsewhere in the sandstone rarely attained 1 1/s, although several of them overflowed when pumping ceased.

The majority of water boreholes in Glasgow penetrated the Lower Limestone, Limestone Coal or Upper Limestone formations. Sustainable abstraction rates were very small (less than 1 1/s) except where old mine workings were encountered. In 1985 a series of four bore-holes were drilled at Govan for water supply to the Glasgow Garden Festival which was held the following year. The first borehole [NS 5668 6475] , 198 m deep, penetrated thin coaly mudstones, shale and sandstones belonging to the Upper Limestone Formation. The hole was tested at 9 1/s and the borehole specific capacity is 0.9 l/s/m. Three production boreholes followed [NS 5645 6490] ; [NS 5680 6487] ; [NS 5730 6482] , each with a production yield of only 1 l/s. The first borehole made contact with water-bearing fissures, yields from the other boreholes being entirely dependent on the intergranular permeability of the formation.

Water quality from the Carboniferous sedimentary rocks is poor, characteristically very hard and reducing with abundant iron and manganese in solution. Shallow mine workings encouraged dewatering of much of the near surface strata, allowing oxidation of pyrite to occur. Subsequent flooding rapidly took the newly formed hydrous iron oxides into solution. This process occurred throughout the 19th century; groundwater supplies with deteriorating quality were being abandoned from the 1860s onwards in favour of Loch Katrine mains supplies.

The response of potential users to the poor groundwater quality is illustrated by the brewers. Only three breweries, Hugh Baird & Co., Castle Brewery and J & R Tennant used groundwater, but all had to blend it with Loch Katrine water and none pursued its use longer than was economically necessary.

Groundwater in the drift deposits of the district has never been exploited. Permeable gravels are present around Bearsden and towards Balmore as well as along the valleys of the Endrick and Blane waters near Killearn and the River Clyde in Glasgow. The gravels in Glasgow offer a perched aquifer of limited resource potential, highly vulnerable to urban and industrial pollutants, not least leaking sewers. The gravels in the north, around Killearn, offer a better prospect, although vulnerable to agricultural pollutants; they are, however, only partly saturated and may be covered by alluvial silt and clay.

A particular problem in Glasgow is the presence of methane in solution in the groundwater. A 183 m-deep borehole into the Limestone Coal Formation to the north-east of the university [NS 579 672] was described in 1939 by Dr A G MacGregor as follows:

The bore gives off gas continuously, and to get rid of it a jet is kept burning. The flame is about 4 inches (100 mm) long, and yellowish. An iron pipe sticks up in a concrete trough in the courtyard. The diameter of the pipe is about 4 inches (100 mm); it is plugged at the top with two short bits of piping (0.25 to 0.75 inches i.e. 12 to 18 mm in diameter) let into the plug. The gas burns at the top of these ... tried putting cement over the hole but it bulged up by the gas pressure. The flame is protected from blowing out by a loose cylindrical collar'.

Other borehole records suggest that methane gas was a widespread problem, particularly in water from the Limestone Coal Formation.

Much of the coal beneath Glasgow has been removed by stoop-and-room working. Long abandoned and flooded, slow degradation of the pillars in these workings may lead to sudden ground failure. Cement injection and other packings are used to support both old and new buildings. These supports may divert groundwater to other routes so promoting chemical degradation and erosion of pillars elsewhere. The problem is illustrated by an inspection of 114 bore logs in the Hillhead area where only the Knightswood Gas Coal has been worked. The results of the study carried out by I H Forsyth are given in (Table 8) .

The Knightswood Gas Coal is generally not more than 80 cm thick. However, the presence of water-filled voids and collapsed wastes at shallow depths has serious implications for ground stability. The role of groundwater as an agent of degradation is unclear, but the opportunity exists for chemical interaction between water and rock, as well as physical erosion by flowing water.

The hydraulic properties of the drift are of greater significance in terms of civil engineering, particularly with regard to landfill. Early landfill in the Glasgow area took place in the numerous sand and gravel and brick clay pits, and in building stone quarries in and around the city.

The search for waste disposal sites has now spread to and beyond the city limits. The Wilderness Landfill at Buchley [NS 59 72] was formerly a sand and gravel pit. In the late 1970s it suffered leachate loss through the side walls into the River Kelvin, a problem resolved only by careful engineering as the landfill progressed. Another difficult site was at Balmuildy [NS 574 716] on the flood plain of the River Kelvin. Again, sand and gravel lenses within the clayey drift allowed leachate access to the river. The adjacent landfill at Blackhill Brickworks [NS 578 715] was engineered in the early 1980s to allow leachate access to the Upper Limestone Formation via the sands and gravels. Next door at Summerston [NS 575 714] , the philosophy had progressed by the mid-1980s to containment of leachate; an impermeable liner was used beneath the new landfill. All contemporary landfill development requires full containment engineering.

The hazards associated with the build-up of methane gas (CH ₄ ) concentrations in landfill sites are now well appreciated (Campbell, 1988). Examples where domestic waste infill has subsequently been scaled by fine-grained cohesive soils, are numerous. Plans to extract gas commercially to heat greenhouses from landfill sites at Summerston [NS 575 715] , Wilderness [NS 600 723] and Mavis valley [NS 595 713] were in hand during 1993.

There is a considerable residue of contaminated land in the Glasgow district. A great deal of reclamation and development has taken place on the less-contaminated sites, but several former gas works and chemical installations remain. Although infiltration of undesirable chemicals to the groundwater causes little threat to the few groundwater users, emergence as baseflow to rivers does cause a potential problem. However, in most cases it is almost impossible to trace individual sources of pollution due to the effect old mine workings have on prevailing groundwater flow patterns.

There are no reflection surveys in the district, but surveys in adjacent areas include the IGS82 MV1 and MV2 Vibroseis lines recorded to 6 s two-way time (Penn et al., 1984, figs 3.2 and 4.1). Such surveys generally show coherent reflectors only to 1-2 s two-way time, equivalent to depths of 1.5-3.0 km and seldom below the base of the Clyde Plateau Volcanic Formation (Hall, 1971; 1974; Evans et al., 1988). Penn et al. (1984) found the lavas to be approximately 600 m thick, with their top about 1 km deep. The deepest reflector, at about 2 km, was interpreted as the top of the Lower Devonian.

Andrew (1978) reported an experimental deep reflection survey near Kippen [NS 5807 9130] where depths for the bases of the Devonian and Lower Palaeozoic rocks were estimated as 4.4 km and 8.6 km respectively. The 4.4-8.6 km interval had a velocity of 6.4 km s ^-1 (Table 9) , suggesting that it is Lewisian crystalline basement rather than Lower Palaeozoic rocks. The 4.4 km-deep reflector could then be interpreted as the base of the Devonian, as supposed by Andrew (1978), or the base of the Lower Palaeozoic.

The early LISPB and LOWNET refraction surveys were of low resolution, but provided deep penetration of 20-40 km (Bamford et al., 1978). Although centred east of the district, their results are relevant to the whole of the Midland Valley. LISPB revealed a 5.8 km s ^-1 refractor ('a ₁ '), at a depth of about 3 km, thought to be the top of the Lower Palaeozoic (Bamford et al., 1978; Barton, 1992). A 6.4 km s- ¹ refractor ('a ₁ '), about 7-8 km deep, was interpreted as the top of crystalline basement of Lewisian type.

These surveys were followed by the shorter MAVIS and quarry blast surveys, of which only MAVIS 2 is located in the district (Figure 30) . The MAVIS surveys detailed refractors to depths of about 12 km (Conway et al., 1987; Dentith and Hall, 1989). The quarry blast surveys had the shortest lines, giving detailed information to depths of 3-6 km (Davidson et al., 1984; Dentith and Hall, 1990). These surveys have been interpreted in terms of four main refracting layers, two of which ('a ₀ 'and 'al') are now differently identified from the original LISPB interpretation. This implies a significantly thinner Lower Palaeozoic succession and a shallower crystalline basement. The four layers are as follows:

1. 0.5-3.0 km thick Carboniferous and Upper Devonian strata (velocities in the range 3-5 km s ^-1 )
2. 1-5 km thick Lower Devonian and Lower Palaeozoic strata (4.8-5.4 km s ^-1 )
3. LISPB layer 'a ₀ ', 3-4 km thick, top at 3-6 km, gneissose crystalline basement (5.8-6.2 km s ^-1 )
4. LISPB layer 'a _l ', top 7-9 km deep; an intrabasement refractor, possibly due to a downward increase in metamorphic grade from amphibole to pyroxene granulite facies, (6.4 km s ^-1 )

Interpretations of the MAVIS and quarry blast surveys by Dentith and Hall (1989; 1990) suggested d é collement at the top of the basement and, locally, in layer 1. Faults in layer 1 may be listric, soling out into the possible d é collement at its base (Dentith and Hall, 1990). In the Glasgow district the tops of layers 2, 3 and 4 on MAVIS 2 are at depths of about 2.5, 4 and 7.5 km respectively. The suggested nature of the basement is in accord with that inferred from magnetic indications.

Detailed resistivity and seismic refraction surveys have been used in the area for assessing rock masses for quarrying (Wattananikorn, 1978).