For more than two centuries the rocks of the Southern Uplands have provided successive generations of geologists with both illumination and frustration. This area furnished vital evidence to support Hutton's ideas of petrogenesis and his cyclical view of earth history. However, for much of the succeeding century construction of a credible and consistent stratigraphy and structure eluded the best efforts of the many workers drawn to this region. It was an English-born schoolmaster, Charles Lapworth, inspired by the ideas of foreign practitioners and fascinated by the hitherto-despised graptolite, who applied a technique (zonal stratigraphy) and developed a structural paradigm (isoclinal folding) that provided an elegant and coherent solution to the myriad problems bequeathed by previous attempts.

So persuasive was Lapworth's model that the Geological Survey deployed its two finest mappers, Peach and Horne, to re-examine the entire region, using Lapworth's techniques. The outcome of that 10-year task is the monumental publication commemorated here. That work, and the ideas it espoused, remained virtually unchallenged for more than half a century and contributed to the application and elaboration of then-new geotectonic ideas, such as the geosynclinal concept.

However, in the 1950s the application of new or neglected techniques (way-up criteria, turbidite sedimentology, greywacke petrography, microstructural analysis, etc.) led to drastic reappraisal of prevailing structural and stratigraphic models and introduced a new paradigm (colloquially termed the ‘Southern Uplands paradox’) that envisaged a dominant role for strike-parallel major reverse faults. In contrast to Lapworth's shale-based approach this new model focussed on evidence derived predominantly from the thick intervening greywacke sequences. Investigations led by the Edinburgh and St Andrews schools extended and amplified this new interpretation of Southern Uplands geology, elucidated details of the palaeogeographic setting and the evolution of both the depositional basins and the source areas, and suggested comparisons with other parts of the Caledonian-Appalachian orogen.

The advent of plate tectonics revived interest in the Southern Uplands, first as a candidate subduction-related margin, then in the late 1970s an accretionary prism origin was proposed for the Southern Uplands imbricate thrust stack. The attractions of this hypothesis were manifest and it stimulated renewed activity by academics and the Geological Survey. Significant inconsistencies and perturbations in the simple accretionary prism concept have emerged from these more detailed studies and a range of convergence zone scenarios has been proposed. A Geological Society Meeting on the topic in 1986 furnished much new data and ideas but failed to yield conceptual consensus. Thus, as yet the latest Southern Uplands Controversy remains unresolved.

Changes in the level of geological research activity in the Southern Uplands can be assessed from an analysis of the numbers of relevant publications appearing over five-year intervals. This survey reveals a pattern that broadly accords with the narrative outlined above and supports the concept of two complete and one as-yet incomplete cycles of model development and adoption. Each of these developmental cycles appears to follow the progression of stages in scientific development identified by Thomas Kuhn and common to many scientific disciplines.

Lapworth, at the time of his 1878 paper on the Moffat Series, was the world's foremost expert on graptolites, and in that paper he gave the first great demonstration of the biostratigraphical value of graptolites. Peach & Home's resurvey of the Southern Uplands of Scotland extended Lapworth's ideas and his use of graptolites across the entire region. Peach's graptolitic work for the Survey is discussed: even though he identified a smaller repertoire of graptolites than Lapworth had, and often identified their general horizons rather than exact zone, his results are considered broadly correct. His faunal lists often emphasise the oldest faunas from the Moffat Shale inliers, presumably in order to stress their supposedly anticlinal structure. Subsequent work has seen a great extension of graptolite taxonomy and provided more detailed biostratigraphical subdivision, especially in the Silurian. The model of the Southern Uplands as an imbricate thrust stack is constrained by identifying the youngest (rather than oldest) fauna from the Moffat Shale inliers or, where possible, graptolites from the overlying greywacke formations. Such work has enabled the identification of about 25 thrust tracts in SW Scotland and of out-of-sequence thrusting in the Moniaive and Peebles areas to the NE.

The Southern Uplands is a major Ordovician-Silurian accretionary terrane which developed as a marine basin over a period of c.75 Ma (495-420 Ma). The terrane extends to c. 10,000 km2 in Scotland alone and correlates with the similar-scale Longford–Down terrane in Ireland. Despite tectonic complexity, a detailed lithostratigraphy has been erected. The oldest strata are mudstones, cherts and lavas of mid-Arenig age known only in the Leadhills Imbricate Zone. The next youngest rocks are of similar lithology but of late Llanvirn-early Caradoc age. These oceanic sediments are succeeded by black shales of the Moffat Shale Group which are, in turn, diachronously overlain by huge volumes of turbiditic sandstones, siltstones, mudstones and minor conglomerates (greywackes) of Caradoc to Wenlock age. Overall, the terrane is sandstone-dominated, with other components such as lavas, volcaniclastics and cherts representing only a tiny proportion of the total volume. The conglomerates have a broadly northerly provenance, whereas the sandstones exhibit both marginal (NW and SE), and axial (NE and SW) derivation. During the Ordovician, strongly contrasting sources alternated through time. The youngest sandstones (Hawick and Riccarton groups), are notably rich in detrital biogenic carbonate, a rare component in the Leadhills Supergroup and Gala Group.

Bounded by sutures and demonstrating a unique geological history and structure, the Lower Palaeozoic rocks of the Southern Uplands–Down–Longford form a definitive Caledonian suspect terrane. The geological history of the final closure of Iapetus is encrypted in their structural fabric.

Across the terrane, NW-younging turbidites predominate but graptolites invariably indicate the presence of younger sediments to the SSE. This fundamental Southern Upland paradox is soluble only by recognizing many strike-parallel faults, dividing the terrane into more than thirty tracts, each with its own variant of the stratigraphy and structure, and each having a lateral extent far in excess of what might be expected from the probable mechanical strength of the composing sediments. Structural interpretations of the terrane's unique tectonostratigraphic pattern are critically reviewed and the accretionary prism model, modified by strong sinistral transpression from the late Llandovery onward, is preferred. Transpression was apparently triggered when the converging continents of Laurentia and Avalonia made solid contact, so establishing a mechanically effective coupling of sialic crustal elements beneath and across the closing Iapetus ocean basin.

The geometry of the terrane's internal structural fabric is analysed. Tentative area-balancing calculations indicate a crustal shortening from a basin width of at least 1,000 km to the current terrane width of 75 km. Continuing sinistral transpression was expressed in fault reactivation and the development of a major shear zone. Late Palaeozoic strike-parallel extension produced W-facing half-grabens and the associated rotation may account for the easterly plunge of most fold axes.

The Scottish Caledonides, sited near the triple-junction between Laurentia, Amazonia and Baltica, is divided into at least five discrete blocks, each with a history incompatible with that of the block now lying adjacent to it. With the exception of the Hebridian margin, with its extensional Torridonian basins and Cambrian passive margin sequence, all blocks have undergone terrane-scale movements. The Moine, Central Highland Division and the bulk (if not all) of the Grampian Group have shared a common regional metamorphism, involving thickening and uplift, at c. 800 Ma and c. 450–480 Ma. This is incompatible with their being in the extensional regime that appears to characterise much of Neoproterozoic Laurentia. They, along with the polymetamorphic Dalradian block, now replace a region of passive margin and an unknown width of attendant Iapetus oceanic crust. These metamorphic blocks are grossly out-of-place.

The Midland Valley is a severely contracted block of ancient crust, once fringed by extensive oceanic basins to the N and S. An Ordovician–Devonian arc was founded on this older craton, and supplied sediment to basins on either side of it. This arc, during its Lower Palaeozoic life, matured, finally to yield relatively quartz-rich sediment, but was re-activated during the Devonian. An arc, similar to that of the Midland Valley, also supplied sediment to the Southern Uplands. Metamorphic debris in the Southern Uplands had a provenance in either this arc-basement or in a basement somewhere along the orogen. No metamorphic detritus in the Highland Border Complex has yet been demonstrated to have a Dalradian source.

Much of Scotland was assembled in a strike-slip regime. Evidence for strike-slip tectonics can be seen from the Late Proterozoic through to the Devonian. In periods of transtension, basins opened to accumulate sediment; in periods of transpression, those sediments were compressed and uplifted to yield sediments to successor basins. In the Neoproterozoic, during the phase of transpression, the basins were buried and metamorphosed, but during the Palaeozoic the basins were at a much higher level and escaped metamorphism.

A substantial volume of the Neoproterozoic–Palaeozoic sediment that accumulated in Scotland was derived from two orogens, both of which were sited some distance away. During the Neoproterozoic, the Grenville orogen was the main source, and in later (Devonian) time sediment accumulated in Scotland from the major, Late Palaeozoic continent–continent collision of Greenland–Scandinavia. These two external sources were augmented by a substantial contribution of sediment supplied from the Midland Valley arc or its lateral equivalent and by mild uplifts within the Scottish basements.

The Scottish Caledonides constitute less than 10% of the length of the Caledonian-Appalachian orogen, the rocks of which define one major margin of the Laurentian craton in Neoproterozoic-Palaeozoic times. Scotland was located, however, in a critical position at the tip of a major cratonic promontory bounded by the Caledonian and Appalachian segments of that margin. Isotopic dates from minerals and rocks collected in the Scottish Highlands have been regarded for 40 years as indicating a Neoproterozoic history of compressive orogenesis that is absent in N America and Greenland. They have therefore been taken by some authors to indicate an origin exotic to Laurentia for rocks of the Northern and Grampian Highlands S and E of the Moine thrust belt. An alternative explanation is that the Neoproterozoic rocks in the Scottish Highlands are all related to the two-stage ‘breakout’ of a discrete rift-bounded Laurentian continent from the core of the Rodinian supercontinent, believed to have assembled at the end of the Mesoproterozoic.

Traditional reconstructions of the late Neoproterozoic–Early Palaeozoic Earth oppose the proto-Caledonian/Appalachian margin of Laurentia and the W African craton of the newly assembled Gondwanaland. However, consideration of the global inventory of late Precambrian rifted margins, their relation to Grenvillian orogenic belts and of scale, leads to the hypothesis that the conjugate was the proto-Andean margin of S America. Recent recognition that the Cambrian and Lower Ordovician strata of the northwestern Argentine Precordillera and their underlying Grenvillian basement are unquestionably of Laurentian derivation, while not definitive, does point in this direction. If correct, this means that even the presence of Neoproterozoic orogenesis need not imply an exotic origin, as Neoproterozoic orogens are widespread in S America.

Traditional models show an Early Ordovician lapetus ocean basin approximately 4500 km wide, but the remarkably synchronous Ordovician collision of arcs and other terranes with the Laurentian and Gondwanan cratons from Argentina to the British Isles, suggests that this premise may be incorrect. The Appalachian–Caledonian orogen may rather have resulted from close and complex tectonic interaction between Laurentia and Gondwana, involving intervening volcanic arcs and other terranes. The interaction may have taken place during a clockwise transit of Laurentia around the proto-Andean margin to its late Caledonian–Scandian collision with Baltica, and the final suturing of Pangaea at the close of the Palaeozoic era. A modern analogue may be the interaction between Australia and Asia, involving intervening volcanic arcs and other terranes of the western Pacific Ocean basin, from ~ 50 Ma through the Present, and into the future.

A new model for the provenance, depositional environment and tectonic setting of the Northern Belt of the Southern Uplands is presented. This turbiditic sandstone-dominated sequence was deposited in a sand-rich submarine fan environment, overlying sparse hemipelagic mudstones. The oldest sandstones are rich in juvenile ophiolitic material and record the first clastic input into the Southern Uplands basin. The bulk of the Northern Belt sedimentary sequence, however, is dominated by relatively quartzose sandstones derived from a Proterozoic continental/metamorphic source represented by the Midland Valley terrane of Scotland and Ireland. The quartzose-dominated succession was punctuated by the input of fresh volcanic detritus shed from an oceanic/continental island-arc situated to the W/NW of the Northern Belt basin, with sediment dispersal turning to the NE along the axis of the basin in Scotland. The tectonic setting of the Northern Belt basin remains uncertain. The complex provenance of the sandstones and recognition of major olistostrome units within the Northern Belt succession suggest that it was tectonically active. The onset of clastic deposition within the Southern Uplands terrane broadly corresponds to uplift and erosion of earlier obducted ophiolite in both Scotland and Ireland, possibly in response to collision of Cambrian–early Ordovician island-arc systems with the Laurentian continental margin. If this interpretation is correct, then the possibility arises that the Southern Uplands–Midland Valley terranes record the dismembering of this oceanic/continental island-arc complex within an overall transpressional regime.

We summarise numerical and analogue models of shape fabrics, and discuss their applicability to the shape preferred orientation of crystals in magmas. Analyses of flow direction and finite strain recorded during the emplacement of partially crystallised magmas often employ the analytical and numerical solutions of the Jeffery's model, which describe the movement of noninteracting ellipsoidal particles immersed in a Newtonian fluid. Crystallising magmas, however, are considered as dynamic fluid systems in which particles nucleate and grow. Crystallisation during magma deformation leads to mechanical interactions between crystals whose shape distribution is not necessarily homogeneous and constant during emplacement deformation. Experiments carried out in both monoparticle and multiparticle systems show that shape fabrics begin to develop early in the deformation history and evolve according to the theoretical models for low-strain regimes. At large strains and increasing crystal content, the heterogeneous size distribution of natural crystals and contact interactions tend to generate steady-state fabrics with a lineation closely parallel to the direction of the magmatic flow. This effect has been observed in all threedimensional experiments with particles of similar size and for strain regimes of high vorticity. On the other hand, studies of feldspar megacryst sub-fabrics in porphyritic granites suggest that these record a significant part of the strain history. Thus, the fabric ellipsoid for megacrysts evolves closer to the strain ellipsoid than for smaller markers. This behaviour results from the fact that the matrix forms of the melt and smaller crystals behave like a continuous medium relative to the megacrysts. Consequently, in the absence of these markers, and because the fabric intensities of smaller particles such as biotite are stable and lower than predicted by the theory, finite strain remains indeterminate. In that case, strain quantification and geometry of the flow requires the addition of external constraints based on other structural approaches.

Subduction-related accretion of fault-defined tracts built up the Southern Uplands terrane during the final stages of closure of the Iapetus Ocean (Llandovery to Wenlock). Contrasts in depositional environment and pronounced differences in geochemical composition, provenance studies and metamorphic grade across the Laurieston Fault between the Gala and Hawick groups, suggests that it has a greater regional significance than most other tract-bounding structures. Initiated by underthrusting, and acting as a locus for subsequent sinistral strike-slip, the fault overlies a regional gravity anomaly gradient that is interpreted to be due, in part, to a concealed NW-ward dipping shallow basement surface. This is modelled as an open ramp in the NE that steepens to a near-vertical step along-strike to the SW. A change in structural geometry noted at the Laurieston Fault, with excision of accretionary tracts, is related to a period of oblique closure of the Iapetus Ocean. The youngest Gala Group tracts were accreted during a period of intense transpression to form a regional strike-slip duplex over the shallow basement ramp with termination of the tracts at the Laurieston Fault, its surface expression. The ramp acted as an obstacle to forward-breaking thrust progress, forcing the out-of-sequence thrusting and repetitive thrust imbrication noted in the eastern Southern Uplands. Upper Palaeozoic reactivation of this basement structure may have transferred strain between extensional Permian basins.

The progressive changes in the provenance of Silurian greywacke turbidites in the Southern Uplands terrane reflect geotectonic events at the Laurentian continental margin during closure of the Iapetus Ocean. In the northern Gala Group, juvenile andesitic detritus in some beds gives εNd values no lower than −4·2; more commonly, quartzo-feldspathic greywackes have εNd values in the −5·5 to −6·7 range, produced by the mixing of juvenile plutonic and Proterozoic basement detritus during arc unroofing. In the southern (younger) Gala Group, Proterozoic εNd values range down from −7·7 to −11·2 with only sporadic evidence for a juvenile component. An abrupt change is seen between the Gala Group and its tectonostratigraphical successor, the Hawick Group. In the latter, εNd values have a compact range between −4·7 and −6·6, indicating the renewed dominance of a more juvenile, plutonic provenance. Regional variations in the Sr/Rb ratio suggest that this was more evolved than the source of the Gala Group plutonic material. The Wenlock greywackes of the Riccarton Group have εNd values in the range −5·1 to −7·8, overlapping with the Hawick Group and with coeval greywackes from both the Midland Valley and Lakesman terranes. Overall, the data support proposals that the Iapetus Ocean had effectively closed by mid-Silurian times. Conversely, data from greywacke boulders in the basal Old Red Sandstone conglomerate of the Midland Valley terrane militate against its Wenlock juxtaposition with the Southern Uplands.

Analogue experiments were used to investigate pluton emplacement during transpression in a layered crust. Models consisted of (1) a silicone gum-PbO suspension as analogue magma, (2) a silicone gum-Pb suspension as a basal ductile layer, and (3) an overlying sand pack representing brittle crust. The models were transpressed at 3 mm/hr causing the extrusion of the analogue magma from a progressively closing slot, and its emplacement into the ductile layer. The thicknesses of the layers were critical in controlling the shapes of intrusions and the structures that developed in the brittle overburden. Thicker sand packs led to flattened, symmetrical laccolith-shaped intrusions and the nucleation of one oblique thrust in the sand pack above the extremity of the intrusion. Thinner sand packs led to thicker, asymmetrical laccolith-like intrusions with uplift of the overburden on an oblique thrust, and the formation of a shallow graben in the extrados of a bending fold. Reducing the thickness of the basal ductile layer resulted in a larger number of shear zones in the sand pack, and structural geometries approaching those produced in experiments involving only a brittle analogue crust and no ductile layer. Shear zones in the sand pack were localised by intrusions, and also played a key role in displacing analogue brittle crust to make space for intrusions. The results suggest that tectonic forces may play an important role in displacing blocks of crust during pluton emplacement in transpressional belts. They also suggest that pluton shapes, and the geometries and kinematics of emplacement-related shear zones and faults, may depend on the depth of emplacement. In nature, depending on the structural level exposed in the map plane, faults and shear zones that helped make space for emplacement may not appear to be spatially associated with the pluton.