The area known as Lethen Bar and Clune, SE of Nairn, Scotland, is a classic Middle Devonian locality, which has yielded nodules or concretions, some of which contain fossil fishes with the highest quality of preservation. The locality was largely centred on farm limestone quarries situated around the perimeter of an isolated outlier of the main fishbed. It was first described in the 19th century, although the upsurge in collecting fossil fishes only occurred some 12 years after the quarries were first mentioned in the scientific literature. Our knowledge of the provenance of the locality is based on very limited accounts, which have never been scientifically tested; these accounts also contain apparent anomalies that have never been adequately addressed. Based on these anomalies, the author of a paper published in 1983 proposed that the locality had been untraceable since the late 19th century and that the outcrop was quarried out. In 2005, the present authors recorded the first scientifically detailed stratigraphical section of the fishbed, followed in 2021–23 by detailed field surveys and by a reappraisal of the literature. This work has resolved the 19th century anomalies and allowed us to confirm the locations of old quarries, to give affirmation to two previously unrecognized sites and to show that the outcrop is still present.

Thematic collection: This article is part of the Geology of Scotland collection available at: https://www.lyellcollection.org/topic/collections/geology-of-scotland

The celebrated Middle Devonian fish-bearing nodule locality named ‘Lethen Bar’, 8.6 km SE of Nairn, has an almost apocryphal reputation in vertebrate palaeontology owing it being the source of limestone concretions that contain beautifully preserved fossil fishes of white, crimson and purple bone. Although the quarries had been worked as a source of lime for at least the preceding 10 years, it was not until 1839 that the pioneering geologist, Dr John Grant Malcolmson (1802–44), accompanied by his friend William Alexander Stables (1810–90), latterly factor of Cawdor Castle, discovered that many of the concretions or nodules contained fossil fishes. The first to be found, in the presence of Malcolmson, was by Stables (Gordon 1859, p. 20), the bizarre armour-plated fish Pterichthys (winged fish) now known as Pterichthyodes milleri, and later, the equally bizarre Coccosteus cuspidatus.

The term ‘quarry’ seems scarcely appropriate in describing the limestone workings as their depths ranged from only about 2.0 m (6 ft) to possibly 4–5 m (13–15 ft) (see below).

The locality is commonly thought to be a hill with a continuous outcrop of fishbed around its base and, according to palaeontologist Dr Sheila Mahala Andrews (1939–97) (Andrews 1983, p. 260), the location of the original working was never recorded. The quarries have thus been inferred to be unlocatable for over 140 years, enhancing the locality's legendary status. Technically, this has never been the case, as Elgin schoolmaster John Martin (1800–81) marked the original quarries on two geological locality maps, one of which was published in 1842 by his friend, Elgin town clerk, Patrick Duff (1791–1861) (Duff 1842), and Dr John Horne (1848–1928) of the British Geological Survey (BGS) located all the known sites in 1878 when his party was mapping the area. Horne oversaw the reopening of one of the original and most celebrated quarries, the ‘old Quarry of Lethen Bar’ (Horne 1923, pp. 72–74), henceforth referred to as ‘old Lethen Bar Quarry’. The position of all the old quarries is consequently recorded in the BGS archives.

Since 1880, efforts to find any remaining deposits that might potentially produce the very best-preserved fossil fishes have focused on the remote outlier of the main outcrop, which is situated at Lethen House approximately 2 km to the NW of the outlier. One of the fundamental difficulties associated with this elusive quest is that, for 20 years, Malcolmson's unpublished 1839 account of the geology and topography of the area lay forgotten in the archives of the Geological Society of London (GSL). After his death it was subdivided, resulting in the publication of two incomplete versions of his original manuscript. These two versions of the memoir contain both primary and secondary idiosyncrasies that have skewed some later workers’ understanding of the exact number and nature of the fishbed deposits.

Part of the problem is that, in the early 19th century, the name ‘Lethen Bar’ was given to two adjacent but very different landmarks. Lethen Bar Farm (now Braevail farm) sits on low-lying land on the outlier of Middle Devonian sediments that include the fossil fish-bearing nodule bed. Lethen Bar Hill, on the other hand, the summit of which is 1.38 km SE of the farmhouse, is composed of non-fossiliferous granite and metamorphic bedrock. Against this potentially confusing backdrop, Malcolmson (1859) sketched two geological cross-sections supporting his description of the stratigraphy of the area but, in doing so, immediately renamed Lethen Bar Hill as its neighbour, Cairn Bar Hill, without explanation.

This and other curiosities in his memoir and sections cursorily suggest that he made some errors, but this is not completely true. Dr Hugh Falconer (1808–65), renowned geologist and palaeontologist, assessed the memoir for the GSL when it was received in 1839 and was the first to advocate its importance (Collie and Diemer 1995, p. 68).

It is reasonable to observe that both the 1839 manuscript and the subsequent incomplete memoir (Malcolmson 1859) were one and the same draft that Malcolmson did not have the opportunity to proof-read; the controversy surrounding its disappearance for 20 years and its subsequent piecemeal publication certainly did not help. One frustrating aspect of the memoir's publication is its division into two parts, published in disparate journals in 1859, and each with contributions by others. In some instances, this added context, but it came at the sacrifice of continuity, and the practice was not uncommon at the GSL at this time (Collie and Bennett 1996, p. xxi). The memoir was later published in an almost entire but still compromised form in 1921. The following presents a generalized timeline.

  1. Malcolmson's (1839) manuscript of his memoir was lodged with the GSL that year with four recorded illustrations; however, there may have been more, as, in a letter to Duff dated 2 March 1840 in Elgin Museum's archive (ELGNM MS. G9-11), Malcolmson wrote that he was preparing fossil descriptions for engraving to illustrate his paper. In the GSL editor's notes (Malcolmson 1859, p. 336) a topographic map is listed as one of only four illustrations lodged with the 1839 manuscript; this was not published with any version of the memoir and is referenced only as ‘the map’ (Malcolmson 1921, p. 452). Unfortunately, this map is no longer available in the GSL archives. Furthermore, Malcolmson's (1859, p. 342) memoir refers to three drawings ‘ Figures 1 and 2 and drawing 33’ to which an editorial footnote refers as ‘do not accompany the MS’ and an Appendix referred to by Malcolmson was ‘never prepared’. The remaining three illustrations comprised an anonymous watercolour of the fossil fish Osteolepis macrolepidotus, from an unknown locality, a vague sketch of a fossil plant, and a lithographic plate of 11 original geological section sketches. The original geological sections and lithographic plate are now lost but facsimiles were eventually published (Malcolmson 1859); the watercolour (LDGSL 113_Fish) and sketch (LDGSL 113_Plant) survive in the GSL archives. Publication was initially deferred, awaiting a contribution on the fossils by the celebrated Swiss palaeontologist Jean Louis Rodolphe Agassiz (1807–73); some time after Malcolmson's death in 1844, interest appears to have lapsed, or even to have been ‘suppressed’ (Keillar 1993, p. 12). In 1859, the efforts of the Reverend George Gordon (1801–93) and the eminent geologist Sir Roderick Impey Murchison (1792–1871) finally resulted in the publication of the memoir, but, by this time, the illustrations appear to have become a casualty of the delay. Significantly, Agassiz’ contributions were not included, having already been published in his Monographie (Agassiz 1844-45).

  2. In 1839, at least one facsimile copy of the manuscript was in circulation. According to Andrews (1982, pp. 71–73), Malcolmson's mother was to have a copy for safekeeping: Andrews mentioned (Andrews 1983, p. 251) the manuscript being sent to Moray by Malcolmson. It is possible that Stables acquired a copy of this version or another, which Gordon subsequently used for his paper (Gordon 1859, p. 22). In 1915, a copy was submitted to the Inverness Scientific Society, the donor having been given Gordon's permission some years earlier (Wallace 1921, p. 422). The origin of the 1839 manuscript copy now held by the National Museum of Scotland (NMS), which Andrews generally cited, is unclear but it may be the Stables or Gordon copy, having passed through various hands (Miller 2023, p. A206).

  3. In 1842, the abstract of Malcolmson's memoir was published by the GSL (Malcolmson 1842), without illustrations.

  4. Gordon's memoir was published (Gordon 1859) with the stated purpose of bringing to light the work of Malcolmson (1839); this paraphrases and quotes large, but edited, parts of Malcolmson's (1839) manuscript including almost all of the information on Lethen Bar and Clune, without illustrations and with, unhelpfully, no direct reference to the geological sections. Andrews (1982, 1983) was critical of Gordon's treatment of Malcolmson's manuscript copy, although she did cite Gordon's contributions to the paper.

  5. Later in 1859, Malcolmson's belatedly published and depleted memoir appeared: this, to its detriment, omitted all of the content published in (3) and (4) above but added editorial notes and reproductions of the 11 geological sections originally lodged with the GSL (Malcolmson 1859, pl. XI); this is the only version where the geological sections are published and Andrews referred to this document, but not specifically the sections, in her work (Andrews 1982, 1983).

  6. In 1921, Malcolmson's almost complete, but apparently not proof-read, memoir was republished by the Inverness Scientific Society under the incorrect author's name of A. G. Malcolmson (Wallace 1921). Unfortunately, this version does not include the geological sections, although the text makes references to them. It also contains errors, which caused Andrews (1982, 1983) to doubt the provenance of its content, although she did cite this version (Andrews 1982, p. 73). Regardless of Andrews’ reservations, Wallace (1921) confirmed that a manuscript copy in his possession was submitted to the Inverness Scientific Society in 1915, with Gordon's prior permission, and published, with some editing, in their 1921 Transactions.

The authors have referred to all the above documentation to provide contextual information and valuable cross-referencing for the present study; it should be noted that there remain significant and unqualified mismatches between the various versions, resulting in none of the publications being as robust or complete as Malcolmson (1859, 1921; Gordon 1859) had intended.

Nineteenth century accounts

To put early accounts in context, geological mapping was at an early stage at this time and, even today, relies on inferred or extrapolated data where, for example, outcrops are not exposed or where borehole or excavation evidence does not exist. However, since the 19th century, our ability to resolve field information has improved greatly but there remains no substitute for visual examination of outcrops. In the case of the Lethen fishbed, although direct excavation evidence for the sedimentary rocks was available from the quarries in 1839, robust igneous and metamorphic bedrock data were not. Where topographical maps were in use, they were often inaccurate, although, to some extent, this information could be assessed in the field.

The best preservation of fishes is reputedly confined to specimens collected between the 1830s and the 1880s from only two named localities, old Lethen Bar Quarry and Clune respectively. It appears that Clune was still productive in 1871 as Gordon wrote to Dr John Grigor (1814–86), the benefactor known as The Father of Nairn, that fishes could still be picked up ‘at Clune on the Lethen property’ (Collie and Bennett 1996, p. 108).

Mr Jex, a commercial collector working alongside the BGS party, appears to have acquired abundant well-preserved material from the reopening of the ‘old Quarry of Lethen Bar’ in 1877 or 1878 (Horne 1878). Jex on this occasion was working for R. F. Damon & Co., a natural history dealership based in Weymouth. Jex was highly regarded, having a number of species named in his honour (e.g. Protodus jexi Woodward, 1892) and was specifically recognized by authors (e.g. Traquair 1890). Unfortunately, no other information on Jex has come to light during the present study.

R. F. Damon & Co. was a father and son enterprise established by Robert Damon FGS (1814–89), the father, and Robert Ferris Damon (1845–1929). The company distributed many fine specimens from Lethen Bar and other localities to international institutions and private collectors, and these still emerge today, including Lethen Bar material, easily distinguishable by their ‘From R. F. Damon. Weymouth. England.’ labels.

It is important to remember that Malcolmson and Horne were not contemporaries; their fieldwork at the site was some 40 years apart and Horne had the benefit of Malcolmson's published stratigraphical work (Malcolmson 1859), which he apparently respected greatly. Horne also had the first, accurate Ordnance Survey (OS) maps of the area to which geological detail was added, known by the BGS as field slips (Horne 1878), where the base map was cut into sections to fit a surveyor's map case for ease of handling in the field. Although Horne was working with a party of BGS surveyors he was given sole credit for the field slips, evidenced by the title handwritten on the rear of the documents: ‘Area report as surveyed by John Horne year ending 30th September 1878’.

Recent accounts

The most recent report (Dineley 1999, p. 197) described Lethen Bar as ‘no longer exposed’, and Andrews described in two separate publications (Andrews 1982, p. 71; 1983, p. 259) the locality thus: ‘Lethen Bar, Nairnshire … is a hill … and the fishbed formed a continuous outcrop around it’. Andrews (1983, p. 260) proposed that the Lethen Bar locality has been effectively lost owing to being ‘worked systematically anticlockwise’, concluding that ‘Owing to the method of working there is probably very little, if any, undisturbed fish bed around the outcrop on Lethen Bar’ (Andrews 1983, p. 262). Both these accounts are misleading as, whereas of the most celebrated quarries, old Lethen Bar Quarry, was finally backfilled in the late 1800s, and Clune probably by the turn of the century, the outcrop of the main exposure, at Lethen House Quarry, and the adjacent outcrop were not. Andrews’ (1983) conclusions now appear particularly implausible as they were based purely on sparse and obtusely worded passages in the manuscript of Malcolmson (1839) and the resultant memoirs (Malcolmson 1859, 1921; Gordon 1859).

Here we present a detailed reconciliation of the historical literature with the geology and topography of the area, supported by our own fieldwork, which sheds new light on the locations and working methods of the quarries, and allows a reappraisal of past workers' reactions to the progressive work of Malcolmson (1839, 1859, 1921).

Of the six main fish-bearing nodule beds around the Moray Firth coast, Clava, Lethen, Tynet Burn, Cromarty, Edderton and Gamrie (Fig. 1), only Tynet Burn has been fully documented, by Trewin and Davidson (1999). That account provided the depositional setting and palaeoenvironmental information typical of these nodule-bearing localities, situating them in a nearshore setting periodically transgressed by the rise and fall of water levels in the Middle Devonian Orcadian Lake (about 388 Ma). The terms ‘nodule’ and ‘concretion’ are used interchangeably today but the most correct term is concretion, being, in this case, a calcium carbonate mass that in key cases has formed around a nucleus (Parnell 1983; Marshall and Pirrie 2013), which may be animal or plant remains.

The Lethen fishbed fauna is typical of the Achanarras horizon (Trewin and Davidson 1999) and up to 18 genera of jawed and jawless fishes and one arthropod are now known across all contemporaneous outcrops, with some variations, although Achanarras Quarry is the only locality to yield the entire suite of genera to date.

Apart from Clava and Dipple, the accessible nodule bed localities at Gamrie, Edderton and Cromarty were also logged by Trewin and Davidson between 1996 and 2005, and comparisons were made to the Lethen fishbed outcrop and past descriptions. This new analysis brings additional clarity to the several localities that existed in the Lethen Bar (Nairnshire) and Moray areas and their relationship to each other, and further investigation may yield additional information.

The date of discovery of fossil fishes at Lethen Bar was given by Gordon as 27 March 1839 (Gordon 1859); however, in 1880 he wrote a letter to Horne (Collie and Bennett 1996, p. 147) stating that ‘It was on 27th March 1839 that Malcolmson and Stables first discovered a Pterichthys at Lethen Bar (Clune)’. It is therefore unclear as to which locality was the first to be recognized for its fossils, but the overall evidence suggests that it was Clune (Andrews 1983, p. 262).

Lady Eliza Maria Gordon Cumming

According to Murchison (1859, p. 421), it was at ‘Clune’ and ‘Lethen Bar’ that botanist, artist and aristocrat, Lady Eliza Maria Gordon Cumming (1795–1842) of Altyre House, obtained her very fine collection of Devonian fishes, which Agassiz (1844-45) drew upon for type and figured material (see Andrews 1982, pp. 24–26 for a full account). Lady Gordon Cumming and her daughter were adept water colourists and produced several fine drawings of the fossils for Agassiz, which today are lodged in various institutions (Andrews 1982), including an album of their watercolours now in the NMS (Miller 2023, fig. 61). Despite Lady Gordon Cumming's significant contribution to science, first acknowledged by the GSL (Lyell et al. 1842) and thereafter Agassiz in his celebrated Monographie (Agassiz 1844–45), it is only recently that her true importance has been championed (Trythall 2012; Berla and Turner 2020; Orr 2020).

Lady Gordon Cumming's collection became known as the Altyre Collection, named after her home, Altyre House, where it was originally housed, and it now resides in the NMS after an extended loan (102 years) to the Falconer Museum in Forres (Andrews 1982, p. 73). This, along with Agassiz’ specific recognition of the single locality name, ‘Lethenbar’ (Agassiz 1844–45, p. 14), and his assertion that the quarry was worked solely for its fishes (Andrews 1982, 1983), led to the locality acquiring classic status, and probably to the suppression of the significance of the main outcrop at Lethen House and the other quarries, which, apart from Clune, produced what was regarded as poorer quality material apparently worthy only of processing for lime.

Apart from Agassiz’ eventual publication of the fossil fishes, other record keeping was poor (Andrews 1983); for example, specific details of the quarries from which the fishes were procured have not survived and despite Horne's outline description of two of the old quarries (Horne 1878, 1923, p. 73) the stratigraphy of the outcrops was not recorded in scientific detail, although this was not uncommon in the Victorian era.

Hugh Miller, in The Cruise of the Betsey (Miller 2022, pp. 190–193), recounted a second visit to Clune, where he saw specimens of fishes that had great beauty, but that after Lady Gordon Cumming's premature death in 1842, the quarrymen were no longer setting them aside for collectors. Despite the insistence by Agassiz (1844–45, p. 14) that ‘Lethenbar’ was the sole source of material in the Altyre Collection, it is apparent from the evidence that both Clune and Lethen Bar yielded similar high-quality levels of preservation, perhaps patchily (e.g. specimens in the Sedgwick Museum, Cambridge; H 4687, H 4689, from Clune, and H 4488, from Lethen), and that Clune material is present in the Altyre Collection.

The original Altyre Collection contained a range of preservation styles from the very finest to material that was apparently less appreciated and was relegated to outside storage (Gordon Cumming 1904, p. 39: ‘The poorer specimens were deposited in rows under the veranda, and there remained as familiar objects of our early days’). The strong historical bias towards favouring the finest material (Andrews 1983, pp. 246–247) (Fig. 2) means that the Altyre Collection now in the NMS is probably not representative of the range of preservation modes, having possibly shed the less treasured material (see below). Lady Gordon Cumming's generosity to other collectors, including the GSL (Lyell et al. 1842), will mean that museum collections elsewhere will exhibit some level of bias.

The Cawdor Collection, Nairn Museum

In collections made by others, collecting bias might not be so stark; for example, Lethen Bar material in the Nairn Museum, some of which was collected by Grigor and Malcolmson's close friend, Stables, as witnessed by one unregistered specimen marked ‘WAS’, Stables’ initials.

In December 2021, at the invitation of Nairn Museum, their assemblage of fossil fishes was examined by R.G.D. The Cawdor Collection of approximately 100 fishes, collected by Stables (Andrews 1982, p. 71), was under review, along with many other unregistered Lethen Bar specimens that were available for viewing. The range of preservation quality is wider than that of the Altyre Collection, of which Andrews (1983) ranked 65% of the total as ‘very good’ and 35% as ‘superb’ (see below). The Nairn Collection can be ranked, albeit subjectively, as fair to superb with all intermediate grades probably present, thereby rendering it potentially more representative than other museum collections. Caution is advised in grading preservation based on aesthetics, however, as some of the fishes are preserved in three dimensions and thereby retain the potential for good preservation of bone structure.

The Nairn specimens may help to resolve the puzzle surrounding ‘lost’ Lethen Bar and Clune specimens. Andrews (1982, p. 71) had written that Stables’ collection had become divided and that the best specimens were retained by the museum as the Cawdor Collection in 1844, and that the whereabouts of the remainder was unknown. Subsequently, she visited Nairn Museum in the 1990s and examined both registered and unregistered material, labelling unidentified and unlocalized material with a red marker on their newspaper wrappings. On our appraisal, which was undertaken to assist preparation for curation (Davidson 2021), a considerable amount of Lethen and Clune material remained unlocalized and this task was completed, albeit with the caveat that it is not possible today to segregate the two localities from specimens alone. The historically large number of unregistered specimens allows for speculation that some or indeed all of it may be the missing Stables material; this merits further investigation and is beyond the scope of this paper.

Recently, Lethen Bar and possibly Clune nodules, discovered in museum searches, have been shown to support exceptional preservation (Davidson and Trewin 2005). Several specimens of acanthodian and actinopterygian fishes from Lethen Bar exhibit consistent iron oxide traces deposited at hemoglobin-rich internal organ sites by chemotrophic bacteria (Trewin and Knoll 1999). One such Lethen Bar specimen (NMS G 1891.39.36) in the NMS, exhibits kidney, and putative heart and spleen traces, and another, with a label affixed stating ?Clune (NMS G.1966.40.78), exhibits a liver or gills trace.

Over recent years, several workers have set out to rediscover the fishbed of the Lethen area, either without success or without publishing any findings, with one notable exception.

Dr Sheila Mahala Andrews

Andrews, latterly known as Mahala, obtained her PhD on fossil lobe-finned fishes at the University of Cambridge and was appointed a senior scientific officer at the Royal Scottish Museum, now NMS, and continued her work on lobefins (e.g. Andrews and Westoll 1970), particularly the group from which land vertebrates would evolve. In 1989 she appeared with David Attenborough in the BBC series Lost Worlds Vanished Lives, in the episode Magic in the Rocks.

Andrews (Fig. 3) produced unparalleled work on the discovery of fossil fishes in Scotland (Andrews 1982). Importantly, Andrews drew on the collective works of Malcolmson (1839, 1859, 1921), Murchison (1859), Gordon (1859) and Horne (1923), in her successful quest to geologically segregate the Middle Devonian Lethen Bar locality from Murchison's Upper Devonian locality on the Altyre Burn (Andrews 1983, pp. 243–259).

Her attempts in the same publication to confirm the past locations and mode of exposure of the workings of the Lethen Bar quarries were less successful. This appears to be partly due to the difficult to decipher and piecemeal nature of the evidence available to her, and to an apparent minimal level of fieldwork and available resources. She did, however, forego important lines of evidence by dismissing or setting aside some findings by previous workers (see below).

More importantly, in the course of her work, Andrews (1982, 1983) performed detailed and extensive research in unpublished correspondence, regional archives and manuscripts, yielding invaluable primary information.

Andrews’ map

In her search for the Lethen Bar locality, Andrews produced a map (Andrews 1983, fig. 4), which is reproduced with modifications in Figure 4 of this paper. She stated that her map was ‘based on published and reference maps of the Institute of Geological Sciences’, now BGS, and ‘from the Geological Survey of Scotland 6 inch:1 mile sheets 4 and 5 Nairn’ (Andrews 1983, fig. 4 caption). Her map faithfully reproduces the outline of the fishbed outlier demarcated by Horne on his field slips and map (Horne 1878; BGS 1923). The fishbed outlier is referred to henceforth as the BGS Outcrop.

At no point did Andrews (1983) specifically refer to Horne's (1878) field slips for the Lethen Bar area, which are stamped Nairn 5, SE, NE and SW, but it appears that she may have examined them, as elsewhere in the paper (Andrews 1983, p. 259 and fig. 3), while discussing the easterly Altyre Locality, she mentioned ‘original MS field maps’. She quoted (Andrews 1983, fig. 4 caption) the notes on quarry positions exactly as they are written on Horne's (1878) field slips, with the addition of her hypothetical quarry. It is interesting to note that the observations she quoted for Locality 3 (Andrews 1983, fig. 4 caption) are incomplete as to the thickness and depth of the working, which was to have implications for her analysis of the depths of the quarries (see below).

Reference to Figure 4 indicates that the BGS Outcrop was understood to be a partly continuous but uncertain sub-elliptical structure situated approximately 2.0 km SE of Lethen House. The main outcrop, represented by Lethen House Quarry (Quarry 6, Fig. 4), was, and still is, traceable over 400 m and remains largely undisturbed, possibly owing to an impression that the quality of fish preservation is poor in this part of the outcrop. However, this notion has never been scientifically tested and past accounts stated that even at the best quarries on the BGS Outcrop (Quarries 1 and 2 in Fig. 4) preservation was variable (Andrews 1983).

Stan Wood

According to a sketch map and stratigraphic section sketches in the BGS archives, the renowned Scottish fossil collector, researcher and dealer, Stanley Purdie Wood (1939–2012), known as Stan, excavated four test pits (Wood 2001) adjacent to Malcolmson's (1859) line of section (Fig. 4), but the fishbed was absent in all four pits. It appears likely that he targeted Andrews’ (1983) hypothetical Quarry 7 (Fig. 4). Wood appears to have been misled by Horne's (BGS 1923) tentative (see below) interpretation of the shape of the outcrop (Fig. 4), which is now known to be narrower in the SW of the outlier, and thus he missed the outcrop by digging two of his pits too far west.

However, the true value of Wood's work is that it contributed significantly to the BGS’ new interpretation (see below). It appears that Wood excavated the fishbed more successfully elsewhere, although it appears that the exact position of those pits was not recorded in BGS archives; but the overall fish preservation was apparently deemed poor and generated no further interest.

Several reports contain areas of potential confusion as local place naming evolved; it is therefore useful to understand the derivations of place names in the past, compared with those in use today. For example, Lethen has previously been known as ‘Lethan’, ‘Lethem’ and ‘Lethin’, and even today the misspelling ‘Letham’ occasionally appears. Historically, other variations have been documented; for example, ‘Lethenbar’ (Agassiz 1844–45, p. 14), ‘Lethenbarn’ (Miller 2022, p. 190) and ‘Clewan’ (Sedgwick and Murchison 1829, p. 151).

The name ‘Lethen Bar’ has endured, largely in palaeontological circles, owing to its celebrated status and over the years evolved to embrace the whole area on the BGS Outcrop comprising Quarries 1, 2, 3 and 4 (Fig. 4), which Andrews further extended to include Quarry 6, Lethen House Quarry to the west. On the OS Nairn and Cawdor Sheet 422 (OS 2015; Fig. 5) the area of the BGS Outcrop stretches from the vicinity of Chapel Hill and Easter Clune, SW to a point about 1 km past Braevail Farm, near the ford in Figure 5; its east–west extent is roughly demarcated by Clune Croft in the east and by Wester Clune and Garrowstripe to the west.

Early reports of Lethen and Clune

The very first report of the nodule localities was a statement by Sedgwick and Murchison (1829, p. 151) giving ‘Letham or Clewan’ as a quarry that they surveyed in 1827, which was excavated for concretionary limestone. However, it is not clear from their account whether they were describing the two localities or a single quarry in the Lethen area (i.e. Clune), as 12 years later, in 1839, Malcolmson inferred that old Lethen Bar Quarry was ‘an excavation recently made’ (Andrews 1983, p. 262). John Martin marked it as a quarry on his 1842 geological sketch map of Moray (Duff 1842, pl. 1). Today, much of the area is known locally as Lethen and the land depicted in Figures 4 and 5 forms part of the Lethen Estate, which might offer an explanation as to what Sedgwick and Murchison (1829) meant.

Clune

The locality name ‘Clune’ has historically been applied to an area containing a quarry and two or three croft houses (now demolished) on the east margin of the BGS Outcrop (Figs 4 and 5), which has been absorbed by Easter Clune Farm. There are two important references to Clune: Horne (1923, pp. 72–74) referred to it to identify the site of an old working, stating the locality (Quarry 2 in Fig. 4) is ‘300 yards NE of Clune’; the second reference is on the 1842 geological sketch map by John Martin (Duff 1842). Both authors are referring to the same quarry and croft. The OS Nairnshire Sheet V (OS 1871) shows three groups of buildings marked Clune, known locally as Clune Croft, adjacent to a field corner, slightly to the north of where the same locality is marked on the modern map (OS 2015; Fig. 5). Because this field boundary is unchanged to the present day, applying these data allows Clune Croft to be relied upon as a reference point; we have thus shown this location in both Figures 4 and 5. Wester Clune is not referenced in any literature, despite it being situated immediately west of the line of the outcrop.

The old quarry of Lethen Bar

In the 19th century, the name ‘Lethen Bar’ was, with a few exceptions, generally confined to a single quarry, the ‘old Quarry of Lethen Bar’ (Horne 1878) (Quarry 1, Fig. 4). The origin of the name comes from Lethen Bar Farm, as it was then known, probably because of its geographical association with Lethen Bar hill to the south. The farm is the site of the original Lethen Bar quarry (old Lethen Bar) but, in the mid to late 1800s, the name was changed to Brevail Farm, a transitional spelling of the modern name of Braevail.

Malcolmson's (1859) controversial renaming of Lethen Bar Hill to Cairn Bar Hill is less straightforward; he was the only early author to persist with the Cairn Bar name, the reasons for which are explored below.

Lethen House Quarry is approximately 300 m SW of Lethen House and comprises two adjacent exposures of the main outcrop of the Lethen fishbed on the NW bank of the Muckle Burn; and is prominently marked on the OS 25 inch:1 mile map (OS 1870b). This map presents the outcrop measuring some 75 m, but it is not clear today how much of this extent is a natural outcrop, and what proportion is anthropogenic. This site concurs with Andrews’ (1983) locality (Quarry 6, Fig. 4), which to this day yields fish in nodules, although these lack the vibrant preservation in those from old Lethen Bar Quarry and Clune. The reason for opening this quarry is not clear; it is the only one of the quarries that is marked on maps (ELGNM 1853; OS 1870b). There is no lime kiln marked on old maps and none was found on site, suggesting that it was never exploited for lime although it is very rich in calcium carbonate, its fundamental component. It was open in 1839 and may have been dug by Malcolmson's field party (Malcolmson 1921, pp. 457–458) but appears to have been abandoned soon after.

This quarry was mentioned by Gordon (1859, p. 35), where he paraphrased Malcolmson's (1839) manuscript fairly heavily, and by Andrews (1983), where she placed the quarry on her map (Fig. 4). Horne's field slips for this area could not be located in a search of the BGS archives. Lethen House Quarry appears to have become overgrown at some point, although the adjacent outcrop has remained accessible for many years and indeed is relatively easy to locate. Despite this accessibility, and its status as the main outcrop, the site has attracted only minimal attention over the years (e.g. Parnell 1983). The original quarry was re-excavated by the estate in the early 2000s when trees were felled to remove non-native conifer species, and two of the present authors (J.A. and S.R.W.) ‘rediscovered’ the locality in 2005. There is about 2.6 m of strata exposed containing approximately 70 cm of nodule-bearing shale. However, in a site appraisal in November 2021, the main quarry was again partly obscured by fallen trees, but the adjacent exposure remains accessible. Lethen House Quarry and exposure have provided the most detailed stratigraphical information for the area including the BGS Outcrop (see below).

In 1878, the BGS surveyed the whole area containing the Main and BGS outcrops (Fig. 4), producing the subsequent map (BGS 1923) with its accompanying memoir (Horne 1923). It has been crucial to this study to refer to Horne's field slips (Horne 1878), as these contain additional primary information.

On Andrews’ map (Fig. 4), Quarry 1 was identified by Horne as ‘old Quarry of Lethen Bar’ (Horne 1878, 1923, pp. 72–74), Quarries 2, 3 and 4 were places where limestone nodules in clay were worked; Locality 5 is a lime kiln, the only one mentioned by Andrews (1983), which presumably served Quarry 3. During our survey a second lime kiln (Locality 8, Fig. 4), which appears to have served Quarry 4, was identified to the east in the scrubland flanking the adjacent field; both these locations, in addition to Locality 9 in the south, and Locality 10, which serviced Quarry 2, are lime kilns marked on OS map Nairn Sheet V, 1871 (OS 1871).

In direct contradiction, and despite quoting Horne's (1923) fixed quarry positions in his own words, Andrews (1982, 1983, p. 262) proposed the method of quarry working as an enterprise that ‘systematically’ followed the outcrop around a hill, and thus the term ‘Lethen Bar Quarry’ indicated only the current position of workings, and not a fixed locality. Even though previous authors named Quarry 2 (Fig. 4) as Clune Quarry, Andrews (1983), did not and, significantly, did not record any visit to the east of the outcrop; Horne referred to it as ‘Clune’ and ‘the quarry at Clune’, and as a place ‘where nodules were worked three hundred yards NE of Clune’ (Horne 1878, 1923, pp. 72–74).

Horne (1878) positioned Quarries 3, 4 and a BGS trial pit on the flanks of the only hilly feature on the BGS Outcrop, the northernmost Chapel Hill upon whose SW flank Easter Clune Farm is situated. His party did not actually inspect these workings, stating that they were largely covered up and under cultivation when he passed through, and that outline details of their stratigraphy were derived from ‘information obtained at that time’ (Horne 1923, pp. 72–74). However, Horne did publish a section through Chapel Hill (Horne 1923, p. 75, fig. 2), correctly depicting the fishbed as a trough structure whose infilling sediments created localized raised topographical relief.

Andrews (1983, p. 260) drew particular attention to what she deduced to be intensified working of the fishbed near Easter Clune Farm ‘where the outcrop broadened’ on the SW flank of Chapel Hill, highlighting the presence of a single lime kiln, but did not show any contour lines on her map to suggest that the area was particularly hilly (Fig. 4), which is obvious on site. A large stream cutting in front of Easter Clune Farm demarcates a significant broadening of the fishbed to the south and is a probable natural exposure; this was presented as a significant feature on Horne's field slips (Horne 1878, Easter Clune FS1, Lethen Bar FS1), with no annotation or further mention.

Owing to Andrews’ (1983) hypothesis of a continuously moving working, she concluded that, at the time the Altyre fossils were collected, the outcrop was probably being worked on the west or NW side (Andrews 1983, p. 262); this accounts for her theoretical Quarry 7 (Fig. 4). As a result, Andrews (1983, p. 262) concluded that ‘there is probably very little, if any, undisturbed fishbed remaining’ around the BGS Outcrop. However, her hypotheses were based on an oddly worded statement from Malcolmson's (1839) manuscript (see below).

In the mid-1900s many different symbols were employed on OS maps to denote the position of lime kilns. On OS Nairnshire sheet 5 (OS 1871) kilns are marked in general as a symbol with no text label. These symbols comprised a circle with the bottom shaded to represent the kiln stoke hole (see OS 1871 and Fig. 5 for examples). Closer examination of OS (1871) reveals that more than one symbol style was in use on this document, both a simple circle (at Garrowstripe) and the stoke-hole variation at all the other localities. There appears to be no reason for this hybrid style, but a circle was also used to denote a water well. The OS 25 inch:1 mile map (OS 1870a) consistently utilized a simple circle labelled ‘Old Limekiln’ for all the kiln localities around the BGS Outcrop, including Garrowstripe.

As the quarry workings were primarily focused on lime production, the presence of a lime kiln is good evidence that an old quarry existed nearby, even when the quarries themselves cannot be traced. Reference to Figure 5 demonstrates that up to six kilns were distributed, apparently strategically, around the area; five of these are recorded on Nairn OS Sheet V (OS 1871). Three of these kilns are situated adjacent to recorded quarry locations (2, 3 and 4 in Fig. 4); the kiln at Quarry 2 [NH958512] is well preserved and can be inspected today (livestock permitting) but that at Quarry 3 [NH956516] (Locality 5, Fig. 4) is in enclosed, difficult to access land, although the farmer confirmed it is still intact; the kiln at Quarry 4 [NH959515] could not be closely examined on our surveys owing to livestock restrictions but was viewed from a distance.

The two kilns (Fig. 5) south of Andrews’ (1983) hypothetical Quarry 7 cannot be directly associated with any quarry revealed by this study, which is intriguing. Stan Wood, on his sketch map (Wood 2001) recorded the occurrence of limestone fragments in the meadow, in the vicinity of the northernmost of the two kilns. This corresponds to the stoke-hole symbol on OS (1871) [NH944504]. Wood (2001) annotated this locality on his sketch map ‘Site of old kiln?’ implying that in 2001 that there was little trace of it left. The kiln to the south is taken from Horne's field slip (Horne 1878, Lethen Bar, FS3); the placement of two kilns in proximity seems odd and Horne's field slip, although deliberate, is vague as to the exact site of the southernmost kiln, as he did not mark it with his customary ‘x’. Reference to the otherwise trustworthy OS 25 inch:1 mile map (OS 1870c) shows that no such kiln was recorded, in which case this could be a BGS mapping error.

The presence of at least one kiln and limestone fragments in this area [NH944504] could be evidence of nearby workings or it may simply be a depository for commercially viable material from old Lethen Bar Quarry, which is known to have yielded limestone bands (see below). Although there is no evidence of a kiln in close association with old Lethen Bar Quarry, this may have been due to the unsuitability of the flat local terrain. In general, kilns were built into a bank to allow top loading of limestone and extraction of lime from the bottom, and thus location and proximity of a kiln to old Lethen Bar Quarry may have been dependent on local topography, as demonstrated by the kiln adjacent to Quarry 2.

The previously unnoticed kiln at Garrowstripe [NH946509] presents an enigma owing to its aberrant marking and its potential for evidence of an unknown quarry. It appears on Horne's base map, OS Nairn sheet 5 (OS 1871) as a simple circle (not the stoke-hole variant) with no label, and therefore is easy to overlook or mistake for a well. It is labelled ‘Old Limekiln’ on the 25 inch:1 mile map of Nairn (OS 1870a) and marked with a partial circle. It was situated immediately east of the line of the outcrop, at [NH946509], approximately 300 m north of Andrews’ (1983) hypothetical Quarry 7 in an area where no quarry or old lime kiln was reported by any fieldworker or author, and no quarry appears on the 25 inch:1 mile map (OS 1870a). Garrowstripe was originally a small farm and by 1869 had become absorbed into neighbouring Braevail Farm according to parish records (Anon. 1869).

Andrews (1983) only cursorily considered kiln evidence, doubtfully implying that the single kiln that she mentioned near Quarry 3 (Fig. 4, Locality 5) signified more intense working in this area, while attaching no significance to the remainder. Ironically, in setting aside kiln evidence, an opportunity to reinforce her argument for her hypothetical west flank quarry was foregone (Fig. 5).

The fishbed traverses two farms, both on gently sloping meadowland with more undulating terrain to the NNE. There are two important hills SE of the fishbed that are of interest to the present study (OS 2015). In the past, as today, the name ‘Lethen Bar’ specifically applies to a discrete hill in scrubland, 1.38 km from Braevail farmhouse; this is classified as a Hump, Tump (200–299 m) on the Database of British and Irish Hills and has a trigonometry point and robbed cairn at its summit [NH953493], 83 m above the elevation of the farmhouse. The south part of Lethen Bar Hill marks the northern extent of the Ardclach Granite Pluton, and its northern slope lies in the Neoproterozoic metasediments of the Badenoch Group (BGS 2012). It has a prominence from the eastern valley floor of 31 m and is directly intersected by Malcolmson's (1859) section. Lethen Bar is closest to the fishbed outcrop and to the farm that bore its name in the 19th century.

A second, unclassified, hill of only 20 m prominence but of far greater width exists 1.2 km across the valley [NH970492] to the east of Lethen Bar Hill. It is entirely centred on the Badenoch Group of the metasedimentary bedrock (BGS 2012) (Fig. 6); on the OS map (OS 2015), it is identified as both Cairn Bar and Shaw Hill, which is also a feature of some older OS maps. Interestingly, Cairn Bar does not have a cairn at its summit whereas Lethen Bar does.

To explore the question of whether the BGS outcrop is a hill or moorland, it is useful to divide the BGS Outcrop into two areas, comprising two adjacent farms (Figs 4 and 5). The SW sector is situated on Braevail Farm, containing Quarries 1 and 7 (Fig. 4) and at least two lime kilns. The NE sector is centred on Easter Clune Farm and includes Quarries 2, 3 and 4, and a large stream cutting adjacent to the farm, plus three lime kilns in the surrounding scrubland (Figs 4 and 5).

Braevail Farm, SW sector

The OS map for the area (OS 2015) (Fig. 5) shows meandering contour lines on a gentle upward gradient to the SE that gradually continues to Lethen Bar Hill, 1.0 km to the SE of the fishbed (BGS 1923). Horne, in the BGS memoir of the area (Horne 1923, pp. 72–74), did not use the term ‘hill’ to describe any part of the SW sector of the outcrop, referring instead to ‘Lethen Bar moor’ and to ‘Lethen Bar Farm’.

Our October and November 2021 site surveys and the OS data confirm that the Braevail area containing old Lethen Bar Quarry is a gentle sloping moor or meadow, covering a wide area, and that no hill exists on Braevail Farm as drawn by Malcolmson (1859).

Easter Clune Farm, NE sector

Site surveys in 2021 confirmed that Easter Clune Farm straddles both the same gently sloping terrain as Braevail Farm and Chapel Hill, which forms raised ground to the extreme NE. Easter Clune Farm and Quarries 3 and 4 (Fig. 4) are partly incorporated in this area of modest undulating relief, with Quarry 2 (Clune Quarry) further to the SE amongst the sloping meadows.

The BGS Outcrop is therefore mostly gently sloping meadowland, with no southwesterly hill, its only raised relief being in the extreme NE, 1.46 km from old Lethen Bar Quarry.

Between 1987 and 2010, the BGS re-surveyed the area and, in 2012, produced an updated solid geology map and new section of the area (BGS 2012), partly to address previous inconsistences, an extract of which is given in Figure 6a and b; new evidence identified by Wood (2001) also greatly enhanced understanding of the depositional structure of the area. The BGS Outcrop is interpreted as being deposited over three fault blocks and exhibits more discontinuities than recorded by Horne (BGS 1923) in the NW and south. The BGS recognizes five historical and extant fossil localities (BGS 2012, Localities 6, 7, 8, 9 and 10) but does not name the quarries; interestingly, Andrews’ (1983) Quarry 4 (Fig. 4) is not included, nor is the trial pit to the east where nodules were found. This seems to be an oversight as they were pinpointed on Horne's field slip (Horne 1878, Easter Clune FS1).

The fishbed along Malcolmson's (1859) section line is narrower in this area than Horne (BGS 1923) depicted (Fig. 4); the line also contains a tilted or uplifted fault block that places gneissose rocks very near the surface (Fig. 6a). This resulted in little of the fishbed being preserved in this area; this revelation was, in part, gleaned from Wood's excavations (Wood 2001), whose approximate positions are marked as purple asterisks in Figure 6a.

Wood's two pits to the NW were curtailed after the fishbed was found to be absent, although it appears that he was digging too far west (compare Fig. 4 with Fig. 6a); the two southern pits encountered gneiss basement (Wood 2001). This disruption, the sparse nature of the outcrop and lack of topographical relief of the nodule bed are in direct contrast to how Malcolmson (1859) confusingly depicted the fishbed. The BGS provided a WNW–ESE cross-section (Fig. 6a, section 3) across the farm boundaries in BGS (2012), an extract of which appears in Figure 6b, which confirms a gently sloping terrain with an average up to 5° dip over this 10.5 km line of section. The BGS section (Fig. 6b) from Lethen House Quarry to the ESE confirms that the BGS Outcrop is a localized trough structure in a gentle slope, rather than a hill, resting on conglomerate of varying thickness and containing sandstones that have been eroded to the level of the surrounding conglomerate. It should be noted that the BGS has used isometric scaling on this section, which provides a faithful representation of the topography (see discussion below).

In the 1839 manuscript for Malcolmson's (1859) memoir of the geology of Morayshire, while discussing Lethen House Quarry, he described the Lethen House fishes as ‘a few poorly preserved fishes of the same fauna as Clune’ (Andrews 1983, p. 262), thereby indicating significant lateral variation in quality of preservation and providing further evidence that Clune predates the old Lethen Bar site.

In his part published memoir, Malcolmson (1859, pl. XI, figs 4 and 5 and p. 343) gave a generalized sketch of a section of the area between Lethen House and Coulmony, via Lethen Bar Hill, which he termed Cairn Bar. However, this sketch has no scale and taken at face value can be highly misleading. The memoir to the 1923 map (BGS 1923) provides a far more accurate account of the geology of the area (Horne 1923). Perhaps crucially, Horne (1878) pinpointed only four static quarries on Lethen Bar moor where nodules were collected and a test pit in the extreme NE corner where nodules in clay were seen.

Malcolmson (1859) and Horne (1923) both identified that the BGS Outcrop and Lethen House Quarries were exposures of the same fishbed; Horne referred to them as the Lethen fish-band, stating that Lethen House was the main outcrop as Malcolmson had previously determined. Despite Horne's (1923) specific and assertive account of four static quarries, gleaned from field study, Andrews (1983) adhered to Malcolmson's earlier manuscript (Malcolmson 1839) from which she developed her hypothesis of a continuously moving method of quarry working.

Reference to 19th century topographical and geological maps in the National Library of Scotland (NLS) shows that it would have been difficult in the mid-19th century to correctly match the terrain to the geological data available. The topography of the area was drawn inconsistently by different surveyors (Collie and Bennett 1996, p. xii) with the most common shortcomings being misplaced, absent or exaggerated hills. A geologist performing field survey work could therefore be confused or misled when obliged to utilize these maps as templates, regardless of the good reputation of the land surveyor.

Aaron Arrowsmith (1750–1823) and John MacCulloch (1773–1835)

On the Treasury commissioned Geological Map of Scotland by John MacCulloch (1836), apparently highly regarded by Malcolmson (1859, p. 338), the extent of the Ardclach Pluton was distorted and placed too far north. MacCulloch was using Aaron Arrowsmith's accepted topographic map as a base (Arrowsmith 1807). Arrowsmith's map contains erroneous topographic detail in the Lethen area, placing a significant hill due east of Lethen House at Braevail, where no such hill exists. MacCulloch was obliged by the Treasury to use the map of Arrowsmith (1807) and complained regularly of its inaccuracy (Bate 2010, p. 153). Because he was using Arrowsmith's map as a template, the geological map of MacCulloch (1836) features the same erroneous hill, composed of granite and gneiss, centred on the area of Braevail. Evidently, MacCulloch recognized the geology and topography of Lethen Bar Hill but placed it too far north in accordance with Arrowsmith's accepted but erroneous cartography (compare Arrowsmith 1807 with MacCulloch 1836).

The first geological sketch map of Morayshire was drawn by John Martin in 1836 for the Highland and Agricultural Society of Scotland and published with his Gold Medal winning essay ‘On the Geology of Morayshire’ (Martin 1837). A second edition of this map by Martin, which now included Nairnshire, was published by Patrick Duff 5 years later (Duff 1842, pl. 1); all the quarries then known were marked by Martin, with varying degrees of accuracy, by a circle and dot symbol and the map depicts a scale of two tenths of one inch to 1 mile; he also used a dash symbol to mark other key places. Subsequently, the Elgin and Morayshire Literary and Scientific Society commissioned a third edition (ELGNM 1853). Unfortunately, at the time of writing, this map is heavily faded and earmarked for conservation. This map depicts old Lethen Bar Quarry, Clune and Lethen House Quarry, again in the correct geographical context on a more accurate scale than edition 2 (Duff 1842, pl. 1) and includes faunal listings.

Duff (1842, pp. v–vi and 59) held Martin's 1842 map in high regard, writing that it is a ‘very correct geological map of the county’ and ‘more accurate as to localities than any yet published’, indicating that it was the best map available at the time and thus could also be used to trace localities.

Martin had been apprenticed as a land surveyor in Morayshire and would have been well placed to undertake geological mapping as his local knowledge would have been central to the relative accuracy of the data (Fig. 7).

Be that as it may, Andrews (1982, p. 65), having examined the third edition (ELGNM 1853), in Elgin Museum, dismissed it, and thereby effectively the 1842 edition, out of hand, as ‘very inaccurate’, perhaps not recognizing that maps of the period tended to be (Collie and Bennett 1996, p. xii). Andrews (1982, p. 65) referred to it as ‘a map made by him in 1859’ but it is marked in Martin's hand ‘1853’.

Andrews (1982) did not acknowledge the second edition map in Duff's book Sketch of the Geology of Moray (Duff 1842, pl. 1) although she dedicated a paragraph to discussing the other plates (Andrews 1982, p. 31). Sadly, Andrews’ (1982) dismissal of Martin's work resulted in her foregoing important clues, only citing unrelated text references from Duff's (1842) book (Andrews 1983, p. 260).

However, there is no doubt that, by modern standards, there are discrepancies in the topography and in the spatial and distance scales used by Martin (Duff 1842, pl. 1). Significantly, the igneous feature at Lethen Bar is absent in all three editions, and there is no hilly topography immediately south of the Muckle Burn (Burn of Lethen) as shown by Martin.

Adjacent to the Muckle Burn Martin has marked a quarry, with his characteristic circle and dot symbol, as ‘Lethen Bar’ (Old Lethen Bar Quarry). This contradicts Andrews’ (1983) comment that its location was never recorded, despite Duff (1842) noting that locality information was the main strength of Martin's map. Adjacent to this quarry, Martin has placed a dash, as he did for key places, which presumably signifies Lethen Bar Farm buildings (now Braevail). A second quarry appears to the NE on the sketch map, marked with the same circle and dot symbol, as ‘Clune’, along with a dash to the SSW where Clune Croft was located. Despite the map's shortcomings, the position of Clune Quarry is representative of its directional relationship to old Lethen Bar Quarry and to Clune Croft in the same sloping terrain. A dash denoting the site of Lethen House is annotated to the NW, again with the correct relationship to both Lethen Bar Farm and Clune Croft. Despite the map's shortcomings, it appears that Martin made a deliberate attempt to represent the relative positions of these landmarks but added erroneous topography with inaccurate distance information. Martin drew both quarries on the NW slope of a single NE to SW elongated hill, although this may not be his invention as some coeval maps of the area present this feature in the same but mistaken way. Martin appears to place old Lethen Bar Quarry close to the summit of the hill (Fig. 7) whereas Clune Quarry is depicted nearer the base. His juxtaposition of a quarry on a hill draws comparisons with how Malcolmson (1859) viewed the topography (see below), raising the possibility that both Malcolmson and Martin may have been influenced by a pre-existing map, especially given that they were firm friends (Andrews 1982, p. 22).

Importantly, edition 2 of Martin's sketch map is the earliest mapped report of the existence of both old Lethen Bar and Clune quarries and is repeated in edition 3, indicating that they were fixed, unconnected and contemporaneous, again in contradiction to Andrews’ (1983) hypothesis of a continuously moving working.

Since their discovery until very recently, except for the BGS, all authors, to varying extents, have produced, endorsed or left unchallenged unexacting location details as to the nature and whereabouts of the Lethen Bar fishbed localities.

The earliest mention of the outcrop is by Sedgwick and Murchison (1829, p. 151), who inferred that ‘Letham or Clewan’ was a single locality, probably in reference to Clune. However, they incorrectly gave the location as ‘about six miles west of the Findhorn’, whereas the sites of Lethen and Clune are both approximately 3 miles due west of the River Findhorn.

The details of John Martin's pioneering but flawed maps are discussed above but of particular significance is that the second edition (Duff 1842, pl. 1) appears to be one possible source of the incorrect idea that old Lethen Bar Quarry fishbed locality is at the summit of a hill.

As stated above, MacCulloch (1836) appeared obliged to accept Lethen Bar Hill as being too far north. This hill is significant, as he depicts it as being composed of the Ardclach Granite pluton to the south and of gneissose rocks to the north, which agrees with the BGS mapping of Lethen Bar Hill (BGS 1923; Fig. 6a). Situating the hill too far north would place it in abutment to both old Lethen Bar and Clune Quarries, creating implications for Malcolmson's (1859) later sections and mapping.

Malcolmson's (1859, pl. XI) memoir, including the sketches of his geological sections, could be considered naive by both today's standards and when compared with the very best contemporaneous work (Palmer 2020) in his representation of both topographical and distance information that he could not have seen in the field, especially as he does not provide explanatory scale bars. Bewilderingly, he provided three totally erroneous and different positions for Clune: the first is given in the text as being ‘on the eastern slope of Cairnbar’ (Malcolmson 1921, p. 453; Gordon 1859, p. 34), a locality that does not exist; the second is in his section (Malcolmson 1859, pl. XI, fig. 5) where he depicted Clune at the base of the NE-facing slope of Cairn Bar Hill, which it is not. Finally, again in the text, he described Clune as being ‘a mile to the eastward’ of old Lethen Bar Quarry (Malcolmson 1921, p. 454; Gordon 1859, p. 36), where the bedrock is gneiss and thus clearly non-fossiliferous. The correct location of Clune is half a mile to the NE of old Lethen Bar Quarry, situated between a well and a lime kiln, in gently sloping, west facing, meadowland (Fig. 5).

Horne (1878, 1923) produced much improved and new data on the stratigraphy and terrain in which the Lethen Bar quarries were dug, effectively superseding Malcolmson's (1859) work. By 1878, Horne was enjoying the benefits of working to BGS standards, with an accurate OS map, and was thus able to apply horizontal and vertical scale bars to his sections, all major improvements on Malcolmson's (1859) methods only 40 years earlier. Horne correctly delineated Lethen Bar Farm from Lethen Bar Hill, both on the 1923 map (BGS 1923) and in the memoir (Horne 1923), and never referred to either Cairn Bar or to a hill in the SW with a continuous outcrop around it. Horne also provided a section through Chapel Hill (Horne 1923, p. 75, fig. 2) to the extreme north and correctly drew the outcrop as a trough structure whose infilling sediments produce the low relief of Chapel Hill, the only hilly topography associated with the fishbed. In only referring to Lethen Moor, the four quarries thereon and the granite intrusion at Lethen Bar Hill, Horne (1923) effectively flagged the inconsistencies in Malcolmson's (1859) topographic naming and description. Horne did not comment on this and went on to endorse Malcolmson's (1859) general stratigraphic work (Horne 1923, pp. 18–21).

Andrews (1982, 1983) did not acknowledge Horne's (1923) work as supplanting Malcolmson's (1839) early account of Lethen Bar, which it clearly did. Instead, she focused her analysis on Malcolmson's (1839, 1859, 1921; Gordon 1859) work and correspondence, and, to a lesser extent, on personal fieldwork and other published references, in the development of her hypotheses.

Andrews (1983) appears to have become confused as to the affinities of the Lethen Bar fishbed, Lethen Bar Hill, and Cairn Bar. On the first page of her 1983 paper (p. 243), she stated that ‘Lethen Bar (or Cairn Bar) is a hill some 8–9 km SE of Nairn’ and repeated this statement with further elaboration on page 259. This is true only for the fishbed itself, which lies 8.6 km SE of Nairn (Railway Station, built 1855) in a gently sloping meadow, rather than on a hill; Lethen Bar Hill and Cairn Bar Hill are 9.98 and 10.74 km respectively from the same landmark, so her statement is clearly mistaken. In the same paper (Andrews 1983, p. 259) she recounted that ‘Lethen Bar, Nairnshire (the name is often given as Cairn Bar [sic]) is a hill covering quite a wide area, and the fishbed formed a continuous outcrop around it’. Again, this is incorrect and must be derived from Malcolmson's paper (Malcolmson 1859, pl. XI, fig. 4) in which his section was published, as this is the only primary source of such information that has come to light (see below). Although Andrews (1983) cited this paper, curiously she did not specifically mention any of his sections that she must have seen. These are significant discrepancies, which confirm that, despite visiting the area, albeit perfunctorily, Andrews (1983) firmly believed that the fossil locality named Lethen Bar was a hill 8–9 km SE of Nairn, around the base of which the Lethen Bar outcrop was situated, as only Malcolmson's (1859, 1921) section and text, when taken literally, support this interpretation.

Andrews’ misunderstanding of the topography is difficult to comprehend (Andrews 1983), as contemporaneous OS maps (e.g. OS 1967, 1972) clearly show the topographic detail and meandering contour lines of Braevail indicating gently sloping terrain, and the swirling lines demarcating Lethen Bar and Cairn Bar hills to the SE of the fishbed. Notwithstanding this, she produced a planarly accurate map (Fig. 4) based on archival data and BGS (1923) yet omitted any detail of topographical relief in support of her hill. Unfortunately, in appending a hypothetical and uncorroborated quarry locality (Quarry 7 in Fig. 4), and embracing Malcolmson's (1839, 1859) findings verbatim, Malcolmson's (1839, 1859) findings verbatim, Andrews (1983) unwittingly served only to perpetuate uninterpreted 19th century information.

Perhaps the most surprising and recent consequence of this mélange of contradictory information was that the renowned and determined fieldworker Stan Wood (2001) was persuaded to invest considerable resources in earth-moving equipment, excavating the inferred fishbed adjacent to Malcolmson's section and the quarry hypothesized by Andrews (1983) (Locality 7 in Fig. 4) but found no fishbed strata present.

Wood (2001) produced a sketch map and sections marked with the positions of his test pits on his hand-drawn line of fishbed, superimposed on an enlarged OS base map (probably the 1:25 000 Pathfinder series). No accompanying field notes came to light in the BGS archives that might confirm the rationale for Wood's test pit locations but, critically, Wood provided very useful data, otherwise unavailable, to the BGS and the present study.

An examination of Malcolmson's circumstances and field methods might help us better understand how his Moray Firth fieldwork has been interpreted over the years. Malcolmson had initially returned to his hometown of Forres in 1837 in chronic ill health, but with ambitions to return to India, where he had served as an Army Medical Officer. Working both in the field in Moray and Nairn accompanied by Gordon and Stables, and on his memoir, he was also making trips to London and Caithness between autumn 1838 and late 1839. He appears to have been under self-imposed time pressure, making many discoveries in a very short period; he discovered Dipple, Tynet and Lethen bar between 14 November 1838 and 27 March 1839, before returning to India in 1840, where he died in 1844. Malcolmson was surprisingly well connected in scientific circles and had met both Richard Owen and Charles Darwin; various accounts of Malcolmson's life have been given by Wallace (1921), Andrews (1982) and Keillar (1993). The most comprehensive and recent account of Malcolmson's life is in Mike Taylor and Ralph O'Connor's critical study of Hugh Miller's The Old Red Sandstone (Miller 2023, p. A291).

Malcolmson's fieldwork, scientific caution and generous acknowledgement of the work of others were revered by his peers (Andrews 1982, pp. 22–26). Malcolmson (1921) revealed an intense and comprehensive style in his descriptions of the landscape and outcrops. One who gained particular benefit from Malcolmson's generous nature was Hugh Miller, who, since 1830, had been discovering fossil fishes near his home at Cromarty; due to his effective isolation from learned fellows Miller was struggling with their affinities. When Malcolmson learned of Miller and his collection, he made his way to Cromarty to visit Miller and ignited in him a new understanding of this young science. This set Miller on a much more informed course and helped to establish him as a geologist in both the practical and theoretical sense (Miller 2023, p. A291).

Malcolmson's mapping practices

Malcolmson (1859), with a few exceptions, drew some geological section sketches for his memoir with no apparent attempt to provide accurate or relative scale, with inconsistent measurements, and with exaggerated landforms, all without explanation (Malcolmson 1859, pl. XI, figs 4 and 5, reproduced as Fig. 8a and b in this paper), which seems contrary to his reputation for rigour. His apparently deliberate, exaggerated angles of slope and dip (see below) could be misleading, especially to the non-geologist; elsewhere, in pl. XI, he appears to recognize the importance of representing accurate dip angle, faithfully presenting the gentle dip in the smaller scale sections of Dipple and Tynet Burn.

It is likely that Malcolmson (1859) was influenced by the displaced topography on MacCulloch's map, and it appears that he was using his version of the mapping practice of exaggerating the vertical scale, a process known as vertical exaggeration (VE) (Weijermars 1997). This may be an attempt to fit very long sections, in this case up to 11 km, into a narrow format while retaining topographical detail.

Implementation of VE creates difficulties for a non-geologist; data presented in this way require interpretation using the VE factor, which, nowadays, should be annotated on the map. Exaggeration of the vertical scale increases the angle of dip and slope and therefore the height of raised landforms (creating the impression of hills) and the depth of topographic lows, shown by the impossible depth of the River Findhorn (Fig. 8a).

It should be noted that the BGS has used isometric scales (Fig. 6b) to provide an accurate representation of the dip angle and of the topography across the fishbed. Unfortunately, Malcolmson (1859) did not provide justification or explanation for his methods, although this may have been lost when his manuscript and sections were divided up for publication. Horne (1923, p. 75, fig. 2) also used VE but helpfully marked the horizontal scale in miles and the vertical scale in feet.

Consequently, there are gaps in our understanding of how and why Malcolmson (1859) arrived at his conclusions, which need to be considered. It should also be appreciated that, because of the infancy of their disciplines, variations by different workers were not unusual in the early 19th century. Innovative styles in geological and landform description were evident, for example, in the work of William Smith (1769–1839), who had previously implemented VE (Palmer 2020, pp. 30–31); Martin, Malcolmson and their peers were of their time, as were their maps and sections.

Malcolmson's sections of Lethen Bar and Clune

We have retained the general term ‘gneiss’ for the metasedimentary rocks to allow correlation with Malcolmson's (1859) terminology, and the subdivisions of the Old Red Sandstone shown by Malcolmson reflect what was generally believed at the time.

We know that Malcolmson (1839) was working with a topographical base map, but the actual document that he was relying on has not been identified. In a letter to geologist William Lonsdale (1794–1871) in the GSL archives dated 28 August 1839 (GSL/L/R/4/266-266a), Malcolmson stated that he was annotating geological features onto a map, which could be the missing map detailed in (1) in the timeline above. In the same letter to Lonsdale, he referred to the sections of Lethen and Clune (Fig. 8a and b) as showing ‘where the sandstone and limestone are principally exposed’, suggesting that Lonsdale may have received a copy of Malcolmson's (1839) sections with the letter. Malcolmson was also working with maps borrowed from Gordon (Collie and Bennett 1996, p. 7). It is not clear as to what extent Malcolmson (1839) relied on a geological map as a template or whether he gave precedence to his field observations; there is evidence that Malcolmson respected MacCulloch's (1836) delineation of the geology of the area and thus used his map (Malcolmson 1859, p. 338). His adoption of MacCulloch's misplaced features may have influenced some of the geological and topographical inconsistencies evident in his sections.

Taken out of context, Malcolmson's (1859) section (Fig. 8a) gives the immediate impression that the Lethen fishbed is a prolific outcrop around a hill, and that it caps a second significant hill, which we have already shown to be incorrect. As discussed, this anomaly is best explained by his use of his rudimentary form of VE. Other anomalies are less easily explained: the hill of granite and gneiss that he controversially renamed ‘Cairn Bar Hill’ is correctly Lethen Bar Hill, with the intervening exposure of gneiss omitted (Fig. 6a); he also showed old Lethen Bar Quarry on the eastern margin of the hilly outlier containing the fishbed, but this locality is Lethen Bar Farm (Braevail), which, as discussed above, lies in a gently sloping meadow.

Malcolmson (1859) omitted directional data from this sketch (Fig. 8a), but the compass direction and relative distances can be deduced from two landmarks, rendering it a straight-line section. The section intersects ‘Lethen’ with a distance around Lethen House of 0.81 km and concludes at ‘Coulmony’, again with a lateral extent of about 1.0 km. To include both the fishbed and the granite intrusion shown, this analysis assumes that the ‘best-fit’ line has Lethen House Quarry as the point of intersection (labelled Lethen), and that Coulmony House lies about 4.75 km to the SE as the second point of intersection (Figs 6a and 8a). This section only just intersects the fishbed and does so in an area where we now know that the fishbed is at its most sparse and disrupted (Fig. 6a); the outcrop is thus not continuous and prolific as shown by Malcolmson (1859). This section intersects the igneous intrusion at Lethen Bar Hill but does not intersect Cairn Bar as shown in Figure 8a as this hill is 1.0 km east of the line of section (Figs 6a and 9). The position of the granite intrusion at Lethen Bar on Malcolmson's (1859) section (Fig. 8a) is significant, as he drew it abutting the edge of the conglomerate underlying the fishbed, omitting the intervening 600 m of gneiss, just as MacCulloch (1836) mistakenly showed it on his geological map.

Whereas the anomalies in Figure 8a could reflect the prevailing methods and liberal presentation style of the early 19th century, Malcolmson's second section (Fig. 8b) has similar inconsistencies in scale and relative distance, but having already discussed his presentation method, we shall describe only its pertinent, but erroneous features. The section (red section line in Fig. 6a) is drawn to the ‘north by east’ (i.e. SSW–NNE), across the outcrop through Clune Quarry, across Chapel Hill, to Boghole, terminating at Culbin Forest (BGS 2012; OS 2015). This line omits Chapel Hill and the northern part of the fishbed, highlighted with the white arrow in Figure 8b. In place of Chapel Hill, Malcolmson (1859) drew the fishbed continuing to dip at an overstated angle NNE from Clune, and not re-emerging; this is incorrect as evidenced by Horne (BGS 1923) and as shown in Figure 6a. When referring to Clune in the text, Malcolmson (1921, p. 454) alluded to ‘Five hundred yards higher up the hill’ where quarries were previously worked in the area; the only hill in the vicinity to include quarries is Chapel Hill, but, as this feature is omitted from the section (Fig. 8b), it appears that he may not be discussing this landmark.

At the opposite south by west end of the section, he drew an upward change in gradient, creating a hill that follows the erroneous angle of dip of the gneiss bedrock; he labelled this Cairn Bar Hill, immediately to the south of Clune, which reflects the displaced hill on the map of MacCulloch (1836). Again, no such overstated topography exists in this direction, the terrain being a gentle slope, and the nearest hill, Lethen Bar Hill, is situated 1.3 km to the south. Given that Malcolmson's section in Figure 8a is more relevant to the present study, this will be the focus of the remaining discussion.

Cairn Bar and Lethen Bar

Today, Lethen Bar Hill, despite being the higher of the two hills, is difficult to see from afar as it is narrow and masked by trees; Cairn Bar is a broader feature with less tree cover and thus appears more prominent. The latter is rarely named on maps, but Figure 9 (OS 2015) shows that Cairn Bar Hill is separated from the high ground of Lethen Bar Hill by a shallow valley; a road passes through the valley and the demarcation of the two hills is obvious in the field.

The annotated 2012 geological map (Fig. 6a) also shows that the two hills are separate features and that the granite exposed on Lethen Bar does not extend as far as Cairn Bar, as depicted in Malcolmson's (1859) section. In his section, Malcolmson presented the geology of Lethen Bar Hill (Fig. 8a) but mistakenly labelled it ‘Cairn Bar Hill’, leading to the confusion that persists to the present day. He also wrote that ‘the hill of Cairnbar, which is in part composed of gneiss: and on the south side, near Coulmony, of granite’ (1921, p. 452), but again his description matches Lethen Bar Hill rather than Cairn Bar Hill (Fig. 6a).

A reappraisal of Malcolmson's section

Significantly, both Falconer and Murchison (Andrews 1982) supported the initial publication of Malcolmson's manuscript (Malcolmson 1839), and the published memoir (Gordon 1859; Malcolmson 1859) was subsequently endorsed by Horne (1923). Horne never criticized Malcolmson's (1859) interpretation of the Old Red Sandstone stratigraphy or his sections of the area in question. If he thought that there were fundamental errors in Malcolmson's work, why did he not say so? The simplest explanation is that Malcolmson's (1859) depictions were taken as accepted practice and deliberate, which were interpreted accordingly.

As discussed in the introductory paragraphs, the name ‘Lethen Bar’ was formerly given to both Lethen Bar Farm, where the fishbed lies, and to Lethen Bar Hill, a separate metamorphic and igneous feature 1.38 km SE of the farmhouse. Malcolmson (1859) may have recognized the potential for confusion and thus chose not to use the same name for both sites, renaming the latter as Cairn Bar Hill, perhaps encouraged by the cairn on the summit of Lethen Bar Hill. However, if this was a deliberate action, his lack of explanation only produced greater uncertainty. Malcolmson's attempt to delineate the two landmarks is supported by his memoir (Malcolmson 1921, p. 454), in which he wrote that, from a position on the NW bank of Muckle Burn (an apparent reference to his section in Fig. 8a), ‘If these shales were prolonged across the stream in the line of the dip they would pass over the conglomerate and strike the hill of Cairnbar [our italics] near an excavation recently made on the farm of Lethenbar where the finest fish have been procured’ (Figs 4 and 8a). This obtuse statement created considerable confusion for Andrews (1983), discussed below. The fishbed lies below the surface of the gently sloping meadowland in a trough in the underlying rocks (Fig. 6b), which crucially Malcolmson could not have known. He may have decided to accentuate the local relief so that he could show the fishbed as a continuous feature and the use of an exaggerated vertical scale would facilitate this false interpretation.

Malcolmson's (1859) section also shows, incorrectly, that the granite intrusion abuts the fishbed. A possible explanation for this is that he believed the intrusion to be displaced to the north, as depicted on MacCulloch's (1836) map, allowing him to forego some fieldwork as MacCulloch, whom he trusted, had already mapped the granite in this area. However, this map incorrectly places the north slope of the hill adjacent to Clune and the granite abutting Lethen Bar Farm (Braevail), accounting for the omission and foreshortening of the gneiss expanse on Malcolmson's section (Fig. 8a); Figure 6a shows the full extent of the gneiss outcrop along Malcolmson's line of section.

Over the years, Malcolmson's (1859) account of the stratigraphy has led to much confusion, especially as his renaming and redrawing of features was never qualified or explained when his memoir was first published. Subsequently, his naming of the topography at old Lethen Bar was not supported by the BGS; by 1878, Horne had reverted on his field slips to the ‘old Quarry of Lethen Bar’ without comment and never referred to Cairn Bar. Andrews (1983), on the other hand, regarded the two names as interchangeable, which, both geographically and geologically, is incorrect.

Horne (1923) and Andrews (1983), both competent scientists themselves, appear to have held Malcolmson in the highest esteem, despite coming to different conclusions about his work. Their disparate scientific backgrounds may explain the differing interpretations that they placed on his memoir.

Dr John Horne's background was as a BGS field geologist and Lead Surveyor; in mapping the area he would have realized very quickly that Malcolmson (1839) had been misled by MacCulloch's earlier map. Horne's field party (Horne 1923) re-mapped the granite correctly and, in the process, reverted to the correct name of ‘Lethen Bar’ for the granite and gneiss hill. He probably also recognized that, in depicting a hill with a nodule bed around its base, Malcolmson (1859) was applying VE to his section, which would have required only simple adjustment for Horne to interpret Malcolmson's work.

Dr Mahala Andrews, on the other hand, was a twentieth century museum-based palaeontologist: her approach was based on a review of the literature and she had no reason to doubt (or reinterpret) Malcolmson's (1839, 1859) text or sections. She appears to have taken these at face value, although she implied in her synopsis (Andrews 1983, p. 243) that she lacked both the time and resources to invest in a more detailed study of this area, notably performing only a superficial level of fieldwork. Consequently, to the detriment of her analysis, her findings on the environs of the old localities occupy only three pages at the end of her 1983 paper.

For whatever reason, Malcolmson's misinterpreted account of the geology (Malcolmson 1859) quickly became embedded in palaeontological folklore, despite the much later, more accurate work by Horne (1923). This resulted, at an early stage, in other researchers faithfully reiterating Malcolmson's (1839, 1859) representation as to the nature and position of the celebrated fishbed, on top of a hill (e.g. McBean 1845; Groome 1884).

Andrews’ hypothesis; a continuously moving Lethen Bar?

Andrews’ (1983) Quarry 7 (Fig. 4) was based only on her hypothesis that ‘Lethen Bar’ was simply the name given to whichever operational site was open as the workings progressively followed the outcrop in an anticlockwise direction, stating (Andrews 1983, p. 260) ‘Thus we might expect that the site of operation would shift considerably with the passage of time’. She made no such reference to the two quarries on Chapel Hill or, indeed, to Clune Quarry in this respect, the site of which has been unchanged since as far back as 1827.

Malcolmson's (1859) misleading topographical statements appear to be the origin of Andrews’ (1983) hypothesis, suggested in part by her placement of Quarry 7 adjacent to his line of section (Fig. 4). Following Malcolmson (1839, 1921, p. 454), she extended the fishbed from the north bank of Muckle Burn to ‘strike Cairn Bar hill on the farm of Lethen Bar’ (Andrews 1983, p. 262); she thus deduced that this would be on the ‘west or northwest side of the outcrop’. As a direct consequence of this argument, Andrews (1983, p. 262) concluded that, in 1839, the westerly Quarry 7 (Fig. 4) was ‘probably’ the operational site of old Lethen Bar Quarry, thereby suggesting that the quarry working had traversed around the outcrop in an anticlockwise direction from the north, culminating in cessation of quarrying ‘in the south in 1880’. Andrews also wrote (Andrews 1983, p. 247) that fossils continued to be collected from Lethen Bar ‘as known from abundant material collected between 1850 and 1880’.

Andrews visited the ‘region of the west side of the outcrop’ (Andrews 1983, p. 262) in 1974, but did not specify the OS grid references or localities examined. She observed an area of rough pasture with an uneven surface, which she took as evidence of exhausted quarry workings, concluding ‘All appearances thus indicate that the bed in this region was exhausted’. At best, this provides only circumstantial support for her theory.

Horne's field slip (Horne 1878, Lethen Bar FS3) covers the area containing Quarry 7; there is no annotation on the line of the fishbed where Andrews (1983) fixed her hypothetical quarry, indicating that nothing of interest was found, but several annotations indicate significant deposits of boulder clay immediately to the west of the line. Horne's map (BGS 1923) indicates that the western edge of the outlier was an uncertain geological boundary (indicated by a dashed line in Fig. 4), a fact that Andrews highlighted as ‘significantly?’ (Andrews 1983, p. 262). Without explanation, Andrews and, to an extent, Wood seem to have set aside this uncertainty when they were working in the field: Andrews (1983, fig. 4 caption) marked Quarry 7 as the ‘possible region of Lethen Bar Quarry’ on the uncertain fishbed line (Fig. 4), as supported by her ‘evidence’ from her 1974 field visit.

The modern BGS map (BGS 2012) shows that Horne's caution was justified; the map depicts the SW sector of the outcrop with more certainty as about 250 m east of Andrews’ (1983) hypothetical Quarry 7. Unfortunately, it appears that Wood (2001), apparently influenced by Andrews’ hypothesis (Andrews 1983), dug two of his test pits adjacent to Quarry 7, too far west of the fishbed (Fig. 6a). Further to the north, near Wester Clune, the line of the fishbed is displaced by two faults (compare Figs 4 and 6a), making the western region of the fishbed difficult to locate or follow in the field.

The only physical evidence for potential west flank quarrying is the mapped presence of two, possibly three, lime kilns to the SW and west (Fig. 5). In the SW, a lime kiln is marked on Horne's base map (OS 1871), and Wood's (2001) report of limestone fragments in the area supports this finding. Horne's ambiguous entry of a second SW kiln on the field slip (Horne 1878, Lethen Bar FS3) seems unreliable, as it does not appear on any map; notwithstanding this, Wood's fieldwork and excavations have shown that no trace of a quarry has been found. The possibility remains that the SW kiln(s) serviced old Lethen Bar Quarry as this was the nearest suitable terrain in which to build a kiln. Old Lethen Bar Quarry is known to be a prolific producer of non-nodular limestone (Horne 1923), which would explain Wood's limestone fragments.

The kiln at Garrowstripe to the north (Fig. 5) presents an enigma as previously discussed. It is likely that excavation would be required to investigate whether a quarry existed and what its extent might have been, which is beyond the scope of this study. Nevertheless, kiln evidence cannot be disregarded, and the possibility remains that a quarry did exist here, perhaps to service only the needs of the small farm, which might further explain why it may have gone unnoticed.

However, even though the previous existence of west flank quarries appears plausible, it provides little, if any, support to a moving quarry theory.

Historical evidence for static quarries

The position of the ‘old Quarry of Lethen Bar’ (Quarry 1 in Fig. 4) is clearly annotated on Horne's field slip (Horne 1878, Lethen Bar FS1) and was found to be both present and productive when it was reopened in 1878, contradicting Andrews’ (1983) suggestion of continuous trenching having exhausted the outcrop. In a series of articles about the geology of Moray for the Elgin Courant in 1841, Duff referred to the Altyre Collection as having been collected from the ‘shale or fishbeds of Clune, Lethenbar, and Lethen’. Martin's 1842 map shows that the quarries of Lethen Bar and Clune were some distance apart, confirming that these sites were separate and contemporaneous at the time; Constance Gordon Cumming (1904), in recalling her mother's exploits, referred to ‘Lethen Bar Lime Quarries’ in the plural, again indicating that there was more than one working site.

Both Duff (1842) and Miller (2022) reported old Lethen Bar and Clune quarries as being open on the east flank of the BGS Outcrop in 1841–42 and in 1844 respectively. Duff (1842, p. 34) described the method of working of both quarries, and Miller, sometime between 1844 and 1856, referred to Clune Quarry as having been active ‘for a considerable number of years’ (Miller 2022, pp. 190–193). Miller went on to describe old Lethen Bar Quarry as being ‘a neighbouring opening in the same bed’, again implying that they were not part of a continuum and confirming an easterly rather than westerly position for the workings.

The evidence for isolated, static quarries seems overwhelming, as evidenced by Table 1, which illustrates the published references to the BGS Outcrop Localities Lethen Bar (old Lethen Bar) and Clune Quarries chronologically; both localities were consistently associated with each other by various authors over at least four decades. Andrews’ (1983) hypotheses of Lethen Bar Quarry being sited on the west of the outlier in 1839, having moved there following the outcrop from the north (1983, p. 262), is contentious and this study has not found any evidence for a moving quarry named Lethen Bar, or for any exploitation of the fishbed between quarries. Our fieldwork supports Horne's mapping of the area, which showed that quarries were being worked at specific, fixed localities (Horne 1923).

There is disagreement between historical and modern accounts of quarry depths. Andrews (1983) described the depth of excavation of the nodules, generalizing that they were collected from just under the topsoil down to a depth of 4.5–6 ft. This is not true for all the localities recorded by Horne (1923). He logged that the fishbed at old Lethen Bar Quarry lay under 7 ft of superficial deposits with a total depth of 13 ft, and the quarry adjacent to Easter Clune Farm (Fig. 4, Locality 3) was said to lie under 9 ft of clay (Horne 1878, Easter Clune FS1).

Duff (1842, p. 34) noted old Lethen Bar and Clune quarries as being ‘worked by removing the vegetable mould’ and that at the quarry face ‘the fishbed … rests on schistose sandstone, highly micaceous, which again lies on a coarse conglomerate of a dark red colour’. Although this might have been correct for Clune in his time, it appears at odds with Horne's description of old Lethen Bar being under 7 ft of superficial deposits (Horne 1923). This, however, may be explained by the method of working the quarry, where it was opened in portions (Andrews 1983, p. 261) perhaps encountering differing thicknesses of superficial deposits. Miller (2022, pp. 190–194) on a second visit to Clune observed a quarryman removing nodules from one notable section near the top of the exposure, all of which contain ichthyolites. He described the deposit as being immediately below the moory soil and trenched down to 6 ft but did not give further details of the outcrop. Notwithstanding Duff's observations, this information generally correlates with the thickness and depth of the main Lethen House outcrop, which Horne's party logged (Horne 1923; see below). However, as discussed above, there are depth variations across the localities.

The intermittent presence of superficial deposits and boulder clay across the area (Horne 1878) thus governed the depth of individual workings, rather than the thickness of the fishbed, which is broadly uniform across the sites.

A survey of the entire area was conducted between October 2021 and June 2022 (R.G.D.) to trace historical sites and to find evidence of old workings on the line of the outcrop. Our successful fieldwork provided some OS grid references for the sites of fishbed excavation and for the quarries identified by Horne (1923). These are shown as Localities A–H in Figure 10, in conjunction with the BGS (BGS 2012) fossil sites (Fig. 6a). We have adopted an alphabetical system for the static quarries, starting in the SE and proceeding around the outcrop in an anticlockwise direction, to clearly delineate our findings from those of past workers.

Locality A is the ‘old Quarry of Lethen Bar’ as pinpointed by Horne (1878, Lethen Bar FS1), long since ploughed over, but, according to the farmer, it lies adjacent to an area where there is a local concentration of clay in the topsoil. The fishbed is reported to lie under 7 ft of superficial deposits presumed to include boulder clay (Horne 1923, pp. 72–74). Wood requested permission to excavate at this locality in 2001 but was refused owing to seed planting.

With the assistance of the Lethen Estate forester, an old, isolated working was located by the roadside, close to the edge of a wooded area on the Braevail Farm boundary (Quarry B in Fig. 10), which is the newly recognized site of ‘Clune Pits’. The two adjoining marsh-filled pits are approximately 20 m by 8 m and are surrounded by heavily disturbed ground. These are well known to local people but have not attracted the interest of geologists since the time of the BGS survey in 1878. Horne marked this locality on his field slip (Horne 1878, Lethen Bar FS1) with an ‘x’ and annotated ‘red B. Clay, S.S. g narl …? sch, gneiss, g rec, SIL’ were seen, thereby ruling out a gravel pit; because of their relatively large size it is unclear why or by whom these pits were dug. Horne's field slips may provide insights into the working methods of his party in 1878; elsewhere on the outcrop Horne described digging a ditch (Horne 1878, Easter Clune FS 1), (Fig. 10, Locality D) with the annotation ‘ls nodules in clay seen when making ditch’. He also described seeing ‘grey quartzite and gneissose rocks in ditch’ at a point about 220 m ESE from Clune Quarry (Horne 1878, Lethen Bar FS1), both of which suggest smaller scale investigative practices. Based on this observation, it seems unlikely that the BGS dug these large pits (Locality B), because of the time and effort required for such little information gain, and no similar pits, except for Clune Quarry (Locality C), were found by the present study. Nevertheless, the proximity of these pits to the line of the fishbed provides compelling contemporaneous evidence for trial pits targeting the fishbed, or active limestone working, although no kiln was found in the vicinity, or on old maps.

Locality C is the site of Clune Quarry, again mapped by Martin (Duff 1842) and with considerable accuracy by Horne (1878, Lethen Bar FS1). Today, all that remains of the workings is an isolated, wide marsh-filled pit about 80 m by 20 m situated directly on the line of the fishbed, flanked by an old well to the SW and a lime kiln to the north (OS 1871). On occasion, a stream intermittently flows through the old quarry, which has the lime kiln set in its bank a short distance to the north. The position of the well has changed over the years but continues to be associated with the fishbed via a new well situated on the bank across the quarry pit from the old well to the NE. This well supplies water to Easter Clune Farm where the operation of a lime treatment plant indicates that a source of carbonate still exists in the subsurface at this locality.

Location D is the aforementioned ditch dug by Horne's party in which nodules in clay were seen (Horne 1878, Easter Clune FS1). Horne did not provide a depth, but the inference is that this opening was of a shallow nature.

In the north of the BGS Outcrop, the historical positions of the quarries can be obtained from Horne's (1878) field slips and from the memoir for this area (Horne 1923, pp. 72–74). His quarries are shown as Localities E and F in Figure 10, and these correspond approximately to Andrews’ (1983) Quarries 4 and 3 respectively; both sites have an adjacent kiln. Whereas little information was recorded for Quarry E, Horne (1878, Easter Clune FS1) provided greater insight into Quarry F, writing that ‘The best ls [limestone] is said to have been found under 9 ft of clay’ and that ‘the nodular bed is said to be 4 ft thick’ and ‘underlain by sandstone’.

Although Localities E and F are difficult to trace today, hillshaded LiDAR (light detecting and ranging) data available on the NLS website (NLS 2021) may provide a position for Locality E. Both Locality B and Locality C can clearly be seen but the LiDAR scans are incomplete in this area and there are no data for the ground containing Locality F. However, in the vicinity of Locality E pinpointed by Horne (1878, Easter Clune FS1), there is a hillshaded inflection indicating a localized change in elevation of the scrubland terrain.

Quarry F may be associated with the ‘Gravel Pit’ and ‘Old Limekiln' annotated on a single map, the 25 inch:1 mile OS map (OS 1905) in an area where Horne recorded no gravel, only ‘red clay with …. schist stones’ (Horne 1878, Easter Clune FS1). Despite referring to Horne's (1878) quarry positions, Andrews (1983) placed Quarry F (Fig. 4, Locality 3) due west of the kiln whereas Horne placed it to the NNW (Horne 1878, Easter Clune FS1). Nodules found at these northernmost localities in the past were not reported to exhibit the fine preservation of fishes evident in those from old Lethen Bar Quarry and Clune.

At Easter Clune Farm, there is a deep, elongated cutting about 20 m from the farm buildings, on the opposite side of the road (Location G in Fig. 10). This significant feature sits directly on the fishbed where it broadens out towards the south and measures about 6 m wide by up to 100 m long and is 3–4 m deep. A narrow stream cuts through the floor of the pit to the NW, which indicates that, rather than being a worked locality, this is a meander scar through superficial deposits and is thus a probable natural exposure of the fishbed. This feature was first shown on Horne's field slips (Horne 1878, Lethen Bar FS1, Easter Clune FS1), apparently mapped directly but without annotation or mention in the accompanying memoir (Horne 1923). However, on the map (BGS 1923), this southerly deviation of the outcrop is curtailed, suggesting a transposition error. This feature appears on Andrews’ (1983) map, as drawn by Horne (1878) on his field slips, and on the most recent map (BGS 2012), (Fig. 10), where it precisely follows Horne's contours. It is unclear whether this site could be productive as it is very overgrown; excavation would be required to determine whether it would yield nodules today and its proximity to the farm may preclude this. Location H is Lethen House Quarry.

We have not been able to confirm the location of any quarry on the western margin of the BGS Outcrop; Wood's two southwesterly pits were too far west to allow him to find evidence of earlier activity (Fig. 6a). Most of the ground overlying the fishbed in this area is prime agricultural land and the current farmers were unable to say whether there is evidence for past disturbance in the meadows and there are no NLS LiDAR data available for the west flank. One small patch of woodland remains SW of Wester Clune towards Garrowstripe (Fig. 5), but this is difficult to access; a short search in 2021 found no evidence of pits but the kiln at Garrowstripe, along with the kiln about 400 m to the SW (Fig. 5), remains significant. Horne (1878) covered this area in detail in the field slip (Horne 1878, Lethen Bar FS3) but reported no quarries. Nevertheless, the existence of two west flank kilns suggests that at least one more quarry is still to be located.

Today, the most accessible exposure of the Lethen fishbed is the main outcrop at Lethen House Quarry and the downstream exposure (OS 1870b), adjacent to the Muckle Burn, on its NW bank (Fig. 11). This exposure can be traced some distance on either side, although at present only the overlying fine sandstone can be seen, the lower units being masked by scree. It appears that, despite Horne describing this exposure as containing ‘the best section’ (Horne 1923, p. 23), Andrews (1983, p. 249) did not investigate Lethen House Quarry, as she only mentions in passing ‘nodule bearing shales as seen at Dipple, Tynet and Lethen House’ and does not describe the stratigraphy either in Horne's memoir (Horne 1923, p. 73) or at outcrop. Apart from Horne's (1923, pp. 72–74) basic descriptions of Lethen House and old Lethen Bar quarries, accounts of the fishbed deposits in the wider Lethen area are very limited and no scientifically detailed sections from any of the Moray Firth fish-bearing nodule beds existed in Andrews’ time (Andrews 1983, p. 247).

The outcrops show a fluvial to lacustrine cycle representing periodic transgression by rising lake waters; members of the Achanarras fauna, Osteolepis, Dipterus, Orcadacanthus (previously Mesacanthus; Newman et al. 2023) and Coccosteus are present and fossil plant debris is common in the fine sandstones. Nodule morphologies vary from smooth to calcite encrusted, flat bottomed to round to lenticular, and all show laminations to varying extents, discussed in detail below.

To aid description, we have differentiated the fossiliferous horizon into Top, Middle and Bottom units (Fig. 12). The Top Unit contains fossil plant debris and is underlain by 70 cm of a laminated shale (the Middle Unit) containing laminated nodules, occasionally bearing fossil fishes, and isolated calcite veins; the nodules vary in morphology from the upper to lower parts of this unit. The Bottom Unit comprises 70 cm of fine laminated micaceous sandstone with fossil plant debris and overlies a red medium or coarse sandstone containing mud clasts and pebbles. The Middle Unit of laminated shale with concretions is remarkably rich in fish remains: six specimens were found by trowelling the surface (this site is not a designated Site of Special Scientific Interest (SSSI)).

The only historical account of the stratigraphy at Lethen House Quarry is in the BGS memoir (Table 2). The information presented in Figure 12 and Table 2 shows fair agreement in the overall stratigraphy and thickness of the two sections, although there are differences in the subdivision and description of the sediments. Unit 1 corresponds well to our basal red medium or coarse sandstone, whereas Unit 2 is comparable with our Bottom Unit. Units 3 and 4 equate to our Middle Unit, and Units 5 and 6 to our Top Unit. The reason for the descriptive differences is unclear, as Figure 12 and Table 2 both refer to the same quarry face. Horne's (1923) account seems erroneous but may reflect the terminology of lithofacies used in the 19th century; no standard for sediment grain size existed until the late 1800s (Udden 1898) and thus interpretations made by fieldworkers were subjective. The modern map (BGS 2012) shows an underlying sandstone at this locality (Fig. 6a), rather than the conglomerate described in historical accounts of the area; both Unit 1 (Table 2) and our basal sandstone (Fig. 12) support the BGS data.

Horne (1923, pp. 72–74) gave an indirect description of the fossiliferous nodule band at Quarries D and E (Fig. 10) ‘from information obtained at that time’, as both sites were backfilled when he carried out their survey. He gives the thickness of the nodule bands as ‘about four feet thick … covered by clay and shale with occasional limestone nodules’. This is consistent with the range seen in the thickness of the Lethen House exposures and in Units 2–4 in Table 2. Similar observations were made by both Duff (1842) and Miller (2023) at Clune. Horne (1923) also described the reopened Lethen Bar Quarry (Locality A in Fig. 10), which was dug by Mr Jex. This section is deeper than others and is subdivided into two richly fossiliferous bands separated by a limestone band (Table 3), with the nodules and shales varying ‘in appearance’ from layer to layer. Horne (1923) described the sediments underlying Sub-zone 1 (the lower nodule band) as calcareous sandstone but did not elaborate further. This is the first definitive account of multiple nodule-bearing bands at Lethen Bar and indicates that significant local stratigraphic variation may occur.

Evidence of limestone bands at both Lethen Bar and Clune

Table 3 shows that Horne's (1923) Sub-zones 2 and possibly 4 are limestone bands, differentiating this section from that of Lethen House in Table 2. Interestingly, Malcolmson (1839) made a reference to ‘Clune limestone with ichthyolites’ (Andrews 1983, p. 262) and Gordon (1859, p. 36), quoting Malcolmson, descibed Clune as a locality where ‘nodules and flat slabs of limestone occur’. It is curious that neither Duff (1842) nor Miller (2022, pp. 191–194) mentioned these limestone bands in their descriptions of the quarry faces. It is therefore feasible that significant lateral variation occurs and that the limestone bands were present only intermittently as the quarries were widened. In 2022, on a field survey of Clune, a few fragments of limestone slabs with crystalline calcite veins were found up to 1 m from the mouth of the lime kiln (NMSUK G.2023.15.1). As there is no other local source of limestone, this confirms the old accounts that the Clune outcrop contained limestone bands, at least intermittently. Similar limestone bands are assocated with the fishbeds at Edderton, at Cromarty (R.G.D., personal observation) and at Tynet Burn (Trewin and Davidson 1999).

The current site at Lethen House Quarry and the adjaent exposure do not contain limestone bands. Fish preservation here is of a lower aesthetic quality than that seen at old Lethen Bar and Clune, suggesting that the presence of limestone bands may be linked to the enhanced preservation of fishes at only these localities.

Comparison with other Moray Firth nodule localities

During the period 1996–2005, two of the present authors (N.H.T. and R.G.D.) logged the nodule beds exposed around the Moray Firth, except for Clava and Dipple Brae, owing to access issues; this showed general agreement with previous published sections (Miller 1841; Horne 1923; Dineley 1999), confirming that there are local depositional variations across the different outcrops.

Lethen House Quarry and adjacent exposure

The Lethen House exposures [NH936515] present a depositional sequence containing a fluvial to lacustrine cycle, with a basal sandstone deposited in a stream or river setting, fining upwards to a laminated shale containing concretions formed in shallow to deep lake conditions, when the ancient lake level rose and transgressed the shore (Fig. 1); the fishbed does not contain limestone bands and is overlain by fine sandstone. The fishbed crops out at other places to the west of Lethen House Quarry and these outcrops are shown in BGS (2012).

Tynet Burn

To date, Tynet Burn [NJ383616] has been investigated in the greatest detail and shows the most complete evidence for several lake transgressions (see Trewin and Davidson 1999, figs 4 and 5 for full details). The outcrop contains two fish-bearing nodule beds separated by approximately 18 m of strata, which includes alluvial plain, fluvial channel and both shallow and deep lacustrine deposits. The lower nodule bed contains only rare fish-bearing nodules and the upper nodule bed contains fish-bearing nodules in shales and a platey limestone. The latter has three distinct units (Bottom, Middle and Top), which are overlain by mudstones interbedded with massive limestones, sandstone beds and limestone slabs. Permission to re-excavate the site would be required if additional in situ material was to be collected today.

Gamrie

The easternmost site is at Gamrie [NJ796635] where the Den of Findon SSSI is set in the east bank of the ravine to the west of Findon Farmhouse. The outcrop comprises a single fluvial to lacustrine cycle of shale and mudstones underlain by conglomerate; the erosional surface of the top layer of mudstone is overlain by breccia. The shales and mudstones are 2.6 m thick and the fish-bearing nodules are confined to the lowermost 1.2 m. The flat to round, smooth nodules exhibit similar morphologies to Lethen House with fibrous calcite coats and calcite veins, but no limestone bands; a description and section has been given by Dineley (1999). Several outcrops have been recorded throughout the area (e.g. Corbie Den), and specimens have been collected on the beach below the outcrops.

Cromarty

On the Black Isle, the Cromarty SSSI beds [NH797672] are exposed on the south foreshore of the Cromarty Firth, 400–800 m east of Cromarty and at Eathie [NH810666], on the SE coast of the peninsula, to the north of an old fishing station. The Cromarty beds dip almost vertically and are mostly in the intertidal zone. Consequently, these are frequently covered by seaweed and shingle, making analysis difficult; experience has shown that the outcrop is best seen after a major storm. An exposure of nodule-bearing shale occurs at the top of the beach in a low cliff, below the footpath at a point where its gradient steepens; this may be one of Miller's (1841, fig. 5) 19th century quarries. Since Miller's time, nodules have been collected only from the shingle at both localities, now designated SSSIs. The sequence at Cromarty rests on conglomerate or breccia and, as exposed, records 32–35 m of sedimentation. This shows at least four fluvial to lacustrine cycles comprising sandstones, shales, limestones, nodule bands and platey limestones. In the frontispiece of his 1841 book, Miller included section sketches showing the general stratigraphy of the nodule-bearing units of Cromarty and of Eathie (Miller 1841, fig. 5 and sections 3 and 4). Although his Cromarty section sketch has no scale, it can be matched to the present low-tide exposure with some effort. Further outcrop is occasionally seen in moving shingle immediately to the west of the Cromarty beds, but this would take significant work to uncover.

Miller did not map the occurrence of nodules on the north shore of the Cromarty Firth where nodules can be found in the shingle and no fishbed section was seen on our survey. These differ from those found on the south shore as some are encrusted with fibrous calcite but contain the same fauna.

In a letter to Duff dated 27 January 1839 (Elgin Museum MS (2) G3/4), Miller gave a more detailed section of the Eathie Fishbed exposed in 2 miles of low cliff face. This outcrop has not been visible in recent years but surveillance, especially after a storm, may provide future opportunity; exposures of the fishbed may occasionally be seen in the intertidal zone if conditions are right and nodules can be collected in the shingle.

Edderton

At the Edderton (Black Park) SSSI [NH678832], there are at least two fishbed exposures on the east and west banks of the Alit Muigh-Bhlaraid burn at Black Park. The eastern outcrop is the main exposure and shows a fluvial to lacustrine sequence of 2.3 m lake-bed sediments underlain and overlain by sandstone. The lacustrine sequence contains abundant fossil-rich concretions in the lower part of the fishbed, succeeded by laminated shales with occasional nodules interbedded with limestones and calcareous siltstones; isolated crystalline calcite bands are also present (see Dineley (1999, pp. 189–193) for a section). Uniquely, the rare, armoured fish Rhamphodopsis threiplandi is recorded as relatively abundant at this site, where it occurs in one specific shale bed at the top of the lacustrine unit (MacFadyen 1994).

Dipple

Dipple Brae SSSI [NJ330588] is now obscured by scree and is difficult to access; hence it not being examined in detail by N.H.T. and R.G.D. The site is 1.6 km west of Fochabers, 300 m SW of Sweethome Cottage; details have been given by Dineley (1999, pp. 205–207), including a section based on ‘MS data by M. A. Rowlands’. This shows a single fluvial to lacustrine cycle, returning to fluvial conditions, containing a limestone overlying a sandy limestone at its base. It is slightly younger than the outcrops above and thus yields a different fauna.

The Lethen House nodules, or concretions, vary from smooth to calcite encrusted and from oval to irregularly shaped. They occasionally contain plants and fish in the classic, but less aesthetic, Lethen Bar pink, crimson and white preservation(s). Although they lack the vibrance of the quintessential Lethen material, there remains the potential for excellent preservation, especially in the three dimensionally preserved armoured fishes. The pink and crimson staining results from chemotrophic bacterial action on decomposing tissue while in the soft sediment phase (Trewin and Knoll 1999; Davidson and Trewin 2005).

Preservation of Lethen Bar fishes in the Altyre Collection

Malcolmson (1839) identified old Lethen Bar as the single source of the finest preserved fishes ‘Lethen Bar … from which the finest fish have been procured’ (Andrews 1983, p. 262). Based entirely on the Altyre Collection, Andrews (1983, pp. 246–247) gave four different, but restricted, preservation styles based on nodule morphology, colour and mineralogy (Fig. 13). This shows that all her defined nodule types (Types A, B, C and D) have surface encrustation, which includes fibrous calcite and shale. Andrews (1983, pp. 247–248) described fish preservation in Type A nodules as ‘very good’ and Type B as ‘superb’, and alluded to them originating from two ‘richly fossiliferous beds’ beds at Lethen Bar’ based on Horne's section (Andrews 1983, p. 247) and the material itself. Interestingly, she did not believe that Types C and D came from Lethen Bar (see below).

Andrews observed that her ‘very good’ Type A nodules were ‘typical of one of the main beds at Lethen Bar’ (Andrews 1983, p. 247) and that her ‘superb’ Type B nodules ‘likely … were a local variation or from a different level at Lethen Bar’ but did not elaborate further (Andrews 1983, p. 248).

The source of the best preserved fossils?

Despite Andrews’ hesitation, there may be justification for speculative correlation with Horne's (1923, pp. 72–74) section (Table 3). He described the nodules from Sub-zones 1 and 3 as ‘richly fossiliferous’ and the fishes from Sub-zone 1 ‘being in a better state of preservation than those in Sub-zone 3’. This could indicate that Andrews’ ‘superb’ Type B nodules may have originated from Horne's (1923, pp. 72–74) Sub-zone 1 and that the ‘very good’ Type A specimens derived from Sub-zone 3, with the caveat that no other evidence appears to exist.

Lethen nodule types and diversity

By confining her research to the Altyre Collection now in the NMS, it appears that Andrews (1983, pp. 246–248) was unable to distinguish a comprehensive range of nodule types (Fig. 13). Andrews (1983, p. 246) categorized Type A ‘very good’ and Type B ‘superb’ as 90% of the collection, with Types C and D, both ‘very good’, forming the remaining 10%. It is difficult, therefore, to rationalize the fate of the ‘poorer’ and ‘less beautiful’ material (Gordon Cumming 1904, p. 39; Andrews 1982, p. 72) relegated to outside storage. The above Types A–D classification of nodule groups is not directly supported by our samples, observations in the field, or by examination of specimens in other institutions conducted for this study. Lethen material in the Nairn Museum Collection shows more diverse nodule morphologies, confirming that fishes occur in several nodule types. However, it only became possible for Andrews to view this collection some years after her classification was published.

The spectrum of Lethen House nodule morphologies ranges from large (around 250 mm long), very irregularly shaped, fish-bearing nodules to small, thin (e.g. 60 mm long by 8 mm thick), oval examples. Other taphonomic variations are present in the Lethen House nodules; for example, occasional flattened nodules showing breakages re-cemented by invading calcareous shale. The calcite-coated types are found predominantly in the top half of the shale unit, with round to lenticular nodules in the lower half. It should be noted that because of the recent excavation no additional digging was necessary for this study; the fresh exposure was logged, and samples were collected in situ and from spoil. A new excavation may yield nodules with the full range of preservation styles, but again is outside the scope of this work.

Information gleaned from the Lethen House exposure proves that fibrous calcite encrusted nodules found in situ are either totally enclosed or that calcite encrustation occurs only on the upper surface of nodules; this may be useful in determining the orientation of some museum specimens.

Comparison of nodules with those from other localities

Andrews (1983, pp. 247–248) identified Types A and B as Lethen nodules but cited potential historical alternatives for the provenance of the latter, eventually arriving at the benefit of the doubt, now supported by the fact that the Nairn Museum Collection is known to contain Types A and B localized to Lethen Bar. Somewhat surprisingly, she concluded that Types C and D were ‘very probably from the two most productive bands at Tynet Burn’. She intimated that, at time of writing (Andrews 1983), experience of identifying the provenance of nodule bed specimens was limited to museum collections. Since then, field activity has recommenced, renewing our understanding of these samples. Although it is perfectly reasonable to think that the Altyre Collection contains some Tynet nodules, the two productive bands at Tynet, the top and bottom units respectively, yield nodules with very different morphologies (Trewin and Davidson 1999). Nodules from the Top Unit are regularly shaped with clear laminations and are shallow to deep in cross-section. Those from the Bottom Unit are irregular, flat and, in many cases, cracked and re-cemented with calcite; Andrews (1983) made no mention of these distinct features. The fibrous calcite coating of her Type C nodule is drawn in the same way as that for the Type B example, but this is not a feature of Tynet nodules. Tynet Bottom Unit nodules have internal calcite only where it infills natural cracks and breakages; there may only be very low-relief extrusion of calcite localized to the crack itself (Fig. 14), rather than the fibrous jacket shown in Type C in Figure 13.

For Type D nodules, the evidence for a Tynet origin is less clear. Andrews (1983) has not shown any fibrous detail in the two small patches on the lower surface, indicating that the adhering material is shale, as depicted for her Type A nodule (Fig. 13); unfortunately, adhering shale is not particularly diagnostic.

Regrettably, no specimen numbers were given by Andrews (1983), preventing a more secure verification but, on this evidence, it can be presumed that nodule Types A, B and C exhibiting calcite encrustation are all from the Lethen fishbed and that Type D material is possibly from Tynet. An objective reappraisal of both the Altyre and the Nairn Museum collections would give a better understanding of preservation types, especially if supported by re-excavation of the Lethen House site, but this is also beyond the scope of this study.

To the untrained eye, some Lethen and Tynet nodules can appear very similar (compare Fig. 2 with Fig. 15), especially as vivid white bone, including bony plates and scales, may show varying degrees of crimson staining, which is a key feature of material from both sites. Tynet nodules vary from pea-sized to around 50 cm and occasionally occur outside the three main units in thin nodule-bearing bands. Nodules from this locality occasionally contain more than one fish but can also contain a single scale.

Nodules from the Gamrie fishbed have a dark grey matrix surrounding black bone. In nodules from the Den of Findon the matrix is weathered to brown, and the bone is buff-coloured owing to interaction with the low-pH groundwater; fibrous calcite also encrusts occasional nodules at this locality.

In the Edderton (Black Park) fishbed nodules around 3–60 cm in size occur in shades of grey to brown inset with black bone. They can be shallow or deep in cross-section, but the largest deep-section nodules are confined to the lower part of the exposed outcrop. Many of these deep nodules are extremely hard and resist splitting by hammer. They may also contain more than one fish.

Nodules from the Cromarty fishbed vary considerably in both colour and hardness from hard, dark grey nodules with black bone to softer brown nodules with brown or black bone; some of those from the shingle on the north side of the Cromarty Firth are calcite encrusted. The size range is around 3–40 cm, and they can be deep to moderately deep in cross-section. The grey nodules can be extremely hard, making it difficult to expose the fossil cleanly.

The Dipple fishbed is different in character, containing bulbous ‘execrences’ of nodular material that is usually flat in cross-section (Parnell 1983). The nodules have a reddish-brown matrix enclosing degraded, but well-preserved white bone, and range in size from around 5 to 18 cm.

The inconsistencies regarding the provenance of the Lethen Bar Quarries have persisted for so long and seem to be due to the unreserved acceptance of previous endeavours. Little effort has been made to challenge, verify or explain these earlier findings, especially with respect to mapping and fieldwork. Notably, Andrews (1983, p. 262) performed only a ‘visit’ to the west flank of the fishbed (BGS Outcrop), and in the development of her hypothesis generally gave precedence to unpublished literary material over more authoritative BGS mapping data. Whereas Andrews (1983) confirmed five old quarries citing the BGS memoir (BGS 1923), the present study has identified seven localities including a BGS trial pit where the fishbed was worked. Furthermore, from kiln evidence, there is a probability that one, or possibly two, quarries existed on the west flank.

Lethen House Quarry is the only site that was abandoned and apparently not exploited for either lime or fossil fish. Malcolmson in his 1839 manuscript (Andrews 1983, p. 262) recorded that the quarry showed poorer fossil preservation than that seen at Clune, but preservation quality would have been important only to collectors; of greater interest to the lime workers would have been its commercial, rather than scientific value. The Lethen House outcrop is traceable as a linear feature over a wide area and could be worked only by digging downdip into the outcrop (Fig. 6b); trees and overburden would have had to be removed, making the site harder to work by hand than the shallow pits elsewhere. This may have made the quarry unviable and, as there was no incentive to exploit the outcrop, it was abandoned intact.

The serendipitous rediscovery of Lethen House Quarry, and the survival of the adjacent exposure, has generated renewed interest in the locality, supported by the publication of new information since Andrews’ (1983) paper, discussed above. A rigorous overhaul of historical accounts, coupled with new fieldwork, has also contributed to this revival. The quarry is of historical value and its preservation would ensure the conservation of one of the few accessible and exposed Middle Old Red Sandstone fish-bearing nodule localities. Apart from the quarry, there is potential for the discovery of new scientific data from the main outcrop. The exposed face is some distance from both the Muckle Burn and from the farm track; removal of overburden would be manageable with earth-moving equipment, making excavation feasible, given the appropriate permissions. Historically, the BGS Outcrop (Fig. 10) has shown that the occurrence and preservation of fossils is highly variable, and the main outcrop should be expected to show the same variability.

Despite the inferred expanse of undisturbed fishbed, any future attempts to find well-preserved fishes in an old, or even new, exposure of the fishbed on the BGS Outcrop will require excavation, perhaps on a large scale. This would potentially also help assign provenance to existing specimens. The Lethen Bar SW sector, centred on Braevail Farm, is now mostly cultivated land and, thus, there may be little probability that any work would be possible in this area.

In the NE sector, the reopening of Clune Quarry would presumably be precluded owing to the risk of contamination of the well providing Easter Clune Farm's household water supply. Rough pasture and uncultivated land is cut by streams around Chapel Hill and Easter Clune Farm and may reward the most determined investigation. Horne (1878) described a BGS test pit where nodules in clay were seen (Locality D in Fig. 10) but well-preserved fishes have not been recorded from the quarries in this area and thus they may only have been sources of locally commercial quicklime.

‘Poorly preserved’ fish-bearing nodule material has recently been recovered by Wood, suggesting that perseverance and major effort may yield better material. Site monitoring may allow opportunities that arise elsewhere (e.g. extensive ploughing, drainage work and/or the sinking of wells) to be exploited.

The anomalous accounts of Lethen Bar being the site of a continuous fishbed around the perimeter of a hill appear to emanate from erroneous early maps by Arrowsmith (1807) and MacCulloch (1836), which heavily influenced John Grant Malcolmson. His geological sections (Malcolmson 1859) presented the geology and topography of the Lethen Bar area somewhat naively. Despite Horne (1923) offering corrections, subsequent workers were consequently misled, especially Andrews (1983), who seemed to accept Malcolmson's (incorrect) findings at face value.

Figure 6a and b shows the correct geological, geographical and topographical relationships of the relevant landmarks, allowing Malcolmson's sections to be seen in context.

As well as reaffirming the surviving Lethen House Quarry, this work has provided new information on the positions of four old fishbed localities from field evidence. The authors have also provided detailed locality information for the sites of previous quarries, and a newly appreciated probable natural exposure, and for a pair of newly recognized pits adjacent to the BGS Outcrop. Evidence from old lime kilns has been uniquely implemented to identify old quarry localities and presents the possibility that at least one, possibly two, additional quarries existed on the west flank of the fishbed.

There is little or no evidence in the literature or in the field to support Andrews’ (1983) hypotheses that Lethen Bar was a continuously moving enterprise following the outcrop on the outlier and the probability that the entire fishbed was thus quarried out. Conversely, the evidence indicates that there was no connection between sites and that all were worked in isolation, and that some of these were active over many years and could be reactivated today given the appropriate permissions.

The evidence shows that only two of the seven quarries, pits and exposure known produced fossils with the finest preservation, namely old Lethen Bar and Clune Quarries, on the eastern margin of the outlying BGS Outcrop. It appears that the presence of limestone bands, possibly transiently and documented only at these localities, may be associated with enhanced preservation.

A new stratigraphic analysis of the Lethen House Quarry correlates extremely well with historical descriptions of the nodule-bearing laminites, and their stratigraphy, from the quarries previously described on the BGS Outcrop. The Lethen House Quarry outcrop is thus representative of the general stratigraphy of the Lethen and Clune area and, as the only surviving reference point for the Lethen fishbed, is pivotal to our understanding of the fishbed stratigraphy of the area.

The authors wish to thank R. Hoskin (Lethen Estate), J. Fraser (Easter Clune Farm) and P. Smart (Braevail Farm) for access permission, and I. Stirling (Lethen Estate) for on-site assistance. We are grateful to R. Paton (NMS retired) and R. Thomas (Nairn Museum) for access to material in their care. We also wish to thank M. Taylor (NMS retired), G. Berkenheger and F. Thom for valuable discussion, D. Stewart for helpful information, L. Albornoz-Parra (BGS) for database searches, C. Lam (GSL) for assistance, C. Auton (BGS retired) for information, and D. and N. Longstaff for figure preparation. R.G.D. also wishes to thank the Trewin family for the donation of material from the late Nigel Trewin's estate. We thank the editor, Y. Candela (NMS) and reviewers A. Wright (Elgin Museum) and M. Newman for their constructive comments and suggestions that improved the manuscript.

RGD: conceptualization (supporting), data curation (supporting), investigation (lead), methodology (lead), writing – original draft (lead), writing – review & editing (equal); NHT: conceptualization (lead), data curation (lead), investigation (lead); JA: conceptualization (supporting), writing – review & editing (equal); SRW: conceptualization (supporting), writing – review & editing (equal)

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License (http://creativecommons.org/licenses/by/4.0/)