Squamate fauna from the Denver Basin shows major ecosystem disruption across K/Pg boundary

Abstract

The Cretaceous/Palaeogene (K/Pg) boundary marks the most recent and the second-most severe mass extinction in Earth’s history. Marine K/Pg boundary sections that preserve uninterrupted sedimentation are abundant, and thus the global marine response to this extinction event is well understood. Similar terrestrial sedimentary sequences that record the before and after of the extinction and that are also geochronologically well constrained are rare. We describe and compare the non-marine squamate diversity immediately before (ca 57 kyr) and after (ca 128 kyr) the K/Pg mass extinction from two geographically and chronostratigraphically well constrained localities in the Denver Basin, Colorado. The latest Cretaceous squamate record there consists of 235 isolated fossils and is extraordinarily diverse, with 27 operational taxonomic units, whereas the earliest Palaeocene record is depauperate in both richness (3 taxa) and abundance (20 isolated fossils)—a ca 91% drop in specimen abundance and a 93% extinction rate. Integrating the Denver Basin squamate fossil record with that of the Williston and Powder River basins of Montana, North and South Dakota, and Wyoming, provides a regional extinction rate of between 75 and 84% for squamates in the Western Interior of North America and suggests this group was more severely impacted across the K/Pg boundary than other small-bodied vertebrates.

1. Background

The Cretaceous/Palaeogene (K/Pg) mass extinction is the most recent of the ‘Big Five’ mass extinctions [1] and was triggered by the impact of a massive asteroid off the Yucatán Peninsula [2]. During the K/Pg mass extinction (ca 66 million years ago), ca 75% of all species went extinct, including all non-avian dinosaurs (e.g. [2–4]). This event completely changed the trajectory of the evolutionary tree of life (e.g. [5,6]) and resulted in a total rebuilding of terrestrial ecosystems from dinosaur- to mammal-dominated communities. The effects of the mass extinction on various marine invertebrate clades are well documented from several different marine environments [3,7–9]. There are fewer terrestrial fossil records that preserve the primary markers (e.g. shocked minerals, spherules, iridium anomaly) for the asteroid impact, as well as a robust fossil record on either side of the K/Pg boundary, and thus comparably detailed and geographically dispersed records of terrestrial ecosystems do not exist.

Most studies on the impact of the K/Pg mass extinction on land come from the Williston Basin in North America (see the Geological Society of America (GSA) volumes dedicated to this basin for a summary [10,11]), which is one of the few basins that preserves abundant fossils on either side of the K/Pg boundary denoted by primary markers (see also [12] for a K/Pg study on the Hanna Basin, Wyoming). Numerous detailed studies have documented the severity and selectivity of the K/Pg mass extinction in plants [13–16], fish [17,18], amphibians [19], mammals [20–24], turtles [25–27], birds [28,29], crocodilians [30] and lizards and snakes [31,32] from the Williston Basin. These studies have shown the severity of the extinction varies by clade—some groups underwent very little extinction (e.g. crocodilians and turtles), while most other groups show high rates of extinction (e.g. mammals, birds, squamates) across the K/Pg boundary.

While the abundant fossil record on either side of the K/Pg boundary in the Williston Basin allows robust estimations of extinction and selectivity, a confounding aspect in these estimates is a formational change between the Cretaceous-aged Hell Creek Formation and the largely Palaeocene-aged Fort Union Formation coincident or nearly coincident with the K/Pg mass extinction [15,33,34]. The formation change represents a distinct change in facies and inferred palaeoenvironment. The Hell Creek Formation is dominated by fine-grained, immature, hydromorphic palaeosols, and single-storied fine- to medium-grained fluvial sandstones and lignites are rare. The overlying Fort Union Formation is characterized by abundant carbonaceous shales, lignites, coals and variegated beds, representative of ponded-water environments, and large multi-storied fluvial sandstones [34–36]. This change in facies and formations has been attributed to a base-level rise of the Western Interior Seaway and a shift from small meandering rivers and associated floodplains in the Hell Creek Formation to a ponded-water and marshy environment in the Fort Union Formation [14,15,34,35]. It is unclear what impact the palaeoenvironmental shift in the Williston Basin has on K/Pg mass extinction estimates. However, it is clear that some fauna [25,37] and flora [14] are preferentially found in certain facies and the facies changes abruptly across the K/Pg, which combined suggests the palaeoenvironmental shift will impact K/Pg mass extinction estimates for some clades. Furthermore, certain lineages like the turtle Hoplochelys disappear at the K/Pg boundary in the Williston Basin but are common in Palaeocene sediments in the Denver (Colorado) and San Juan (New Mexico) basins [38,39] and it is impossible to determine if this is a true localized extinction (extirpation) in the Williston Basin or a result of the shift in facies/palaeoenvironments that is coincident with the K/Pg boundary.

In order to provide a more robust extinction and selectivity estimate for terrestrial plants and animals across the K/Pg boundary, field areas that represent geographical and environmental settings different from that of the Williston Basin need to be developed. Here, we describe a new squamate fauna from two localities that closely straddle the K/Pg boundary identified by primary markers (e.g. shocked minerals, iridium anomaly) in the West Bijou study area within the Denver Basin.

(a). Geological setting/background

The West Bijou study area is located 60 km east of Denver, Colorado, USA, along the eastern edge of the Denver Basin (figure 1a,b) and is ca 800 km south of the better-studied Williston Basin (figure 1a). Whereas the sediments of the Williston Basin were deposited distal from the mountains (ca 450 km to the Sevier Orogenic Belt front) and proximal to the Western Interior Seaway (a few to tens of kilometres), those of the West Bijou study area were deposited ca 80 km from the Laramide Orogenic Belt front and distal (hundreds of kilometres) from the Western Interior Seaway. The Denver Formation exposed in the West Bijou study area is dominated by lignites interspersed with fine-grained sandstones, and drab mudstones (figure 1c), which reflect a low-energy fluvial depositional environment where perennially high water tables contributed to the formation of swamps, lakes and hydromorphic soils [42,43]. The study area has a well preserved K/Pg boundary section (figure 1c,f) with primary indicators of the asteroid impact, including shocked minerals, iridium and a boundary clay [41,44], and there is no change in facies across the K/Pg boundary (figure 1c; [43,44]). The magnetochron [41] and palynostratigraphic biozone [45,46] boundaries are well resolved in the West Bijou study area. The West Bijou study area is remarkable in the number of ash beds preserved in lignite (often 10+ ash beds per metre of lignite; figure 1f). Many of these beds contain volcanic zircons that have been dated, including those closely associated with the K/Pg and magnetochron boundaries (figure 1f; [41]). In addition to abundant vertebrate [40,47,48] and plant [43,49,50] fossils that span the K/Pg boundary, organic biomolecules are also well preserved in the West Bijou study area and document a long-term warming and transient cooling ‘blip’ at the very end of the Cretaceous (ca 30 000 yr or ca 20 cm prior to the K/Pg boundary) that has been attributed to Deccan Trap volcanism in India [51].

Figure 1.

Geographical and geological context of the West Bijou study area within the Denver Basin. (a) Map of the sedimentary basins mentioned in text. (b) Map showing the position of West Bijou along the eastern margin of the Denver Basin. (c) The simplified stratigraphic section shows the position of DMNH Loc.2560 and DMNH Loc.3648 relative to the K/Pg boundary which is identified by several primary markers (e.g. iridium anomaly, shocked minerals [40]). ‘A’ denotes volcanic ash layer; letters below the right column represent the following grain sizes and lithologies: ‘C’ coarse sand, ‘M’ medium sand, ‘F’ fine sand, ‘V’ very fine sand, ‘S’ silt, ‘Cl’ clay, ‘L’ lignite. (d) Google Earth image showing the geographical position of each locality relative to the K/Pg boundary (distance between both localities is 425 m); the section of the K/Pg boundary documented in [41] is shown in (f). (e) Field photo of DMNH Loc.2560. (f) Photo of the K/Pg boundary at West Bijou highlighting the number of radiometrically dated ash beds above and below the K/Pg boundary. Scale bar shows 10 cm increments and ‘Ma’ = million years. (g) Field photo of DMNH Loc.3648; arrow pointing at fossiliferous layer. Map in b modified from Lyson et al. [39], stratigraphic section in c modified from [40] and part f modified from [41].

The present study focuses on two fossiliferous microvertebrate localities: the Cretaceous-aged DMNH Loc.3648 (figure 1g) contains abundant microvertebrate remains preserved in a fine-grained, buff-coloured sandstone unit, interpreted to represent a meandering fluvial channel facies, that is ca 4 m below the nearby K/Pg boundary. Using the Clyde et al. [41] age model and sediment accumulation rate of 142 yr cm⁻¹, this corresponds to ca 57 kyr prior to the K/Pg boundary (figure 1c; [41,51]). The locality has produced over 3500 isolated microvertebrate specimens, where the majority of those remains belong to gar fish and crocodilians with over 1000 specimens each and 235 squamates (table 1), as well as a Lancian (70.6−66 Ma) North American Land Mammal ‘Age’ (NALMA) fauna. In addition to abundant, disarticulated microvertebrate remains, the locality preserves disarticulated remains of a large ceratopsian dinosaur including two edentulous dentaries and several caudal vertebrae.

Table 1.

Distribution of microvertebrate remains from the Late Cretaceous Loc.3648 and the early Palaeocene Loc.2560 from the West Bijou study area. Percentages reflect portion of two datasets (with and without removal of the two most abundant clades). Squamates are highlighted in italic type.

clade	Loc.3648	(%)1	(%)2	Loc.2560	(%)1	(%)2
Chondrichthyes	60	1.65	5.35	115	0.68	8.73
Actinopterygii	1525	41.82	—	11 071	65.06	—
Amphibia	336	9.21	29.95	650	3.82	49.35
Squamata	235	6.45	21.00	20	0.12	1.52
Testudines	91	2.50	8.11	27	0.16	2.05
Crocodylia	1000	27.42	—	4628	27.20	—
Dinosauria—Ornithischia	207	5.68	18.45	—	—	—
Dinosauria—Theropoda	105	2.88	9.36	—	—	—
Mammalia	85	2.33	7.58	500	2.94	37.97

¹Percentage relative to total abundance data.

²Percentage after removing the two most common taxa (Actinopterygii and Crocodylia).

The Palaeocene-aged DMNH Loc.2560 (figure 1e) preserves abundant microvertebrate remains preserved in two stacked, fine-grained, buff-coloured sandstone units, interpreted to represent a meandering fluvial channel facies (see [40] for more details), that is 9 m above the nearby K/Pg boundary (figure 1c). Using the Clyde et al. [41] age model and sediment accumulation rate described above, this corresponds to ca 128 kyr post-K/Pg boundary (figure 1c; [40,41,51]). The locality has produced over 17 000 fossil vertebrate specimens (table 1), where the majority belong to gar fish (>8500) and crocodilians (>4600), as well as 20 squamates and a Puercan 1 NALMA (66−65.7 Ma) fauna [40]. The locality is ca 425 m to the west of DMNH Loc.3648 (figure 1d).

2. Material and methods

The two microsites were excavated using hand tools, then dry-sieved on site using inter-nested screens (mesh numbers 4 and 30, i.e. 4 and 30 openings per inch (2.57 cm)), and finally wet-sieved in inter-nested screens with mesh sizes 4, 18 and 30. DMNH Loc.2560 was excavated in 2014 and DMNH Loc.3648 in 2008−2010, 2013−2014, 2016, 2021 and 2024. Numerous volunteers and interns at the Denver Museum of Nature & Science (DMNS), as well as undergraduates at Colorado College and the University of Michigan, picked the microfossils from the screened matrix. Squamate fossils were imaged using a Keyence microscope, and the resulting images were processed in Adobe Photoshop and assembled into plates in Adobe Illustrator.

Fossil richness: diagnostic elements (mainly tooth-bearing elements; electronic supplementary material, table S1) were identified to the genus and species level using published references (see electronic supplementary material, document S1) and the existing collections at DMNS; previously identified specimens were checked against the same references. Photographs of the specimens were shown to colleagues with expertise in fossil squamates to assist with identifications. Elements that could not be identified to that level, either owing to lack of diagnostic features or because they may represent unnamed taxa, were organized into morphotypes at the lowest taxonomic level possible after consulting published references. Each unique genus-level identification was counted to produce minimum alpha diversity for each locality. Maximum alpha diversity was created by combining morphotypes and named taxa.

Richness and abundance data were used as a starting point, and three datasets were constructed for further analysis of taxonomic richness and abundance. Dataset 1 treats each unique morphotype as a new operational taxonomic unit (OTU) and represents maximum alpha diversity. The second dataset does not include unique morphotypes and only uses previously described taxa (minimum alpha diversity). The third dataset only uses higher taxonomic units (e.g. Scincoidea). All three cases underwent individual rarefaction analysis to evaluate how the diversity might change with the amount of sampling and the probability that additional sampling may recover additional taxa. Species evenness and richness were compared between both localities using Shannon’s H index (e.g. [52–54]), Simpson’s 1-D index (e.g. [55]), evenness E index and equitability J index [56]. The Berger–Parker dominance B was calculated to evaluate species dominance, and the improved Chao’s richness estimator (iChao1) [57,58] and the abundance-based coverage estimator (ACE) were used to estimate true species richness [59]. iChao1 and ACE should be close to the observed species richness if all taxonomic units were recovered from the localities. Finally, a permutation test (n = 999) for all diversity metrics was used to evaluate statistical significance of the differences between maximum and minimum diversity scenarios [60]. All calculations and analyses were performed in the software PAST5 [61], and the results are summarized in electronic supplementary material, tables S2–S4.

Extinction rate: squamate extinction rate for the Denver Basin was calculated by dividing the taxonomic diversity post-extinction by that before the extinction for minimum and maximum alpha diversity. Extirpation rate for taxa that survived the end-Cretaceous mass extinction elsewhere in North America but that are not among the observed survivors in DMNH Loc.2560 was also calculated. Finally, the extinction rate among North American squamates was calculated using the total diversity pre-extinction and the total diversity of immediate survivors (i.e. taxa that occur in Puercan 1 NALMA strata).

3. Results

(a) Fossil richness

Both the latest Cretaceous and early Palaeocene localities at West Bijou showed an even distribution of squamate taxa and neither is dominated by any one individual taxon or morphotype (table 2). The Late Cretaceous (latest Maastrichtian) squamate fauna (DMNH Loc.3648) produced a total of 235 individual squamate remains, of which 158 were anguid osteoderms that may belong to any of the proposed anguid morphotypes and taxa. The locality has also produced 36 tooth-bearing elements of various squamates, two diagnostic cranial elements and two diagnostic vertebrae (electronic supplementary material, table S1). Eight taxa were identified to the genus and species level, and an additional 19 morphotypes were recognized for a total of 27 OTUs (table 2 and figure 2). The most diverse clades at this locality are Polyglyphanodontia, with eight OTUs, and Anguimorpha (Anguidae+Varanoidea), with ten OTUs. Finally, among the morphotypes we identify a potential lacertoid (figure 2l).

Table 2.

Summary of diversity indices that are discussed in text. Datasets: 1, maximum diversity (all OTUs); 2, minimum diversity (previously named taxa only); 3, supergeneral taxa only (see §2 for details). Indices highlighted in italic type reflect significant difference between datasets 1 and 2. The bottom two metrics are modelled diversity estimates. OTU, operational taxonomic unit.

	Loc.3648			Loc.2560
	dataset 1	dataset 2	dataset 3	dataset 1	dataset 2	dataset 3
no. OTUs	27	8	8	3	3	3
no. specimens	40	20	38	4	4	4
Simpson’s 1-D	0.9679	0.8737	0.8307	0.8333	0.8333	0.8333
Shannon’s H	3.438	2.089	1.9	1.29	1.29	1.29
evenness E	1.153	1.009	0.8357	1.211	1.211	1.211
equitability J	1.043	1.004	0.8694	0.9464	0.9464	0.9464
Berger–Parker B	0.125	0.25	0.2632	0.5	0.5	0.5
iChao1	66	8.475	9.351	4.5	4.5	4.5
ACE	83.23	9.162	8.773	6	6	6

Figure 2.

Late Cretaceous squamates from DMNH Loc.3648 (a–aa) and early Palaeocene squamates from DMNH Loc.2560 (bb–dd). Polyglyphanodontia: (a) DMNH EPV.105814—cf. Chamops segnis, dentary; (b) DMNH EPV.107933—cf. Socognathus unicuspis, dentary; (c) DMNH EPV.105812—cf. Meniscognathus altmani, dentary; (d) DMNH EPV. 141547—cf. Obamadon gracilis, dentary; (e) DMNH EPV.141559—Polyglyphanodontia_A, parietal; (f) DMNH EPV.105813—cf. Polyglyphanodontia_B, dentary; (g) DMNH EPV.141594—Polyglyphanodontia_C, dentary; (h) DMNH EPV.107919—Polylgyphanodontia_D, dentary. Scincomorpha: (i) DMNH EPV.121447—Scincoidea_A, dentary; (j) DMNH EPV.107913—Scincoidea_B, maxilla; (k) DMNH EPV.57445—Cordyloidea_A, dentary. Lacertoidea: (l) DMNH EPV.141563—Lacertoidea, maxilla. Xenosauridae: (m) DMNH EPV.105792—cf. Exostinus lancensis, isolated tooth; (n) DMNH EPV.141579—Xenosauridae_A, maxilla. Anguimorpha: (o) DMNH EPV.94418—Odaxosaurus piger, dentary; (p) DMNH EPV.107967—Anguimorpha_A, maxilla; (q) DMNH EPV.121449—Anguimorpha_B, maxilla; (r) DMNH EPV.109622—Anguidae_A, dentary; (s) DMNH EPV.141556—Anguidae_B, parietal; (t) DMNH EPV.121448—Anguidae_C, dentary. Varanoidea: (u) DMNH EPV.107966—cf. Colpodontosaurus cracens, dentary; (v) DMNH EPV.141545—Varanoidea_A, dentary; (w) DMNH EPV.141572—Varanoidea_B, dentary; (x) DMNH EPV.107953—Varanoidea_C, dentary. Ophidia: (y) DMNH EPV.59231—cf. Coniophis precedens, vertebra. Incertae sedis: (z) DMNH EPV.107968—Incertae_sedis_A, maxilla; (aa) DMNH EPV.121323—Incertae_sedis_B, isolated tooth. Scincomorpha: (bb) DMNH EPV.103125—cf. Palaeoscincosaurus middletoni, dentary; Xenosauridae: (cc) DMNH EPV.99641—cf. Exostinus lancensis, isolated tooth. Anguimorpha: (dd) DMNH EPV.105630—cf. Odaxosaurus piger, dentary. (l,n,q) Lateral view; (e,s) dorsal view; (y) anterior view; all other specimens in medial view. Scalebar for each specimen, 1 mm.

Rarefaction analysis of the squamate remains from the Late Cretaceous locality (DMNH Loc.3648) suggested that diversity at the OTU (dataset 1; maximum diversity) level is undersampled, whereas rarefaction and subsequent analyses for dataset 2 (named taxa only) suggested low potential for recovery of additional diversity of named species (figure 3; table 2). The diversity permutation test showed that Simpson’s 1-D, Shannon’s H and Berger–Parker dominance B are statistically significantly different in both datasets, highlighting the impact of dataset composition (i.e. inclusion of unique morphotypes versus inclusion of only named taxa) on analysis of proportional species abundance (table 2). Evenness E and equitability J are similar, i.e. analyses of species evenness were not influenced by dataset composition (table 2). Our final dataset, using only higher taxonomic units, provided a rarefaction curve that reached its asymptote, suggesting higher taxonomic units have been fully sampled (figure 3c). The diversity indices agreed with the other two datasets in that taxon abundance (Simpson’s 1-D, Shannon’s H) is proportionate, there is no dominant taxon (Berger–Parker B), and the specimens are distributed evenly among all taxa (evenness E, equitability J; table 2).

Figure 3.

Rarefaction curves. (a) Rarefaction of all operational taxonomic units (OTUs; dataset 1) at DMNH Loc.3648 (green) and DMNH Loc.2560 (orange), showing greater potential for finding additional diversity at both localities. (b) Rarefaction of specimens that can be identified to the genus level (dataset 2) at both localities, showing potential for additional diversity at the Maastrichtian locality DMNH Loc.3648, but plateauing at the Palaeocene locality DMNH Loc.2560. (c) Rarefaction of all supergeneral taxa (dataset 3), showing that both localities are close to representing true supergeneral taxonomic diversity.

The early Palaeocene (Danian) squamate fauna (DMNH Loc.2560) produced 20 squamate elements (table 1). Of those, 11 were anguid osteoderms and nine were tooth-bearing elements, of which four could be attributed to Odaxosaurus piger, Palaeoscincosaurus middletoni and Exostinus lancensis (figures 2 and 4). Minimum and maximum alpha diversity for this locality is three taxa. The rarefaction curves for datasets 1 and 2 showed a potential for additional diversity, although both curves and the diversity metrices indicated that the missing diversity is very low (likely three or fewer taxa according to the ACE metric; figure 3 and table 2). The rarefaction curve for dataset 3 closely matched the beginning of the same curve of DMNH Loc.3648 (figure 3), but because the diversity indices were the exact same between all three datasets, little additional diversity is expected (table 2).

Figure 4.

Diversity of North American squamates across the K/Pg boundary. Thick black lines and bold names represent taxa present at DMNH Loc.3648 and Loc.2560; thin grey lines represent taxa that have been reported from the Maastrichtian and Palaeocene records of North America, but that currently are not known from West Bijou. K, Cretaceous; Pg, Palaeogene; Maas., Maastrichtian. Modified from [32].

(b) Abundance and extinction rate

There was a ca 91% reduction (235 Cretaceous fossils to 20 Palaeocene fossils) in relative abundance of squamate remains across the K/Pg boundary. Squamate fossils made up 6.45% of the Late Cretaceous dataset, whereas they made up 0.12% among the early Palaeogene vertebrate fauna (table 1). When accounting for dominant taxa (crocodilians and gar in both localities), the percentage of squamates in the Late Cretaceous increased to 21%, whereas that of squamates in the early Palaeocene increased to 1.52% (table 1).

Of the 27 squamate OTUs present in the Late Cretaceous locality, only two were found in the Palaeocene locality, providing an extinction rate of ca 93%. We note that the ophidian Coniophis was present in the Late Cretaceous locality (DMNH Loc.3648) and is known to survive the K/Pg mass extinction elsewhere in North America [32] but was not recovered from the Palaeocene locality at West Bijou. This suggests a local extirpation rate of ca 4%. Finally, the scincoid P. middletoni was known from the Palaeocene locality (DMNH Loc.2560) but was not part of the Late Cretaceous fauna (figure 2bb; electronic supplementary material, table S1).

4. Discussion

Unlike that of the Williston Basin, the depositional environment at West Bijou did not change across the K/Pg boundary [42,43,48], providing an opportunity to analyse vertebrate relative abundance and diversity pre- and post-extinction within the same depositional environment. Our data suggest a dramatic ca 91% decline in relative abundance of squamates across the K/Pg mass extinction event (table 1).

Integrating the squamate diversity of West Bijou with that reported by Longrich et al. [32] provided a more regional extinction rate for squamates from the Western Interior of North America. Assuming taxa that first appeared in the earliest Puercan 1 NALMA Palaeocene fossil record were survivors of the K/Pg mass extinction and not newly evolved forms suggests an 82% extinction rate, whereas assuming taxa that first appeared in the earliest Puercan 1 NALMA Palaeocene fossil record were newly evolved forms suggests an extinction rate of 87%. The West Bijou data showed a depauperate early Palaeocene squamate fauna and an undersampled Late Cretaceous fauna (dataset 1), which suggests the extinction rate of 93% is an accurate estimation. These results are similar to that of Longrich et al. [32] (76% species-level extinction) and, combined, these data paint a bleak picture for squamate survivorship across the K/Pg boundary (table 2).

Whereas the Longrich et al. [32] dataset provided a comprehensive North American squamate inventory to compare the squamate record at West Bijou, it was less chronostratigraphically resolved and therefore unable to suggest either a gradual or catastrophic extinction. West Bijou squamates were reliably placed within 57 kyr before and 128 kyr after the K/Pg boundary using a tightly constrained chronostratigraphic framework that includes a well-documented K/Pg boundary identified by primary markers (e.g. iridium spike, shocked minerals), palynostratigraphic biozones [45,46], well-resolved magnetochron boundaries [41], and U–Pb-dated volcanic ash beds (figure 1); West Bijou, therefore, indicates a relatively sudden extinction event that occurred over no more than 128 kyr. The specimens in Longrich et al. [32, suppl.], on the other hand, ranged in age from over 300 kyr to 1.8 Myr before the extinction and no age model was given for the specimens of the post-extinction fauna [32, suppl.], which allows either a gradual or catastrophic squamate extinction scenario. In the Williston Basin, the facies changes from lower to upper Hell Creek and from Hell Creek to Fort Union formations and the hypothesized rise in sea-level that underlies those changes [14,15,34,35] impacted faunal and floral composition even before the K/Pg extinction event [62]. The apparent sea-level rise and the accompanying changes in the ecosystem inhabited by snakes and lizards could explain a more stepwise extinction pattern shown by Longrich et al. [32]. West Bijou was more mountain-proximal but also a more stable depositional environment during the K/Pg boundary interval than that in the Williston Basin [14,15,34,35,42,43]. Because our dataset focused on a single ecosystem with a persistent depositional environment and habitats, the absence of some known survivors from the early Palaeogene at West Bijou indicates that local and regional extirpation may affect the picture of the severity of the K/Pg mass extinction.

The Cretaceous squamate fauna differed between the Williston and Denver basins, where polyglyphanodontian taxa dominated in the former and the Denver Basin was dominated by polylgyphanodontians and anguimorphs (Anguidae+Varanoidea) (figure 4). The K/Pg mass extinction affected both basins similarly and all squamate clades saw a drastic reduction in diversity (figure 4). Polyglyphanodontia went extinct in the Denver Basin, as well as elsewhere in North America [32] (figure 4). Early Palaeocene squamate faunas in North America were dominated by skinks and anguids (figure 4) [32]. This suggests an extinction selectivity favouring smaller-bodied insectivores and food generalists; larger-bodied (varanoids and polyglyphanodontians) squamates, macrocarnivores (varanoids) and largely herbivorous squamates (some polyglyphanodontians) preferentially went extinct (figure 4). This matches previous analyses of body-size selectivity in squamates [32] and dietary selectivity in mammals [24]. However, other ecological parameters that have been identified as playing a role in survivorship in other clades (e.g. substrate preference in mammals [63] and mode of life in turtles and crocodiles [64]) cannot be addressed with the present study.

The level of extinction among squamates at West Bijou was severe compared with that of other North American taxonomic groups [24,65] and indicates that squamates were affected more than other small terrestrial vertebrates (e.g. turtles, amphibians), but were comparable to extinction levels reported for mammals (75%) [24] and birds (80% of lineages) [29]. Unlike some amniote clades that had post-K/Pg boundary ‘bloom’ taxa (e.g. the mammals Mesodma thompsoni, Procerberus formicarum and Peradectes cf. pusillus [24]), there were no such examples among squamates as all taxa were present in low relative abundances early in the Palaeocene. Most squamates are diurnally active and are dependent on environmental parameters such as warm temperature for optimal growth to reproductive state [66,67], and the long period of darkness (e.g. [68]) and drastic cooling (e.g. [69,70]) that accompanied the impact may have been especially detrimental to the survival and early post-K/Pg recovery of squamates.

5. Conclusions

Two localities that are 425 m apart in the West Bijou study area east of Denver, Colorado record the changes in squamate diversity across the K/Pg boundary, the most recent mass extinction event. Exceptional chronostratigraphic control of the depositional system that persisted across the boundary without facies change sets the West Bijou study area apart from coeval fossil records across North America, e.g. the Williston Basin. Comparison of the pre- with the post-K/Pg boundary squamate faunas revealed high extinction rates (93%) compared with the total record for North America (76%). A 91% reduction in relative abundance of squamate fossils across the K/Pg boundary in the West Bijou study area further emphasizes the devastating effect of the mass extinction. This study showcases not only the selectivity of the K/Pg mass extinction, as most surviving squamates are small-bodied insectivores or food generalists, but also the suddenness of the extinction event (less than 128 kyr). However, our data also highlight the effects of local and regional extirpation on accurately reconstructing extinction rates as one known surviving taxon is present in the Late Cretaceous record but missing from the early Palaeogene record at West Bijou. Thus, our dataset represents a chronostratigraphically well constrained snapshot of the latest Cretaceous and early Palaeocene squamate fauna from North America.

Ethics

The Savory Institute and Byron Shelton gave permission to the Denver Museum of Nature & Science to conduct fieldwork and fossil extraction in the West Bijou study area. The West Bijou study area is private land, and therefore no collecting permit was necessary. The collected material was accessioned into the DMNS collections (accession numbers: 2010-25, 2010-31, 2014-173, 2015-25, 2016-28 and 2021-96) and the specimen numbers are provided in the manuscript.

Data accessibility

All primary data are included in the manuscript. Additional tables listing individual specimens with taxonomic and osteological identification as well as the tables used for datasets 1–3 have been uploaded as electronic supplementary material (Supplementary_tables). Additional specimen views and descriptions are provided as electronic supplementary material (Specimen_documentation).

The supplementary material is available online [71].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

H.P.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing—original draft, writing—review and editing; T.R.L.: data curation, funding acquisition, methodology, project administration, resources, validation, visualization, writing—review and editing.

Both authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work is supported by an anonymous funder to H.P. and by Lyda Hill Philanthropies, M.L. and S.R. Kneller, and a National Science Foundation (NSF-FRES-2317666) grant to T.R.L.