The fossil site of Glencartholm (OS Grid Reference [NY376795]) is located in the Scottish region of Dumfries and Galloway, 0.5 km north of Glencartholm Farm along 50 metres of the east bank of the River Esk (Dineley Reference Dineley, Dineley and Metcalf1999) (Fig. 1). Fossils were first discovered at Glencartholm in 1879 by the famous collector A. Macconochie, during work for the Geological Survey (Geikie Reference Geikie1881). The fishes were first described by Traquair (Reference Traquair1881, Reference Traquair1884, Reference Traquair1890) and the arthropods by Peach (Reference Peach1882a, Reference Peachb). Further collecting took place between 1933 and 1936 by Moy-Thomas (Reference Moy-Thomas1934, Reference Moy-Thomas1935, Reference Moy-Thomas1936, Reference Moy-Thomas and Bradley Dyne1938) and Schram (Reference Schram1983), when the use of explosives removed all the accessible fossiliferous strata (Lumsden et al. Reference Lumsden, Tulloch, Howells and Davies1967).

Figure 1 Locality of Glencartholm exposure (located on inset map of Scotland) along the east bank of the River Esk.
Glencartholm has been the source of the richest Carboniferous fish fauna in the British Isles, and one of the richest in the world. It is a Site of Special Scientific Interest and is scheduled as a site of international importance by the Geological Conservation Review of the Joint Nature Conservation Committee (Dineley Reference Dineley, Dineley and Metcalf1999). The fauna is remarkable because of the number of complete fishes discovered, and the variety of forms (35 species). It is the only known locality for 20 species of fishes, and the type locality for a further six. The associated fauna includes marine brachiopods and bivalves, crustaceans and scorpions.
Despite its significance the section has not been described in detail. Peach & Horne (Reference Peach and Horne1903) identified two fossiliferous horizons: the ‘Shrimp Bed', containing crustaceans, fishes and molluscs, and the ‘Scorpion Bed', containing an extensive flora, as well as sparse crustaceans and arachnids. The beds occur as a thin unit within the Glencartholm Volcanic Member. This was formerly placed at the base of the Upper Border Group of the Calciferous Sandstone Measures (George et al. Reference George, Johnson, Mitchell, Prentice, Ramsbottom, Sevastopulo and Wilson1976), but it now forms part of the base of the Yoredale Group (Stone et al. Reference Stone, McMillan, Floyd, Barnes and Phillips2012). The Glencartholm Volcanic Member lies between the Holkerian and Asbian and is of mid Viséan age (Stone et al. Reference Stone, McMillan, Floyd, Barnes and Phillips2012). The rocks are poorly exposed, but a complete section was provided by the Archerbeck Borehole (Lumsden & Wilson Reference Lumsden and Wilson1961). This, however, did not encounter the fossil beds of Glencartholm. Even the local extent of the Glencartholm fish beds is unclear because of faulting at the site, and it has been assumed for many years that the fish bed was worked out (Lumsden et al. Reference Lumsden, Tulloch, Howells and Davies1967).
Stan first became interested in fossils in the late 1960s. Advised to direct his attention to the beach at Wardie in Edinburgh, where fossil fishes had been found in the 19th Century, he spent his leisure time for two years on that insalubrious shore before finding his very first fossil (Wood Reference Wood1975; Smithson & Rolfe Reference Smithson and Rolfe2018 [this volume]). Completely unschooled in palaeontology, he nevertheless, within only a further two years, had discovered hundreds of Carboniferous fish fossils, sold the collection to the National Museum of Scotland (NMS), embarked on an intensive programme of self-education, resigned his job and become a fossil hunter by profession.
The fishes of the Carboniferous were therefore his very earliest interest. He read widely on the subject seeking out original references and soon learned of the importance of Glencartholm. It was little wonder that Stan was anxious to see the place for himself. Determined to familiarise himself with the site, he first visited it in the early 1970s, only to find that the original locality was overgrown with trees, that there was extensive faulting and that all the fossiliferous beds had indeed been removed. Not at all deterred, he determined then that he would explore the locality himself one day.
However, many difficulties had to be overcome (S. Wood pers. comm. 2012). Obtaining permission to excavate was problematic. The Buccleuch Estates factor (or property manager) was not cooperative. Despite support from the Nature Conservancy Council and many colleagues, there was covert opposition from some members of the establishment, as happened occasionally during Stan's unorthodox career. Then there were the distractions. Over the years Stan was involved with major explorations at the Dora Opencast Site near Cowdenbeath, in association with University of Newcastle upon Tyne (Andrews et al. Reference Andrews, Browne, Panchen and Wood1977), Foulden (Wood & Rolfe Reference Wood, Panchen and Smithson1985); at Bearsden (Wood Reference Wood1982), in association with the Hunterian Museum; and at East Kirkton (Wood et al. Reference Wood, Panchen and Smithson1985). He set up the business Mr Wood's Fossils in 1983, was involved in putting on a travelling exhibition showcasing many of his discoveries, made appearances on television programmes and became a popular lecturer.
In the meantime, Stan revisited Glencartholm again many times to plan his own excavation there. On one occasion, he was accompanied by Fredrick Schram, who was working on the biota of Glencartholm (Schram Reference Schram1983) and discovered a black argillaceous limestone containing an almost complete actinopterygian fish, Acrolepis ortholepis (S. Wood pers. comm. 2012). The matrix and mode of preservation were similar to that of the ‘Shrimp Bed'. Nearby, black shales were exposed in the east bank to the north on the upthrow side of a fault that separates it from the original collecting site, and a section 4.87 m long was exposed. The fossiliferous unit consisted of fine-bedded black shales with two thin limestone bands, and yielded fossils throughout the section in a manner unlike that described for the original section (Schram Reference Schram1983). Twenty years were to pass before Stan's own exploration finally began in 1994.
1. Methods
1.1. Sources of information
Information was obtained from Stan's Wood's own notes and from conversations with him during 2012.
1.2. Selection of new site for exploration
The original type section in the river bank (Lumsden et al. Reference Lumsden, Tulloch, Howells and Davies1967) was examined to identify the sequence. The base of the fossiliferous sequence was marked by the top of the cherty matrix at the northern (upstream) end. The top was downstream, where large blocks of thick yellowish limestone had worked loose in and around the river bank. Stan predicted that the beds would continue in a forested area 50 m to the east of the riverbed exposure. This was the site selected for the new excavation which began in 1994 and lasted for two field seasons. Figure 2 shows Stan planning the excavation on site with Glencartholm farm in the background. The new locality (Fig. 3A) was named Mumbie Quarry, after the nearby settlements of Upper Mumbie, Nether Mumbie and Mumbie Cotttages.

Figure 2 Stan planning the excavation on site; Glencartholm Farm can be seen in the background.
1.3. Excavation
To allow access for collecting samples, a 1.25-m trench was dug close into the northern edge of the exposure, which allowed an uninterrupted view of the newly exposed rocks in cross-section. As Stan had predicted, it included a sequence with parameters identical to the original in the banks of the River Esk, with a cherty base and a limestone top. The trench was marked out with white index cards at one-metre intervals, number 1 being the metre of the section resting upon (but not including) the underlying chert marker, as shown in Figure 3B. Thus, the cards marked the top of each metre, beyond which was the lower part of the next metre. Rock samples were collected across the exposure and placed in bags labelled according to their position in the section. These are available for sedimentary analysis.

Figure 3 Excavations at Glencartholm: (A) Mumbie Quarry; (B) trench showing metre marking cards; (C) the over-burden put to good use surfacing forest tracks.
To gain access to the bedrock, a great deal of over-burden had to be removed, for which a mechanical excavator was employed. The extent of the excavation is shown in Figure 3A. The over-burden was put to good use surfacing adjacent forest tracks (Fig. 3C).
1.4. Fossil preparation
Fossils were removed from the section using standard hand tools. Mechanical preparation was undertaken using hand tools and, where necessary, breaks were prepared with glue. The vast majority of specimens were only taken out of storage for preparation and initial identification by Stan in 2012. During the years 2000–2006, Stan no longer had any preparation facilities, and from 2006 he was fully occupied with field work in the Scottish Borders (Smithson et al Reference Smithson, Wood, Marshall and Clack2012). It was just a month or so before his death when he became too weak to undertake any more field work that he finally returned to his Mumbie specimens.
2. Results
2.1. Geology
The rocks exposed in Mumbie Quarry dipped to the West at an angle of approximately 45°, occasionally but briefly shallowing or steepening. The dipping surface was essentially flat. An 18-metre section was examined in detail (Table 1). It comprises limestones, mudstones and shales, with occasional layers of tuff. Three separate fish horizons were identified. The top fish bed is a sequence of mudstones, 1 m thick, lying between levels 12 and 13. The middle fish bed is also 1 m thick and lies between levels 7 and 8. Level 7 contains large limey nodules without nuclei, and level 8 consists of mudstones and shale. The bottom fish bed is 10 cm thick and lies within layer 5. Due to the overlying rocks, the bottom fish bed was the least accessible and little could be collected. The middle fish bed was more accessible and a significant amount of overlying rock was removed to expose the surface. A large area of the top fish bed was exposed, enabling significant collecting. The number of vertebrates collected from each bed is shown in Table 2.
Table 1 Log of Mumbie Quarry

Table 2 Distribution of fish taxa throughout the top (T) middle (M) and bottom (B) beds

*This is the large shark found above the top fish bed in metre 14
2.2. Fossils
2.2.1. Catalogue
Tables 3 and 4 list the Mumbie fauna catalogued by Stan. The collection includes ray-finned fish (Fig. 4A), cartilaginous fishes, shrimps, scorpions, horseshoe crabs (Xiphosura), orthocones and conularids. A number of other invertebrates and plants have not been identified and are therefore not included in the detailed catalogue. The collection of fish included 217 individuals, many with part and counterpart. Twenty-nine different taxa were identified by Stan. He was unable to make a positive identification of 16 bony fish specimens and one cartilaginous chondrichthyan fish specimen.
Table 3 Vertebrates from the Mumbie excavation

Table 4 Invertebrates from the Mumbie excavation

*Now assigned to Tealliocaris (Clark Reference Clark2013)

Figure 4 Glencartholm specimens: (A) Rhadinichthys canobiensis from top fish bed, Mumbie; (B) Platysomus superbus from top fish bed, Mumbie.
2.2.2. Repositories for the collection
Sixty-five specimens were acquired by the National Museums Scotland (NMS). This collection includes six ray-finned fish, most of the cartilaginous fishes and xiphosurans, and some of the shrimps and plants. Ten specimens of ray-finned fish were acquired by a private collector. The remaining specimens (numbering over 300) have recently been acquired by the Natural History Museum in London.
2.2.3. Descriptions
Very few of the specimens collected by Stan at Mumbie have so far been described. The new specimen of Onychoselache traquairi (NMS 1998.35.2) provided significant new information which enabled Coates & Gess (Reference Coates and Gess2007) to present a new reconstruction and reassess its phylogenetic position. The new specimens of the holocephalan Chondrenchelys problematica (NMS 1998.35.1 and NMS 2002.68.1) have been described by Finarelli & Coates (Reference Finarelli and Coates2012, Reference Finarelli and Coates2014). This included the first unambiguous evidence of a gnathostome with an extra-mandibular dentition. The teeth of the large undescribed shark (NMS 2000.14.2), found above the top fish bed at Mumbie, were compared with the dentition of the holotype of a ctenacanthiform shark, Goodrichthys eskadalensis, by Ginter (Reference Ginter2009), who tentatively suggests they represent the same species.
There have been no publications on any of the other specimens from Mumbie. The following information has been taken from Stan's own notes (S. Wood pers. comm. 2012). On the fish, he comments that over 220 specimens of ray finned fish were collected from Mumbie Quarry, all from well-documented horizons. Fish were not found outside the confines of the three fish beds (see section 2.1), apart from the large shark collected from the limestone band above the Top Fish Bed. A straightforward comparison between the number of fish species in each of the three fish beds is hampered by collecting variables, the lower bed being the least accessible. During cataloguing, a marked difference was noted in the numbers of Cheirodopsis species between this study and that recorded previously; Cheirodopsis occurring only thrice in this survey, whereas Schram (Reference Schram1983) records 99 examples. It was noted that the deep-bodied form Platysomus (Fig. 4B) does not appear until the Top Fish Bed horizon. The rotten limestone band within the Top Fish Bed does not appear to affect the fauna contained above or below it.
A most notable find among the ray-finned fishes includes one with four outstanding and unusual black shiny pearl-shaped teeth spaced out along the dentary, the largest anteriorly, followed by a gradual size reduction posteriorly. The scales are deeper than broad resulting in quite a thin slender appearance. The fish is fusiform but not deeply so, the head stout. This may represent the first undescribed ray-finned fish recovered from the sequence for some 120 years.
Stan comments “Then came one of those rare ‘eureka' moments inspired by Providence herself. [While preparing the specimens] it slowly dawned on me that we could separate the two lower fish beds from the upper simply by colour definition! The latter are preserved with glossy black scales, the former two with light brown scales. Even the acanthodians can be distinguished too. What a key to the past provided by simple observation. We now know where in the 1880s Macconochie made his collections in the section and, indeed, all the others who followed and extended the fauna”.
There has been no attempt in Stan's notes to do anything other than list and identify, where possible, the remaining specimens of invertebrates and plants. They include 19 individual examples of xiphosurids, which have not previously been reported from Glencartholm, and over 60 shrimps.
3. Discussion
The vast majority of specimens have now been acquired by the National Museums Scotland and the Natural History Museum in London. It is hoped that the collection, together with the field notes, will facilitate future analysis and research.
The excavation at Mumbie produced a large number of new vertebrate and invertebrate fossils, most of which remain to be studied. The three vertebrate specimens described so far have yielded important new information and allowed a reassessment of chondrichthyan relationships and diversity (Coates & Gess Reference Coates and Gess2007; Finarelli & Coates Reference Finarelli and Coates2012, 2014). The description of xiphosurans will provide similar new insights into the morphology and relationships of horseshoe crabs. It is anticipated that detailed study of the rest of the Mumbie collection will provide a greater understanding of the radiation of vertebrates and arthropods in the Early Carboniferous.
The discoveries at Mumbie have shown that the Glencartholm deposit is more complicated than current literature suggests. There are three distinct fish beds, each with a rich fauna, and further fossiliferous horizons are present throughout the sequence, notably the limestone band in metre 14 containing the large shark and the yellow limestone in metre 18 containing the horseshoe crabs. Stan's observation on the differences in scale colouration of actinopterygians from the lower and middle fish beds, and those from the upper fish bed, should allow a reassessment of the temporal distribution of fishes. His findings may also help determine whether other taxa have patterns of distribution similar to that of Platysomus superbus, which appears to be restricted to the top fish bed.
The new discoveries at Mumbie have provided further evidence of the extraordinary diversity of the Glencartholm biota and the outstanding preservation of the fossils (Dineley Reference Dineley, Dineley and Metcalf1999). It is the earliest record of significant diversification amongst actinopterygians, following the end-Devonian extinction, with at least 25 different species (Schram Reference Schram1983). This compares with just four species at Foulden, the best known of the older Scottish Tournaisian localities (Gardiner Reference Gardiner1985). It is expected that detailed study of the actinopterygians from Mumbie will help account for this major radiation during the Early Carboniferous that laid the foundations of the modern fauna.
4. Acknowledgements
Thank you to Tim Smithson, Stan's friend and colleague, for his advice and unstinting support; Neil Clark and Mike Coates for responding to my enquiries about taxonomy and references; Glyn Sattersley for photographs of the excavation; and, most of all, to Stan himself, whose vision, determination and persistence inspired so many people over so many years.