The Western Isles–North Channel (‘WINCH’) deep seismic reflection profile runs through the North Channel across the extension of the Midland Valley into the Firth of Clyde.

A variety of shallow Carboniferous and younger sedimentary basins dominates the upper crustal structure. The deep structure includes a reflective lower crust, bounded downwards by the Moho, and dipping ‘thrusts’. No margins to the Midland Valley are observed equivalent to the bounding faults on land. Thrusts below both the Highlands and the Southern Uplands have a variety of dip directions indicative of zig-zag crustal shortening of 100 km in the Highlands and 60 km in the Southern Uplands.

The Moho varies only a little in depth, despite the gross crustal deformation in the Caledonides. Isostatic recovery of the Moho on unlocking of collided crustal blocks may explain the formation of some sedimentary basins.

I. ABSTRACT: Research interest in the Highland Border Complex has been pursued sporadically during the past 150 years. The results and conclusions have emphasised the problems of dealing with a lithologically disparate association which crops out in isolated, fault-bounded slivers along the line of the Highland Boundary fault. For much of the present century, the debate has centred on whether the rocks of the complex have affinities with the Dalradian Supergroup to the N, or are a discrete group. Recent fossil discoveries in a wide variety of Highland Border rocks have confirmed that many are of Ordovician age, and hence cannot have been involved in at least the early Grampian deformational events (now accurately dated as pre-Ordovician) which affect the Dalradian Supergroup. Such palaeontological discoveries form the basis for a viable biostratigraphical synthesis. On a regional scale, it is apparent that the geological history of the Highland Border rocks must be viewed in the context of plate boundary tectonism along the entire northwestern margin of Iapetus during Palaeozoic times.

II. ABSTRACT: Silicified articulate brachiopods from the Lower Ordovician (Arenig) Dounans Limestone are extremely rare but the stratigraphically diagnostic genera Archaeorthis Schuchert and Cooper, and Orthidium Hall and Clarke, have been identified. In addition, three specimens with characteristic syntrophiid morphology have been recovered. Inarticulate brachiopods are known from Stonehaven and Bofrishlie Burn near Aberfoyle, and have also been previously recorded from Arran.

III. ABSTRACT: Micropalaeontological investigation of the Highland Border Complex has produced a range of microfossils including chitinozoans, coleolids, calcispheres and other more enigmatic objects. The stratigraphical ranges of the species lie almost entirely within the Ordovician and reveal a scatter of ages for different lithologies from the Arenig through to the Caradoc or Ashgill, with a pronounced erosional break between the Llandeilo and the Caradoc.

IV. ABSTRACT: A Lower Ordovician (Arenig Series) silicified ostracode fauna from the Highland Border Dounans Limestone at Lime Craig Quarry, Aberfoyle, Central Scotland, represents the earliest record of this group of Crustacea from the British part of the early Palaeozoic ‘North American’ plate.

V. ABSTRACT: Palaeontological age determinations for a variety of Highland Border rocks are presented. The data are based on the results of recent prospecting which has demonstrated that macro- and microfossils are present in a much greater range of Highland Border lithologies than previously realised. Data from other studies are also incorporated, as are modern taxonomie re-assessments of older palaeontological discoveries, in a comprehensive survey of Highland Border biostratigraphy. These accumulated data demonstrate that all fossiliferous Highland Border rocks so far discovered are of Ordovician age, with the exception of the Lower Cambrian Leny Limestone.

VI. ABSTRACT: The Highland Border Complex consists of at least four rock assemblages: a serpentinite and possibly other ophiolitic rocks of Early or pre-Arenig age; a sequence of limestones and conglomerates of Early Arenig age; a succession of dark shales, cherts, quartz wackes, basic lavas and associated volcanogenic sediments of Llanvirn and ? earlier age; and an assemblage of limestones, breccias, conglomerates and arenites with subordinate shales of Caradoc or Ashgill age. At least three assemblages are divided by unconformities and in theirmost general aspect have similarities with coeval rocks in western Ireland.

The Highland Border Complex probably formed N of the Midland Valley arc massif in a marginal sea comparable with the Sunda shelf adjacent to Sumatra–Java. Strike-slip and thrust emplacement of the whole Complex in at least four episodes followed the probable generation of all or part of its rocks by pull-apart mechanisms.

Narrow, intermittent, fault-bounded outcrops forming the largely Ordovician Highland Border Complex comprise terrigenous-derived turbidities, a dismembered ophiolite, and ophiolite-derived sediments.

New major- and trace-element analyses of the mafic igneous rocks confirm that two main groups exist. One, represented in outcrops the length of the Highland Boundary fault-zone, has a mostly MORB-like chemistry with some trace-element compositions conventionally pointing to genesis above a subduction zone. The other group, found more locally, has an alkalic ‘within-plate’ character. Amphibolites interpreted as ophiolitic ‘sole’ rocks are chemically similar to the MORB-type mafic extrusive rocks. X-ray diffraction of the sedimentary rocks reveals kaolinite to be widespread and this is attributed to tropical weathering of the Highland Border Complex beneath a (?mid-Devonian) unconformity surface. New major- and trace-element analyses show that the turbidities of the Highland Border Complex were derived from a terrigenous terrane similar to that which supplied the Dalradian Supergroup. Inter-lava sediments reflect varied terrigenous (distal turbidites), hydrothermal (iron oxide sediments), mafic extrusive (volcaniclastic silt) and biogenic (jaspers) provenances. The ophiolite-derived Highland Border Complex sediments also have a terrigenous component. Unlike, for example, the early Ordovician rocks of the South Mayo trough (Ireland), coeval differentiated volcanic material is not a significant component of the Complex.

No one existing model adequately explains all the available data. We favour an origin of the Complex in the Ordovician as a small Gulf of California-type marginal basin which was later tectonically emplaced in stages involving a long history of alternating extension, strike-slip and compression.

Palaeomagnetic data define a polar wander path for the Dalradian tectonic block of the Scottish Highlands but resolve no relative motion between the Dalradian and Midland Valley blocks from early Ordo vician onwards. The history of structural episodes (D1 to D4) is traced for the Dalradian rocks of the Southern Highlands, and peak metamorphic temperatures are closely associated with D3 in the 520–490 Ma (early Ordovician) time period, whilst retrograde events are mostly associated with D4. Throughout the structural history, the influence of a lineament or lineaments parallel to the Highland Border is seen and is particularly demonstrated by the formation of the major Highland Border downbend (D4) structure. This monoform is interpreted as lying structurally above a major step in the basement and is associated with uplift of the Dalradian flat belt in the period 460–440 Ma (late Ordovician).

The mechanism of emplacement of the slivers of Highland Border Complex (HBC) rocks adjacent to the Dalradian Supergroup along the Highland Boundary fault zone is considered given that structural and palaeontological evidence indicates the emplacement to be post- D3 in the Dalradian rocks. The steep attitude of rocks along the Highland Boundary fault zone is probably due to vertical movements but either thrust or strike-slip movements are also indicated. Thrust emplacement prior to downbend formation (D4 in the Dalradian rocks) would also precede uplift and have to take place at considerable depth (10–20 km). Thrust emplacement subsequent to downbend formation is difficult to reconcile with the existing structural relations, whilst strike-slip motion offers a straightforward mechanism of placement.

The Dalradian terrane of Connemara was thrust southsoutheastwards about 460 Ma ago (Rb–Sr and K–Ar ages). It rides on a major thrust of post-D age over mylonitised acidic volcanic rocks of putative lower Ordovician age and contains a number of thrusts of similar age. Several major S- to SE-directed thrusts also limit the southeastern margin of the Dalradian rocks in Mayo and Tyrone. It is suggested by analogy with Ireland that during mid-Ordovician times the Highland Boundary fault in Scotland could have been a thrust zone which carried the Scottish Dalradian rocks over a lower Ordovician basement now represented only as fragments in the Highland Border Complex.

A summary of previous gravity and magnetic field interpretations is presented along with a review of the LOWNET and LISPB models. Recently acquired seismological data are compared with the LISPB upper crustal model. The LISPB Layer 1 is found to represent an average of lower velocity (VP = 3·0–3·7 km s−1) Carboniferous and Upper Old Red Sandstone sediments, and higher velocity (VP = 4·0–5·5 km s−1) Lower Old Red Sandstone and Lower Palaeozoic sediments.

On the basis of high P-wave velocities along strike (VP = 60·6–1 km s−1), the LISPB Layer 2 S of Glasgow and Edinburgh is re-interpreted as a quartzofeldspathic crystalline layer of igneous or metamorphic origin, which is also seen to cross the surface of the Southern Uplands fault at around 2·5 km depth.

This interpretation implies that within the southern part of the Midland Valley, the Lower Palaeozoic sediments are nowhere thicker than about 3 km, as postulated under the Silurian inliers, and generally much less.

The Ballantrae complex has yielded its oldest yet determined age of 576 ± 32 Ma from an Sm–Nd internal isochron for a block ‘eclogite’ interpreted as being a component of the mélange-olistostrome unit at Knockormal. This unit is possibly older than the main ophiolitic part of the complex and the ‘eclogite’ block provides evidence for the existence of oceanic crust older than c. 575 Ma on the southern margin of the Laurentian shield.

An Sm–Nd internal isochron from a high-grade metapyroxenite in the metamorphic aureole beneath an obducted slab of mantle peridotite at Laigh Knocklaugh yields an age of 505 ± 11 Ma. This is interpreted as close to the age of metamorphism and is older than the K–Ar age of 478 ± 8 Ma determined for the amphibolites of the aureole which underlie the metapyroxenite. The metapyroxenite had a provenance deeper in the oceanic crust than the amphibolite and the former has been structurally detached and justaposed against the latter in this telescoped metamorphic sequence.

The distribution of Arenig graptolite faunas and zones of differing metamorphic fades help to establish a polyphase imbrication history for the complex. A further complication is introduced by the presence within the imbricate stack of lava sequences of comparable age but with contrasting geochemical signatures of island-arc and oceanic island basalts, and oceanic island sequences of markedly different Arenig ages. Strike-slip mechanisms appear the most likely means of introducing such a disparate assortment of elements into the Ballantrae complex and an origin in a continental margin transform zone is preferred.

Major and trace element, including REE, analytical data are used as bases for interpreting the petrogenesis of the major igneous components of the northern part of the Ballantrae complex which occurs in the southwestern part of the Midland Valley of Scotland. Most of the peridotite, now serpentinised, is similar to ultramafic rocks in other ophiolite complexes. Mean crystallisation conditions, determined on the basis of co-existing orthopy-roxenes and clinopyroxenes for the dominant peridotite and minor pyroxenite were 1060 (±60)°C—20 (±2) kb and 1240 (±89)°C—25 (±25) kb, respectively. These rocks, of mantle provenance, have compositions consistent with being residues after the extraction of 20–30% of tholeiitic material from the mantle. The presence among them of a rock whose REE contents indicate that it is a plagioclase peridotite, point to the tectonic incorporation of the products of a high level magma chamber.

The mafic parts of the complex have tholeiitic characteristics and developed between 1300° and 1100°C. They do not represent primary mantle melt but fractionated material. Clinopyroxene was the main fractionating phase and more than 10% fractional crystallisation is indicated with increase from gabbros, through beerbachites (metadolerites) of a sheeted dyke complex and pillow lavas, to microgabbros and pyroxene diorites. Biotite diorites and trondhjemites represent the most fractionated products, the latter having affinities with ophiolitic plagiogranites.

The beerbachites of the sheeted dyke complex do not all represent the same stage of fractionation. The pillow lavas have REE patterns similar to rocks found in marginal basins but are markedly different from pillow lavas from the Highland Border Complex in Arran, near the northern margin of the Midland Valley.

Doleritic intrusions known to post-date the obduction of the Ballantrae complex during Arenig times record changing magma provenance during the cooling and serpentinisation of obducted peridotite. There are two groups of dolerites with different petrographical and geochemical characteristics.

The earlier emplaced group, which is subordinate, is characterised by amphibole formed under low metamorphic facies conditions and the virtual absence of Fe–Ti oxides. Chemically these rocks are of primitive tholeiitic character and are similar to modern island-arc basalts. They were derived from a strongly depleted mantle source region. The source region of the later emplaced group, which is predominant, was much less depleted. Clinopyroxene and plagioclase, much of it albitised, are the dominant minerals. Fe–Ti oxides also are common and there are pronounced chilled margins and well-preserved ophitic textures. Although showing mildly alkaline tendencies, these rocks originated from a tholeiitic parental magma of “within-plate” type.

Both groups show evidence of rodingitisation and associated alteration related to serpentinisation of the peridotite. During the metasomatic activity, only some elements were mobile, while Al2O3, total Fe, MgO, TiO2, Cr, Ni, Nb, Y and Zr remained almost constant. Assessment of original magma type and geotectonic environment, and demonstration that both groups of doleritic rocks show the products of fractionation, has relied heavily on data for the apparently immobile elements and on petrographical study which identified those rocks least affected by alteration.

The Kirkland Conglomerate and the Benan Conglomerate and associated sediments of the Barr Group (Ordovician—Llanvirn-Llandeilo) of the Girvan district, SW Scotland record the development of two fan-delta systems situated on the northern margin of Iapetus. The intervening Stinchar Limestone represents a shallow marine fan-delta abandonment facies. Subaerial fan-delta deposits are seen at the lowest exposed horizons within the Kirkland Conglomerate. Transgression and eventual abandonment of the fan-delta system is recorded by (1) matrix-rich gravels forming the topmost horizons of the unit. (2) subaqueous distributary channel sands of the overlying transitional sandstone and associated shallow-marine carbonates of the Auchensoul Limestone and (3) shallow-marine sandstones of the Confinis Flags. Following a period of shallow-water sedimentation (Stinchar Limestone), a phase of rapid subsidence occurred during the upper Llandeilo, broadly synchronous with the Nemagraptus gracilis Zone transgression. Lowermost horizons of the succeeding Benan Conglomerate comprise re-sedimented gravels and laterally equivalent ‘outer shelf’ deposits (Superstes Mudstone). Renewal of coarse clastic sedimentation resulted from source area uplift related to granite plutonism, responsible also for the rapid subsidence of basinal areas. Fan-delta progradation is recorded by the occurrence of braided-fluvial deposits and shallow-marine carbonates at higher stratigraphical levels within the unit. Progradation of the fan-delta complex resulted from a gradual reduction in subsidence rates along basin-margin faults.

Fan-deltas of the Barr Group prograded southwards, from a SW—NE-trending faultdelineated basin margin, across a narrow shelf area. To the S of the shelf area defined by Barr Group outcrop, sediments of the Tappins complex accumulated in outer shelf, ?slope and base of slope settings.

The thickness of individual conglomerate units (>150 m—Kirkland Conglomerate, and up to 700 m—Benan Conglomerate), the associated high sedimentation rates, narrowness of the shelf area, and distribution and style of basin-margin faults indicate that oblique-slip motion along the northern margin of Iapetus may have provided a major control over Middle Ordovician sedimentation in the Girvan district.

Old Red Sandstone sediments from Arran in western Scotland yield an array of Pb isotope compositions whose slope corresponds to an age of 1830 ± 400 Ma. These compositions are not consistent with a significant contribution of material from the Southern Highlands, Midland Valley, or Southern Uplands, but are similar to the Torridonian sandstones of NW Scotland. Bearing in mind the probable Laxfordian provenance of the Torridonian assemblage, this points to a source for the Old Red Sandstone in a crustal block of early Proterozoic age corresponding to the age of the Laxfordian episode. Since such a source is not presently in a position to yield the Old Red Sandstone, strike-slip movement along the Highland Boundary fault is a strong possibility.

A model for the early Palaeozoic metamorphic history of the Midland Valley and adjacent areas to the S in Scotland, England and Ireland is based on the results of new field mapping, thin section petrography, electron probe microanalysis, X-ray diffractometry, conodont and palynomorph colouration and graptolite reflectance measurement.

The oldest metamorphic rocks of the Midland Valley of Scotland, excluding xenoliths in post-Silurian lavas, are possibly the blueschist occurrences in the melange unit of the Ballantrae complex. These may be tectonised remnants of (?)pre-Arenig ocean-floor subducted during closure of the Iapetus Ocean. In the early Ordovician, the melange terrane was dynamothermally metamorphosed during obduction of newly-formed ocean crust. The obduction process piled up a thick sequence of various ocean-floor types such that burial metamorphism in parts reached pumpellyite-actinolite facies; elsewhere prehnite-pumpellyite and zeolite facies was attained.

Whilst the Midland Valley acted as an inter- or fore-arc basin during the Late Ordovician and Silurian and experienced burial metamorphism, an accretionary prism was formed to the S. Accretion, tectonic burial and metamorphism of ocean-floor and trench sediment was continuous in the Southern Uplands and the Longford-Down massif of Ireland through Late Ordovician to Late Silurian times. Rocks at the present-day surface vary from zeolite facies to prehnitepumpellyite facies. Silurian trench-slope basin sediments can be recognised in part by their lower grade of burial metamorphism. Greenschist facies rocks of the prism probably lie close to the surface.

The Lake District island-arc terrane of Northern England has an early Ordovician history of burial metamorphism up to prehnite-pumpellyite facies. The Late Ordovician and Silurian metamorphic history is one of sedimentary burial complicated by tectonism and intrusion of granite plutons to a relatively high level. The Iapetus suture is marked by a weak contrast in metamorphic grade.

New major- and trace-element data for granitoid plutons from the Grampian Highlands, the Midland Valley and the Southern Uplands of Scotland are presented and discussed. The study is restricted to ‘late granitoids’ (all younger than 430 Ma); the term ‘granitoid’ is used in a wide sense to encompass all plutonic components of a zoned intrusion of this age, sometimes including diorites and ultrabasic cumulate rocks. The data indicate that as a whole the province is chemically high-K calc-alkalic. Other notable enrichments are in Sr and Ba, and a marked geographical difference in these trace-elements is found between plutons of the SW Grampian Highlands and those of the Southern Highlands, the Midland Valley, and the Southern Uplands. Plutons of the NE Highlands tend to be more geochemically evolved than those further SW and those of the Midland Valley and Southern Uplands.

When petrographical and geochemical data are considered, three plutonic suites are recognised: (1) the Cairngorm suite comprising plutons of the NE Highlands, (2) the Argyll suite comprising plutons from the SW Highlands, and (3) the S of Scotland suite comprising plutons from the Southern Highlands, Midland Valley and the Southern Uplands excluding Criffell and the Cairnsmore of Fleet. It is proposed that the more acidic granitoids are dominantly the products of I-type crustal sources, but certain diorites and the more basic members of zoned plutons have a substantial mantle component. The elevated Sr and Ba levels in granitoids of the Argyll suite may reflect the influence of incompatible-element-rich fluids from the mantle in the petrogenesis of this suite. The relatively anhydrous pyroxene-mica diorites of the S of Scotland suite are richer in Ni and Cr and appear to represent mantle-derived melts. The relationships between these data and already published isotopic data are discussed.

The pre-Carboniferous Midland Valley of Scotland comprises three tectonic elements: an arc, a proximal fore-arc basin and a marginal basin. These tectonic elements have been juxtaposed by strike-slip and thrust faulting, both of which have effected a 300% reduction in the width of the orogenic belt.

Rocks which span Arenig to Late Devonian or Early Carboniferous times and which are found S of the Highland Boundary fault have no clasts of certain Dalradian provenance despite substantial uplift of the Dalradian block at this time. This, combined with other evidence, suggests the Midland Valley to have been remote from this rapidly uplifting terrane. The Dalradian block, eroded down by c. 410 Ma was thrust southeastwards in Late Devonian–Early Carboniferous times. However, this thrust movement was minor, yielding little sediment, but it caused Dalradian rocks to cover the northern margin of the Midland Valley where (1) the source for part of the Old Red Sandstone rocks existed and (2) the faults along which the Midland Valley block was transported to dock against the Dalradian block are thought to be present. The existing Highland Boundary fault is therefore seen as a late Old Red Sandstone reverse fault which covered more significant older structures.