Reconciling Divergent Ages for the Oldest Recorded Air-Breathing Land Animal, the Millipede, Pneumodesmus newmani Wilson & Anderson, 2004: A Review of the Geology and Ages of the Basal Old Red Sandstone Stonehaven Group (Silurian–Early Devonian), Aberdeenshire, Scotland

Abstract

Divergent mid-Silurian (late Wenlock) and latest Silurian–earliest Devonian (Pridoli–Lochkovian) ages have been proposed for the strata bearing the millipede Pneumodesmus newmani, the oldest known undoubted air-breathing land animal, marking a significant event in the evolution of the first land biota. The late Wenlock age is based on physically correlating a non-marine section dated with spores and detrital zircon dates, across a fault, with a distinctly different non-marine section bearing the millipede. The Pridoli–Lochkovian ages are based on vertebrate fossils and detrital zircon dates from this latter section. A review of the available lithostratigraphic, petrological, and biostratigraphic data, plus detrital zircon dating of the basal Old Red Sandstone around Stonehaven, shows that the two dated sections have strata that are of different depositional environments, sources, and ages and that the totality of the evidence firmly dates the millipede-bearing strata as Pridoli–Lochkovian.

1. Introduction

The oldest recorded air-breathing land animal, the millipede, Pneumodesmus newmani Wilson & Anderson, 2004, is important in understanding the evolution of the earliest terrestrial ecosystems [1,2,3,4]. A single specimen was found by Mike Newman in 2001 in the Cowie Harbour fish bed (or Dictyocaris bed as fish are scarce), just east of Cowie, Aberdeenshire, Scotland. The age of the fish bed is controversial. Using the current Silurian and Devonian time scales of Melchin et al. [5] and Becker et al. [6], the fish bed was initially estimated as Downtonian (latest Silurian, Pridoli) based on the fish fauna [7]. Floral evidence from isolated inland exposures later pointed to a late Wenlock–early Ludlow (mid-Silurian) age [8,9]. LA-ICP-MS U/Pb youngest single-grain detrital zircon dates of 414.3 ± 7.1 and 413.7 ± 4.1 (2σ) Ma, from strata bounding the fish bed, re-established the Pridoli–Lochkovian (latest Silurian–early Devonian) age [10], but was followed by a return to a Wenlockian age again, based on further floral evidence and a youngest detrital zircon LA-ICP-MS U-Pb date of 430 ± 6 (1σ) Ma from adjacent, but fault-separated, exposures [4].

In fact, the two divergent interpretations of the age of the Cowie fish bed are easily reconciled by considering the locations of the materials analyzed, which come from separate fault-bounded structural blocks with different lithostratigraphy and fossils. The Pridoli–Lochkovian age comes from strata enclosing the Cowie Harbour fish bed, while the late Wenlock age comes from sediments in a different structural block to the north.

In order to place the earliest recorded air-breathing land animal in context and to establish its environmental setting and age, we review, in detail, the stratigraphy, sedimentology, paleontology, paleoenvironments, radiometric dating, and structure of the late Silurian to early Devonian Stonehaven Group, which hosts the Cowie fish bed. The geology of the Lower Old Red Sandstone around Stonehaven (including the Stonehaven Group) is summarized in Gillen and Trewin [11], McKellar [12], and McKellar and Hartley [13].

2. Structural Units

Around Stonehaven, rock exposures adequate for the determination of stratigraphic and structural relationships are only found in coastal exposures [11,14,15,16]. Away from the coast, it is difficult to correlate strata across separate structural blocks because of the very limited bedrock exposure, with over 90% of Quaternary cover and many faults (Figure 1). Each exposed area thus needs to be examined separately and in detail.

2.1. Inland Exposures

Evidence from the inland exposures can be easily dismissed as they show only diverse small and isolated outcrops in numerous fault splays along the Highland Boundary fault, juxtaposing diverse rock units [17,18]. Apart from the two spore-bearing siltstone exposures (AS12, AS16) [8,9], the strata were stratigraphically assigned purely based on lithology (Figure 1).

The Cowie Harbour Conglomerate Member is mapped in extremely small isolated and faulted exposures to the N and NW, while the inland spore- and plant-bearing localities come from small isolated exposures in and near the Carron water south and west of Tewil farm [8,9] (Figure 1). The large supposed Castle Harbour siltstone exposure mapped by the British Geological Survey (BGS) lies one kilometre west of the spore-bearing localities along the south bank of the Carron water (Figure 1). During a visit in October 2024, only fine-grained greenish-grey and purplish-grey sandstones with no interbedded siltstones were poorly exposed in the overgrown left bank meander cliff, with no exposures in the stream bed and right bank.

2.2. Coastal Exposures

On the Cowie foreshore, two distinct areas with differing stratigraphic sections are separated by a postulated tear fault through Cowie Harbour, which has been inferred for over 100 years based on the stratigraphic differences across it [14] (Figure 2 inset) (see also Shillito and Davies, Figure 9 [19], and Wellman et al., Figure 3 [4]).

The northern block has a steep NNW dipping overturned non-marine fluviatile and lacustrine section with supposed late Wenlock spore assemblages and a maximum youngest detrital zircon Llandoverian date of 439 ± 4 Ma from its lowermost beds, as noted by [4]. It is separated from the southern block by the inferred tear fault along Cowie Harbour.

The southern block has a steep NW dipping section, with a divergent strike and dip to the northern block, and a different section with the Cowie Harbour Conglomerate and the Cowie Harbour fish bed (or Dictyocaris bed), containing Pridoli–Lochkovian vertebrates and maximum youngest detrital zircon Lochkovian–Pragian date of 413.7 ± 4.1 Ma [10]. The Cowie fish bed occurs nowhere else, even where apparently similar enclosing strata are mapped in scattered inland exposures (Figure 1).

To the south of the southern Cowie Harbour outcrop, a covered area of beach sand 500 metres wide, with an inferred fault, separates it from a foreshore outcrop attributed to the Cowie Formation, with another nearby foreshore outcrops just south around Stonehaven Harbour, with the type section of the Carron Sandstone Formation (Figure 1, section E).

The detailed petrology of many of the Old Red Sandstones was reported by McKellar (her Appendix 2) [12], but the locations were not plotted on a detailed geological map. We have plotted the location and constituent values of samples relevant to this on Supplementary Figure S1 and Supplementary Table S1. In some units, very few analyses are available, and the results may be unrepresentative, but they are, however, compatible with those reported by Phillips [20].

3. Descriptions and Interpretations of the Northern and Southern Blocks

3.1. Northern Block

3.1.1. Sediments

The northern block has thin basal pediment breccias resting unconformably on the Highland Border Group basalts (Figure 3) and contains angular clasts of reddened basalts and chert entirely derived from the underlying rocks (Figure 4A,B).

The breccias are overlain by about 90 metres of the Purple Sandstone Member, which consists of brownish to greenish, lithic volcaniclastic sandstones with minor interbedded red and grey siltstones and mudstones, with one thin interbedded andesite lava flow or sill at about +10 m. The lowermost beds are predominantly trough cross-bedded fine- to medium-grained sandstones, with minor parallel laminated sandstones and minor interbedded mudstones and siltstones, and with unidirectional paleocurrent flow to the ESE, with only a 90° spread. Such characteristics are typical of large, low-gradient Platte-type braided streams, far from bedrock sources, with variable flow, transporting medium-grained sand, with an abundance of linguoid bar and dune deposits (planar and trough crossbedding), but with no well-developed cyclicity [21,22].

Petrographically, the basal Purple Sandstone samples (#7–11) are lithic to sublithic sandstones, dominated by acid to intermediate volcanic grains, with variable metasandstone grains, but very little feldspar, very low plagioclase/K-feldspar ratios, very low Qm/Qp ratios, and no biotite (Figure 5, Supplementary Figure S1 and Supplementary Table S1). Sources are thus dominantly acid to intermediate volcanics to the west-northwest.

The higher beds starting just below the andesite are tabular and trough cross-bedded fine- to coarse-grained lithic to sublithic sandstones with abundant often burrowed mudstones and siltstones (Figure 3 and Figure 4C,D). The fining-upward cycles indicate lateral accretion deposits in a lowland channelized river system with extensive floodplain deposits, while trace fossils and fragmentary plant remains indicate a humid climate [23,24,25]. Such characteristics are found in Donjek-type wet climate braided streams, distinguished by fining-upward cycles caused by lateral point-bar accretion or vertical channel aggradation, with longitudinal and linguoid bar deposits, channel-floor dune deposits, and bar-top and fine overbank deposits [22]. Cycles are commonly less than 3 m thick, as in the Purple Sandstone cycles. Though dominantly WNW, the almost 360° spread in paleocurrents [11,16] (Figure 3), however, suggests an intricate meandering channel pattern analogous to multi-channel (anastomosing) rivers on alluvial plains forming under low energy conditions near local base levels [26]. The distinction between sinuous, single-channel “meandering” rivers and straighter, multi-channel “braided” rivers is not clear-cut [27] and may be impossible in the absence of a complex vegetation cover, though the late Silurian–early Devonian may mark the onset of small meandering channels [28].

The overlying Castle of Cowie Member consists of about 74 metres of red medium-grained tabular, trough cross-bedded and planar bedded sandstones, and pebbly sandstones with intraformational red siltstone clasts, interbedded with very minor red siltstones (Figure 4E,F and Figure 6). The lower trough cross-bedded units form in migrating channels, while the dominantly tabular cross beds above indicate downstream-migrating sand flats [23,27]. Minor trough and parallel laminated sands and mudstones/siltstones in fining-upward cycles show relatively uniform southwesterly paleocurrents [11,16].

The absence of paleosols, partly attributable to the lack of an integrated plant cover anywhere outside water-saturated environments [28,29,30], and red siltstones are comparable with the semi-arid unconfined chaotic, vertical accretion braided stream networks of Central Australia with overbank fines accumulating during major floods [31,32,33,34]. Some beds, of uncertain stratigraphic position, have a low diversity trace fossil assemblage assigned to the Pridoli–Lochkovian [17] (Figure 6B).

The Castle of Cowie sandstones (#5, 12, 13) are slightly more feldspathic sublithic sandstones than the basal beds with greater amounts of metasandstone grains, very low plagioclase/K-feldspar ratios, variable Qm/Qp ratios, and biotite (Figure 5). The slight differences indicate somewhat lower chemical weathering due to drier conditions in the Castle of Cowie sandstones.

The overlying clastic sediments, up to 1800 m thick, are a combination of dominant tabular and trough cross-bedded lithic, occasionally pebbly, sandstones with abundant mudstone clasts (Figure 7). These higher sediments are only exposed during low tides and have not had a complete, detailed section measured. Their brownish to greenish colours indicate a return to more humid conditions similar to, but not as wet as, the basal beds with lake deposits, in unconfined chaotic, vertical accretion braided stream networks, marking a return to Platte-type braided stream deposition. Petrographically, the upper sandstones (#2–4) are more lithic sandstones dominated by acid to intermediate volcanic grains, variable but mostly few metasandstone grains, with moderate plagioclase/K-feldspar ratios, low Qm/Qp ratios, and relatively high plagioclase and biotite contents, indicating less chemical weathering (Figure 5).

3.1.2. Fossils

The low diversity trace fossil assemblage from the Castle of Cowie Member consists of Arenicolites and Taenidium, attributed to burrowing arthropods, and indeterminate bioturbation. Though limited, it is, nevertheless, more diverse than that from other continental deposits of middle Silurian age and shares greater similarity with worldwide Pridoli to Devonian-aged ichnofaunas [17].

Land plant spores were found in two horizons in the northern block: i) a grey sandy siltstone (NO88290/87036) from the lowermost mudstones and siltstones; and ii) a large cobble (20 cm in diameter) of dark grey siltstone within a sandstone at the base of the Castle of Cowie Member (NO88420/87136) [4]. The latter only dates the cobble, not the sandstone. Its large size in fine-grained sandstone, without other intraclasts or extraclasts, indicates that it was very locally derived since such a large siltstone clast is hydrodynamically incompatible with the enclosing sandstones, whether it was near-contemporaneous with the sandstone or older. The fact that it survived transport suggests that it was at least partly indurated. Individual taxa lists for the two samples were not published separately, presumably on the assumption that they belong to the same floras and no taxonomic ranges were given for individual species [4].

Range charts are the critical limiting factor for biostratigraphic resolution, which requires detailed local range charts that resolve the first and last appearances of fossil species [35]. However, using the individual species ranges from White’s [36] palynodata compilation (current to 2006) and accepting only those species ranges in marine strata that can be correlated with the standard stages give the ranges of the individual species in the inland (A1) and northern block (A2), as shown on Figure 8.

These do not define a Wenlockian assemblage: 16 out of 18 overlap in the Pridoli. Dyadospora murusdensa and Tetrahedraletes medinsis range only up to the end Wenlock and end Ludlow, respectively, but could be reworked older fossils [37]. Hispanaediscus lamonitii is found in both the Ludlovian [38] and Pridoli [39], but not in between; similarly, Artemopyra sp. A Burgess and Richardson 1991, is found in the upper Wenlock [40] and lower Pridoli [41] of the Welsh borders. Hispanaediscus wenlockensis is found not only in the upper Wenlock and Ludlow [39] but also in the lower Pridoli [42]. The standard biostratigraphic overlapping range method firmly indicates a Pridoli age for both assemblages, and none of the species are exclusively Wenlock.

3.1.3. Detrital Zircon U/Pb Dates

The detrital zircon laser-ablation sector-field-inductively coupled plasma mass spectrometry (LA–SF–ICP–MS) U-Pb dates cited by Wellman et al. [4] come from McKellar [12] and McKellar et al. [43] with 2σ errors. Using McKellar’s [12] grid reference locations, the analyzed zircons came from three samples from the lower beds in the northern block, and one from the Carron Formation in the Stonehaven Harbour section, far to the south [43] (their Figure 14). Unfortunately, only probability density plots for lumped U-Pb data. are given in McKellar [12] and McKellar et al. [43], without the detailed analytical data for the individual zircons (e.g., % discordances, raw U and Pb values) required to adequately evaluate the results obtained [44]. Such details are necessary because ”few, if any, scientific disciplines publish numerical data that are accepted by nonexperts and propagated through the literature as extensively as geochronology”, and “quantifying random and systematic sources of instrumental and geological uncertainty is vital, and requires transparency in methodology, data reduction, and reporting” [45].

Assuming a conventional 10% discordance cut-off was used for acceptable dates, the youngest zircons in the four McKellar et al. [43] samples (with >100 grains per sample), in supposed stratigraphic order, are 439 ± 4, 478 ± 4, 470 ± 7 (basal “Cowie”), and 430 ± 6 Ma (Carron Sandstone Formation. The 439 ± 4 Ma (Llandoverian) date is from the lowermost northern block sediments (petrographic sample #11), while the 430 ± 6 Ma (Wenlock) date is from the isolated type Carron Sandstone outcrop in Stonehaven Harbour (petrographic samples #28) (Supplementary Table S1). The two older dates in the strata in between indicate that no near-contemporary igneous rocks were being sourced at that time.

Detrital zircon dates give only a maximum, not necessarily true, depositional age and need careful evaluation [46] and an understanding of various U/Pb methods and their significance [45,47]. Thus, a sediment cannot be any older than the youngest thing in it, though it can be younger [48]. The Llandoverian detrital zircon date from the basal northern block sediments is compatible with both the Wenlock and Pridoli alternative ages from the spores, though we consider the Pridoli age more accurate.

3.1.4. Summary

The northern block section therefore shows basal pediment breccias overlain by SE-flowing low-gradient braided stream deposits passing up into anastomosing streams under a wetter climate of Pridoli age, succeeded by SW-flowing braided streams of Pridoli–Lochkovian age with variable wetter and drier conditions: all of which have dominantly distant acid volcanic sources. There is no sign of any topographic effect on sedimentation by the adjacent Highland Boundary fault [12].

3.2. Southern Block

3.2.1. Sediments

The southern block has a lower sandstone-dominated section, overlain by the Cowie Harbour Conglomerate Member with volcanic clasts, and the Cowie Harbour Siltstone Member with a fish bed (Figure 1 and Figure 9). These are overlain by sandstones correlated with the Carron Sandstone Formation, whose type section is in Stonehaven Harbour, 2 km to the south across faulted and unexposed terrain. However, the petrology of these two sections is very different, and they cannot be assigned to the same unit (Figure 5).

The lower beds are similar in character and may belong to the same stratigraphic unit as those at the top of the northern section but are displaced westwards by right-lateral movement on the Cowie Harbour fault; this displacement is also indicated by the small right-lateral shears next to the fault in the southern block [15] (Figure 2 inset). Two of three basal sandstones (#1, 14, 15) are extreme volcaniclastic lithic sandstones, with no metamorphic or sedimentary rock grains, moderate Plagioclase/K-feldspar ratios, high Qm/Qp ratios, and relatively high plagioclase and variable biotite contents, indicating predominantly physical weathering of an acid to intermediate volcanic source. They are distinct from any other samples in the succession and must be very locally derived (Figure 5, Supplementary Table S1).

The overlying Cowie Harbour Conglomerate Member marks a major change in sedimentation, forming the base of a thick fining-upward fluvial cycle culminating in the lake sediments of the Cowie Siltstone Member (Figure 9). The conglomerate consists of two thick beds of conglomerate with rounded clasts of andesite and rhyolite, mostly around 10 cm in diameter but with some of boulder size [11], 1986), supplied by local outcrops to predominantly well-sorted conglomerates with more rounded, more distantly travelled pebbles. These are channel deposits of large, permanently flowing trunk streams [22]. The overlying siltstone unit probably represents the filling of these channels, inactive after avulsion and channel migration.

The green tuff bed is a volcanic microconglomerate in which sand- to pebble-sized green tuff fragments float in a green tuff matrix. It marks an apparently isolated volcanic event.

The Red Sandstone member above consists of parallel laminated and tabular to trough cross-bedded fine-to medium-grained sandstones, with paleoflow towards the southwest, like the underlying beds. These relatively thin, unchannelized sandstones alternate with laminated siltstones, resembling the flood-dominated overbank deposits of Central Australia [34]

The Cowie Harbour Siltstone Member consists of tabular to ripple drift cross-laminated fine- to medium-grained sandstones interbedded and passed up into laminated siltstones and mudstones, which bear the Cowie Harbour fish bed (Figure 7). This latter fine-grained unit is thicker than any other such unit in the Stonehaven Group. It marks a relatively long-lived lake depositional environment, as shown by the scattered plant remains in the thicker siltstones and mudstones and the unique fish bed [14].

The entire succession from the Cowie conglomerates to the siltstones marks a rising base level in which successive channel and overbank deposits of a fan delta culminate in the development of a permanent lake [22].

The overlying “Carron Sandstone formation” in this block consists of a thick sequence of red, fine- to coarse-grained, trough cross-bedded and horizontally laminated, locally pebbly lithic sandstone, with lenses of conglomerate, analogous to the type of Carron Sandstone Formation at Stonehaven Harbour to the south, but with distinct though overlapping petrologies. This “Carron Sandstone Formation” has quartzo-feldspathic sublithic sandstones, with moderate plagioclase/K-feldspar ratios, moderate Qm/Qp ratios (~2.5), relatively high plagioclase (>14% of grains), and moderate biotite contents (Figure 5, Supplementary Table S1), indicating greater chemical weathering and/or recycling of sediments.

At Stonehaven Harbour, the type Carron Formation consists of red fine- to very coarse-grained trough festoon cross-bedded and tabular and horizontally laminated lithic pebbly sandstone and sandstone, with thin quartzite pebble beds and fine- to medium-grained planar laminated micaceous sandstone [13] (Figure 10). These river channel deposits record dune migration at the base of channels and the downstream migration of transverse bars in an accreting semi-arid to arid braided fluvial system [22,49,50,51,52].

The type Carron Sandstones is petrologically distinct from the “Carron Sandstones” in the southern block, being volcaniclastic lithic sandstones, with low to high plagioclase/K-feldspar ratios, very variable Qm/Qp ratios, moderate to high plagioclase (>14% of grains), and very variable biotite contents (Figure 5, Supplementary Table S1). This limited sandstone petrography indicates distinct sources for the Cowie “Carron” sandstones and the type Carron Sandstones.

Since the paleocurrents in both sections indicate a dominant palaeoflow direction towards the west to southwest (Hartley and Leleu, 2015 [16]; McKellar, 2017) (Figure 6, Figure 7 and Figure 9), possibly different sources were sequentially provided by strike-slip faulting to the east during sedimentation [53]. Though the two sections are not correlatable, the overlap of sublithic sandstone compositions suggests that the type Carron Sandstones may be a higher sedimentary strata of the southern block “Carron Sandstones”.

3.2.2. Fossils

Fossils occur only in the Cowie Harbour Siltstone Member, with identifiable taxa only in the fish bed. Curiously, only indeterminate plant remains and no spore fossils are recorded, even from the apparently suitable Cowie Harbour fish bed [8,9,14].

The fauna of the fish bed has, however, been studied for many years. Initially considered Downtonian (Pridoli), based on comparisons with fish/arthropod biotas elsewhere [7,14], Wellman et al. [4] assigned it to the mid-Silurian, based on correlation with spore-bearing sediments in the northern block. In view of these divergent interpretations, a detailed taxonomic evaluation and comparison are required.

All the fish taxa in the Cowie Harbour fish bed range into the Lower Devonian and are not exclusively Silurian. Hemiteleaspis heintzi Westoll 1945 is now regarded as only an indeterminable species of Hemicyclaspis [54,55,56] which is a Pridoli to Devonian form [57]. Traquairaspis campbelli Traquair 1912 is found only at Cowie, where a single specimen of a dorsal plate defined the genus [58]. T. campbelli is one of only two species in the Traquairaspidae, as Tarrant [59] removed all the Welsh Border species to other genera. The second Traquairaspis species is found in Pridoli sediments of northern Canada, Poland and Russia [60,61,62]. The birkeniid anapsid, Cowielepis richiei Blom 2008, is confined to the Cowie fish bed, and Blom attributed it to the mid-Silurian only because the Cowie fish bed was then so assigned [63]. Birkenids range into the Lochkovian and are not exclusively Silurian [64]. The vertebrate fauna cannot therefore be assigned to any marine-based biostratigraphic unit. The agnathan taxa have non-overlapping ranges from Ludfordian to Lochkovian and cannot therefore be used for a vertebrate biostratigraphy applicable to the Cowie assemblage [65].

The arthropod, Nanahughmilleria norvegica Kiaer 1911, was assigned by Kiaer [66] to the Downtonian or uppermost Ludfordian [67], but other Nanahughmilleria species range into the Devonian [68]. The original record of N. norvegica comes from the Ludfordian lower Sundvollen Formation [65,66]. Dictyocaris is the most abundant fossil in the fish bed (also called the Dictyocaris bed), and is a long-ranging fossil found in many suitable facies from Llandoverian to Pridoli [69]. Dictyocaris occurs only as abundant, frequently large, thin carbonized fragments in strata bearing well-preserved articulated eurypterids. Its apparent originally cone-shaped morphology was first compared with the living liverwort Marchantia but was subsequently attributed to a cephalaspid fish or an arthropod [14,69]: it is more likely to be a plant than a fish or an arthropod [70,71], but is still enigmatic [72] (see discussion in Brookfield [73]. The three fossil millipedes, Cowiedesmus eroticopodus Wilson & Anderson, Albadesmus almondi Wilson & Anderson, and Pneumodesmus newmani Wilson & Anderson, have no stratigraphic value, except that the assemblage is more diverse than any other Silurian occurrence, including the single millipede species in the Kerrera fish bed in Argyll, Kampecaris obanensis, dated 425.5 ± 4.5 Ma (Gorstonian) [2,73].

None of the non-marine fossils in the Cowie Harbour fish bed can, so far, be assigned to a specific Silurian to Devonian stage in the standard marine-based geological time scale.

3.2.3. Detrital Zircon U/Pb Dates

The detrital zircon dates in the southern section with the fish bed are discussed in detail in Suarez et al. [10]. They are carefully evaluated U-Pb zircon LA-ICP-MS dates, with standard 2σ errors, of 413.7 ± 4.1 Ma (% discordance 0.72) from the green tuff below, and 414.3 ± 7.1 Ma (% discordance 0.65) from a sandstone above the fish bed [10]. These are identical within error and consistent with a maximum Pridoli–Lochkovian age assigned. Attempts to make these dates older by appealing to lead loss are contradicted by the low % discordance, which, in any case, would make them only slightly older [74]. Thus, assuming all the measured ²⁰⁷Pb is common, and that the concordia discordance is due to simple Pb loss, then the dates become 413.6 ± 4.4 Ma and 414.0 ± 7.3 Ma, which are Lochkovian.

3.2.4. Summary

The southern block strata show an upward change from semi-arid westward-flowing braided streams with, unlike the northern block, beds of locally derived basement clasts (indicating volcanics exposed by contemporary faulting), passing up into thicker wetter climate anastomosing streams with marsh and lake beds, followed by coarse locally derived semi-arid braided stream deposits.

4. Conclusions

The northern and southern block strata at Cowie show distinctly different sedimentary successions with fossils, indicating that both were deposited in Pridoli to Lochkovian times. The sedimentology suggests that there were two relatively wet periods represented by brownish to greenish anastomosing channels and overbank deposits, with nearby vascular plant communities, during the Pridoli–Lochkovian, separated in time by drier braided stream deposits.

The late Wenlock age inferred by Wellmann et al. [4] for the Cowie Harbour fish bed can only apply to the lowermost strata of the northern structural block and some disconnected tectonically isolated exposures along the Highland Boundary fault, kilometres away across unexposed terrain, and these may in fact be Pridoli. The Pridoli–Lochkovian age for the Cowie Harbour fish bed is based on vertebrate faunas and maximum detrital zircon dates from the actual strata bearing the fish bed. The various lines of evidence, lithostratigraphic, sedimentological, biostratigraphic, and radiometric, clearly indicate that the Cowie fish bed is Pridoli to possibly early Lochkovian. Using spores or detrital zircon ages to suggest a Wenlock age for the Cowie Harbour fish bed only works by accepting a spurious lithological correlation of the southern block strata with the dissimilar spore-bearing strata to the north of the Cowie Harbour fault and disregarding the known ranges of both the vertebrate fossils and the spore species.

Lastly, if the basal Cowie sediments of the northern block are Pridoli, then the Old Red Sandstone continental sedimentation began in the Pridoli and not in the Wenlock in the northern Midland valley and is contemporary with the basal Old Red Sandstone in the Welsh Borders [75]. This apparently synchronous onset of lowland desert conditions in the United Kingdon, and the onset of more permanent wetter conditions of the fish bed strata, may reflect not only a rise in bass level during a Pridoli marine transgression but also the moderate to rapid warming and development of supergreenhouse conditions during the Přídolí in northern Gondwanaland [76].

5. Future Work

More precise dates for the various Stonehaven units (and beyond) could be obtained by TIMS U-Pb analysis of the youngest of the already LA-ICP-MS dated near-concordant zircons. An objective procedure would be to analyze the youngest concordant zircon grains from many (cheap) imprecise LA-ICP-MS dated ones, with (expensive) TIMS methods. As demonstrated by Garza et al. [47] and Howard et al. [77], while LA-ICP-MS provides a useful preliminary approach for dating, its accuracy and precision are generally lower compared to other methods. Specifically, LA-ICP-MS often yields younger zircon U-Pb dates than those obtained with CA-ID-TIMS, primarily due to undetectable cryptic Pb loss and systematic biases such as matrix mismatch effects. These issues highlight the necessity for future analyses using the more precise CA-ID-TIMS technique. We do not have TIMS facilities in Austin, but if anyone would like to use this method on our dated zircons, we would be happy to provide them. Also, more studies trying to establish non-marine spore ranges from occurrences in marine strata are needed [78]