Palaeontology

Spatiotemporal sampling patterns in the 230 million year fossil record of terrestrial crocodylomorphs and their impact on diversity

Journal: Palaeontology

Manuscript ID PALA-07-18-4277-OA.R1

Manuscript Type: Original Article

Date Submitted by the Author: n/a

Complete List of Authors: Mannion, Philip; Imperial College, Department of Earth Science and Engineering Chiarenza, Alfio; Imperial College, Department of Earth Science and Engineering Godoy, Pedro; University of Birmingham, School of Geography, Earth and Environmental Sciences Cheah, Yung Nam; Imperial College, Department of Earth Science and Engineering

crocodiles, Crocodylomorpha, fossil record bias, Pull of the Recent, Key words: divergence times, diversity, spatiotemporal sampling

Note: The following files were submitted by the author for peer review, but cannot be converted to PDF. You must view these files (e.g. movies) online.

Supplementary Information.xls

Palaeontology Page 1 of 35 Palaeontology

1 2 3 Spatiotemporal sampling patterns in the 230 million year fossil record of terrestrial 4 crocodylomorphs and their impact on diversity 5 6 Philip D. Mannion1, Alfio Alessandro Chiarenza1, Pedro L. Godoy2 and Yung Nam Cheah1 7

8 1 9 Department of Earth Science and Engineering, Imperial College London, South Kensington 10 Campus, London, SW7 2AZ, UK; [email protected], [email protected], 11 [email protected] 12 13 2School of Geography, Earth and Environmental Sciences, University of Birmingham, 14 Birmingham, B15 2TT, UK; [email protected] 15

16 17 RRH: CROCODYLOMORPH FOSSIL RECORD 18 LRH: MANNION ET AL. 19 20 ABSTRACT 21 22 The 24 extant crocodylian species are the remnants of a once much more diverse and 23 24 widespread clade. Crocodylomorpha has an approximately 230 million year evolutionary 25 history, punctuated by a series of radiations and extinctions. However, the group’s fossil 26 record is biased. Previous studies have reconstructed temporal patterns in subsampled 27 crocodylomorph palaeobiodiversity, but have not explicitly examined variation in spatial 28 sampling, nor the quality of this record. We compiled a dataset of all taxonomically 29 diagnosable non-marine crocodylomorph species (393). Based on the number of 30 phylogenetic characters that can be scored for all published fossils of each species, we 31 32 calculated a completeness value for each taxon. Mean average species completeness (56%) 33 is largely consistent within subgroups and for different body size classes, suggesting no 34 significant biases across the crocodylomorph tree. In general, average completeness values 35 are highest in the Mesozoic, with an overall trend of decreasing completeness through time. 36 Many extant taxa are identified in the fossil record from very incomplete remains, but this 37 might be because their provenance closely matches the species’ present-day distribution, 38 rather than through autapomorphies. Our understanding of nearly all crocodylomorph 39 40 macroevolutionary ‘events’ is essentially driven by regional patterns, with no global 41 sampling signal. Palaeotropical sampling is especially poor for most of the group’s history. 42 Spatiotemporal sampling bias impedes our understanding of several Mesozoic radiations, 43 whereas molecular divergence times for Crocodylia are generally in close agreement with 44 the fossil record. However, the latter might merely be fortuitous, i.e. divergences happened 45 to occur during our ephemeral spatiotemporal sampling windows. 46

47 48 Keywords: crocodiles; Crocodylomorpha; divergence times; diversity; fossil record bias; Pull 49 of the Recent; spatiotemporal sampling 50 51 INTRODUCTION 52 53 Our uneven sampling of the fossil record obscures and biases our reading of 54 macroevolutionary patterns (Raup 1972). Whereas most studies have focused on 55 56 understanding and ameliorating fluctuations in sampling through time (e.g. Smith and 57 58 59 60 Palaeontology Palaeontology Page 2 of 35

1 2 3 McGowan 2007; Alroy 2010), spatial bias is also an important factor, but has received 4 relatively little attention (though see Allison and Briggs 1993; Noto 2011; Vilhena and Smith 5 2013; Close et al. 2017). If the geographical distribution of extinct taxa changes between 6 time intervals, fluctuations in global biodiversity (both observed and subsampled) might 7 merely be the result of shifts in spatial sampling patterns (Allison and Briggs 1993; Vilhena 8 9 and Smith 2013). As such, heterogeneity in our sampling of the fossil record needs to be 10 accounted for on spatial scales too, with a greater focus on well-sampled regions and 11 controlling for variation in geographic spread of the fossil record (Bush et al. 2004; Barnosky 12 et al. 2005; Vilhena and Smith 2013; Benson et al. 2016; Marcot et al. 2016; Close et al. 13 2017). 14 The 24 species of living crocodylians (alligators, caimans, crocodiles, and gavials) are the 15 remnants of a once much more diverse and widespread clade. Crocodylomorpha has an 16 17 approximately 230 million year (myr) evolutionary history (Brochu 2003; Irmis et al. 2013), 18 punctuated by a series of radiations and extinctions (Brochu 2003; Martin et al. 2014; 19 Bronzati et al. 2015; Pol and Leardi 2015; Mannion et al. 2015; Wilberg 2017) that appear to 20 be closely tied to fluctuations in temperature and aridity, at least in the terrestrial realm 21 (Markwick 1998; Carvalho et al. 2010; Mannion et al. 2015). However, like most groups with 22 a long evolutionary history, the fossil record of crocodylomorphs is biased. Previous studies 23 24 have attempted to attenuate the impact of sampling biases in order to reconstruct temporal 25 patterns in crocodylomorph palaeobiodiversity (Markwick 1998; Mannion et al. 2015; Pol 26 and Leardi 2015; Tennant et al. 2016), but have not explicitly examined the effects of 27 variation in spatial sampling. 28 The completeness of the fossil specimens themselves yields additional information on 29 fossil record bias that is not captured in those diversity analyses. Several studies have tried 30 to capture information on fossil specimen completeness via the use of coarse taphonomic 31 32 classes, e.g. using a completeness score ranging from 0 (no elements preserved) to 4 (all 33 elements preserved) for discrete regions of a skeleton (e.g. Fountaine et al. 2005; Benton 34 2008; Dyke et al. 2009; Beardmore et al. 2012a, b). A growing body of work has attempted 35 to quantitatively estimate fossil specimen completeness, applying this to an array of 36 Paleozoic to early Cenozoic tetrapods (Mannion and Upchurch 2010; Brocklehurst et al. 37 2012; Walther and Fröbisch 2013; Brocklehurst and Fröbisch 2014; Cleary et al. 2015; 38 Verrière et al. 2016; Davies et al. 2017; Tutin and Butler 2017; Driscoll et al. 2018), but this 39 40 approach has never been applied to an uninterrupted time series of an extant group. 41 Here we present a detailed examination of the quality of the terrestrial crocodylomorph 42 fossil record throughout the group’s 230 myr evolutionary history, exploring how 43 spatiotemporal fluctuations in sampling and specimen completeness impact upon our 44 reading of macroevolutionary patterns. 45 46 MATERIALS AND METHODS 47 48 49 A compilation of all taxonomically diagnosable species of non-marine (terrestrial + 50 freshwater) fossil crocodylomorphs was assembled based on updates to Mannion et al. 51 (2015), Tennant et al. (2016), information in the Paleobiology Database 52 (http://paleobiodb.org), and a thorough review of the literature. Taxa were determined to 53 be non-marine based on previous assessments in the literature (e.g. Mannion et al. 2015), 54 as well as their depositional environments. Only taxa based on body fossils were included. 55 56 Data on higher taxonomic level (e.g. Alligatoroidea, Notosuchia), geography 57 58 59 60 Palaeontology Page 3 of 35 Palaeontology

1 2 3 (palaeocontinent, country, palaeolatitudinal range), and stratigraphy (temporal range, 4 geological formation), were also recorded. Our dataset comprises 393 crocodylomorph 5 species recognized in the fossil record, and is up to date as of the 4th July 2017 (Mannion et 6 al. 2018a). 7 We also estimated the total length (TL) of each species as a measure of body size, 8 9 primarily using formulae available in the literature to estimate TL from several parts of the 10 skeleton. These equations are derived from linear regressions based on specimens of 11 modern taxa. For specimens with complete skulls preserved, we used the formula from 12 Hurlburt et al. (2003) that estimates TL from dorsal cranial length (DCL). For specimens with 13 skulls that only lack the rostrum, we were able to use another formula from Hurlburt et al. 14 (2003) that uses orbito-cranial dorsal length (ODCL). Also, for specimens with femora 15 preserved, we applied the formula from Farlow et al. (2005) that estimates TL from femoral 16 17 length (FL). Finally, for specimens lacking near-complete skulls or femora, we used other 18 preserved elements to coarsely estimate either DCL or FL, and then applied the formulae 19 from Hurlburt et al. (2003) and Farlow et al. (2005). When more than one specimen was 20 available for a single species, we utilised the largest estimated value of TL. Similarly, when 21 we were able to apply more than one formula to estimate TL for a species (e.g. if a skull and 22 femur are preserved for one species), we used the largest value in our analyses. This 23 24 approach enabled us to estimate TL for nearly all crocodylomorphs in our dataset (363 25 species). TL estimates (in metres) were log-transformed prior to correlation analyses. To test 26 for a possible differential influence of completeness on animals of distinct size magnitudes, 27 we also allocated taxa into four coarse size bins (categories): (A) taxa smaller than 2 metres; 28 (B) taxa between 2 and 4 metres; (C) taxa between 4 and 6 metres; and (D) taxa larger than 29 6 metres. 30 Mannion and Upchurch (2010) devised a character completeness metric (CCM) that 31 32 quantifies the amount of phylogenetic data preserved in fossil specimens. A percentage 33 score is provided for the number of phylogenetic characters that can be scored for each 34 taxon, based on all published specimens. We utilized the phylogenetic data matrix of 35 Montefeltro et al. (2013), which has the greatest number of characters (484) of any 36 published crocodylomorph morphological data matrix, and samples taxa and characters 37 from across the tree. Most characters pertain to cranial elements (393), with vertebral (31), 38 appendicular (46), and osteodermal (14) characters providing a much smaller contribution. 39 40 A similar distribution of characters is present in other crocodylomorph data matrices (e.g. 41 Brochu and Storrs 2012). A completeness score was calculated for each species based on an 42 extensive review of the literature and personal observations of fossil specimens (see Fig. 1 43 for an illustration of body regions). An element that is clearly preserved, but not visible (e.g. 44 concealed by matrix), was regarded as complete. With regard to inapplicable characters, we 45 ‘scored’ each species as to whether it had the potential to preserve that feature, e.g. if one 46 humerus was completely preserved, then all humeral characters were considered as 47 48 ‘scorable’. The exception to this was the small number of taxa that unambiguously lack 49 osteoderms: osteodermal characters were excluded from the CCM for these species. 50 Inapplicable characters comprise a very small proportion of the 484 characters, and their 51 treatment has little impact on the calculated CCM for each species. 52 A mean average value of crocodylomorph CCM scores and associated standard deviations 53 was calculated for each time bin. We utilized two different time-binning approaches. Firstly, 54 we used standard stratigraphic stages (though with the Quaternary counted as one interval, 55 56 as some of its stages are extremely short in duration), based on the latest (2017) version of 57 58 59 60 Palaeontology Palaeontology Page 4 of 35

1 2 3 the geological time scale presented by Cohen et al. (2013). However, these stages display a 4 large disparity in temporal duration, ranging from 1 myr to 19 myr in length. The ages of 5 many crocodylomorph fossils are also poorly constrained, meaning that numerous 6 occurrences cannot be assigned to a single stage. As such, our second binning strategy uses 7 approximately equal-length (~9 million years) stratigraphic time bins that group together 8 9 multiple stages, whilst maintaining important geological boundaries (see Mannion et al. 10 2015; Benson et al. 2016). In both cases, we did not exclude species that could not be 11 assigned to a single time bin, i.e. their full stratigraphic uncertainty was used as their 12 temporal range. Both sets of time bins are detailed in Mannion et al. (2018a). 13 Temporal fluctuations in completeness were reconstructed at the global level for both 14 time-binning strategies. Completeness at the palaeocontinental scale, as well as for 10° 15 palaeolatitudinal bands, was reconstructed only for the 9 myr bins. Our palaeocontinents 16 17 comprise: Africa, Asia, Australasia, Europe, India, Madagascar, North America, and South 18 America. Only one crocodylomorph species is known from Madagascar prior to the 19 Maastrichtian: this is from the Middle Jurassic, at which time Madagascar was still 20 connected to Africa, and thus it is considered an African species here. Otherwise, both 21 Madagascar and India lack diagnosable crocodylomorph fossils prior to the Maastrichtian, 22 and thus they are not grouped together as Indo-Madagascar, nor are they incorporated with 23 24 Africa, with rifting having separated these areas by the latest Cretaceous (Ali and Aitchison 25 2008). India is incorporated within Asia from the Eocene onwards, following their collision 26 (Ali and Aitchison 2008). We also calculated completeness for several major taxonomic 27 subgroups and for our four body size classes. 28 A species-level census of observed (=uncorrected/raw) crocodylomorph diversity through 29 time was calculated from our dataset at global, palaeocontinental, and palaeolatitudinal 30 scales. We also produced global, palaeocontinental, and palaeolatitudinal time series of 31 numbers of non-marine crocodylomorph-bearing collections (CBCs), as a sampling proxy 32 th 33 (2424 collections in total), using data in the Paleobiology Database, accessed 29 May 2018 34 (Mannion et al., 2018b). 35 All previous equivalent studies of fossil specimen completeness have been restricted to 36 Paleozoic to early Cenozoic (Palaeogene) taxa (e.g. Brocklehurst et al. 2012; Brocklehurst 37 and Fröbisch 2014; Cleary et al. 2015; Davies et al. 2017), meaning that we cannot 38 undertake a comparison with other groups for the full time interval of interest. We 39 40 therefore limited comparisons with existing completeness metrics to two terrestrial groups 41 that span the Late Triassic through to end-Cretaceous (Maastrichtian): (1) sauropodomorph 42 dinosaurs (Mannion et al. 2010), with large and robust skeletal elements; and (2) pterosaurs 43 (Dean et al. 2016), which are generally small-bodied, with delicate bones. In both instances, 44 statistical comparisons with crocodylomorph completeness were made only at the stage- 45 level, using the data presented in Dean et al. (2016). 46 Correlation tests were used to compare fluctuations in crocodylomorph completeness 47 48 with other temporal and spatial series (i.e. numbers of crocodylomorph species and 49 crocodylomorph-bearing collections [CBCs], completeness of sauropodomorphs and 50 pterosaurs). All of our time series were log-transformed prior to analysis, and the effects of 51 trend and autocorrelation were removed via generalized differencing, using a function 52 written by Graeme Lloyd (http://www.graemetlloyd.com/methgd.html). Bins with zero 53 values were excluded from our temporal and spatial series correlation tests. Spearman's 54 rank correlation coefficients were calculated for each pairwise comparison. 55 56 57 58 59 60 Palaeontology Page 5 of 35 Palaeontology

1 2 3 We used ordinary least squares (OLS) regressions of the completeness scores on our 4 body size data to test for correlation. We also performed these OLS regressions within each 5 of our four body size categories (i.e. A–D). 6 All correlation analyses were performed in R version 3.4.3 (R Core Team, 2017). Time 7 series plots were produced using the R package strap (Bell and Lloyd 2015). The complete 8 9 dataset is available in Mannion et al. (2018a). 10 11 RESULTS 12 13 Global patterns in completeness and taxonomic diversity 14 15 Just six species are known solely from postcrania, whereas 186 species are known only 16 17 from cranial remains. The average CCM of all 393 species is 56%, with values ranging from 18 0.2–100% (Fig. 2). Completeness of Mesozoic species is higher on average (217 species with 19 an average completeness of 60%) than that of Cenozoic species (179 species: 51%) (Fig. 2). 20 Within the Mesozoic, the greatest average values are from Jurassic species (51 species: 21 72%), with notably lower values in the Late Triassic (9 species: 64%) and especially the 22 Cretaceous (159 species: 55%). Completeness in the Palaeogene (81 species: 54%) is greater 23 24 than that of the Neogene + Quaternary (99 species: 48%). 25 At a global level, there is no significant correlation between temporal fluctuations in 26 crocodylomorph diversity and completeness in any of our analyses (Table 1). Whereas 27 diversity is strongly correlated with numbers of crocodylomorph-bearing collections (CBCs) 28 in all of our analyses (Spearman's rho (rs) = 0.72–0.86), completeness and numbers of CBCs 29 are only significantly correlated when using 9 myr bin-level analyses, after applying 30 generalized differencing (rs = 047; Table 1). Below we describe global patterns in 31 32 completeness and taxonomic diversity through the last 230 myr (Figs 3, 4). 33 Crocodylomorph completeness is low in the first stage of the Late Triassic (Carnian), 34 although only two valid species are known from this interval. Three times as many species 35 are known from the Norian, which also has a substantially higher completeness value. 36 Completeness reaches its highest percentage for the entirety of the evolutionary history of 37 Crocodylomorpha in the Rhaetian, although this value is only based on two known species. 38 There is a decline in completeness across the Triassic/Jurassic boundary (203 Ma), although 39 40 CCM is still high for the more species-rich Hettangian–Sinemurian (Figs 3, 4). No diagnostic 41 terrestrial crocodylomorph species are currently known from the last two stages of the Early 42 Jurassic (Pliensbachian and Toarcian), or from the first stage of the Middle Jurassic 43 (Aalenian). Both completeness and number of species is lower in the Bajocian–Bathonian 44 than the earliest Jurassic, but increase in the Callovian and Oxfordian. The last two stages of 45 the Jurassic are characterized by lower completeness, but much greater numbers of species 46 (Figs 3, 4). 47 48 Completeness and number of species decreases slightly across the Jurassic/Cretaceous 49 boundary (145 Ma), with species numbers declining further in the Valanginian and 50 Hauterivian (Figs 3, 4). A much greater number of species is known from the Barremian– 51 Albian, with slightly higher average completeness values. Although there are a similar 52 number of species in the Cenomanian, completeness plummets in this first stage of the Late 53 Cretaceous. Whereas the number of species declines, completeness slightly increases in 54 subsequent stages (Turonian and Coniacian). The number of species approximately doubles 55 56 in the Santonian, although there is a decrease in completeness (Figs 3, 4). Both the number 57 58 59 60 Palaeontology Palaeontology Page 6 of 35

1 2 3 of species and average completeness are considerably higher in the Campanian and 4 Maastrichtian (note that the age of the species-rich Late Cretaceous Brazilian Adamantina 5 Formation is poorly constrained, but is here assigned to the late Campanian–early 6 Maastrichtian based on recent stratigraphic work by Batezelli [2017]). 7 Aside from the Carnian value (see Figs 3, 4), completeness reached its nadir in the first 8 9 interval (Danian) after the Cretaceous/Palaeogene boundary (66 Ma). However, 10 completeness rebounded in the subsequent stages of the Paleocene (Selandian and 11 Thanetian). Both the Ypresian and Lutetian have similar values, prior to a decline in the 12 Bartonian. Completeness increases substantially in the final stage of the Eocene 13 (Priabonian), although the number of species is unchanged from the preceding interval. Just 14 two species are recognized from the Rupelian (early Oligocene), although their average 15 completeness is one of the highest values. Whereas the number of species increases in the 16 17 Chattian (late Oligocene), completeness drops precipitously (Figs 3, 4). 18 These values do not change substantially across the Palaeogene/Neogene boundary (23 19 Ma). Both values increase in the Burdigalian, and the number of species rises in the 20 Langhian and Serravallian. The number of species reaches its Cenozoic acme in the 21 Tortonian, and high values for completeness and number of species are recorded in the final 22 stage of the Miocene (Messinian). Species counts decrease in the Pliocene. Finally, an 23 24 increased number of species is recognized in the Quaternary, although completeness is 25 reduced (Figs 3, 4). 26 27 Palaeocontinental sampling patterns 28 29 The 393 crocodylomorph species have the following palaeocontinental distribution (Fig. 30 5): Africa (43 species with average completeness of 61%), Asia (68 species: 56%), Australasia 31 32 (25 species: 35%), Europe (70 species: 69%), India (one species: 15%), Madagascar (six 33 species: 70%), North + Central America (74 species: 57%), and South America (106 species: 34 48%). The spread of completeness data is relatively similar for all palaeocontinents, with the 35 exception of Europe (Fig. 5), for which most species are known by fairly complete remains 36 (>60%). No palaeocontinent has a continuous (i.e. uninterrupted) record throughout the 37 evolutionary history of crocodylomorphs, and often a time interval is dominated by a single 38 region (Figs 6, 7). There are significant positive correlations between temporal fluctuations 39 40 in South American crocodylomorph completeness and both numbers of species 41 (uncorrected time series and with generalized differencing) and CBCs (generalized 42 differencing only) (Table 1). In contrast, there is no correlation between completeness and 43 the numbers of species or CBCs for the other palaeocontinents with relatively complete 44 time series (i.e. Asia, Europe, and North America). 45 Late Triassic species are restricted to Europe, North America, and South America, and 46 only Africa (restricted to southern Africa), Asia and North America contribute to our 47 48 knowledge of Early Jurassic crocodylomorph species (Figs 6, 7). Middle Jurassic–Oxfordian 49 species are primarily restricted to Asia, with two species from South America, and just one 50 each from North America and Africa. The Kimmeridgian–Tithonian signal reflects a slightly 51 more global dataset, with species known from Asia, Europe, North America, and South 52 America. European and Asian species comprise most of the diversity of Berriasian– 53 Barremian crocodylomorphs (with a small contribution from Africa and South America). The 54 Aptian–Cenomanian record is dominated by African species (primarily from North Africa), 55 56 although Asia, North America, Europe (Albian-only), and South America also contribute, as 57 58 59 60 Palaeontology Page 7 of 35 Palaeontology

1 2 3 does Australasia for the first time (Figs 6, 7). The Turonian–Santonian signal predominantly 4 stems from South America and Asia, although there are additional species from Europe, and 5 one from each of Africa and North America. Campanian–Maastrichtian species are found in 6 Asia (Campanian-only), Europe, North America, South America, and Indo-Madagascar 7 (Maastrichtian-only). 8 9 Paleocene–Eocene crocodylomorph species are known from Asia (now including India), 10 Europe, North America, and South America, with the Ypresian record supplemented by a 11 species from Africa, and four Australasian species known throughout the Eocene (Figs 6, 7). 12 Australasia is the principal contributor to our knowledge of Oligocene diversity, with a single 13 species otherwise known from each of Africa, Europe, and North America during this epoch 14 (note that most of the fossil occurrences in Fig. 6I represent generically or specifically 15 indeterminate materials). Aside from Madagascar, the Miocene record of crocodylomorph 16 17 species is global. This distribution of species is retained in the Pliocene, with the exception 18 of Europe, and Quaternary crocodylian species are known from Africa, Asia, Australasia, 19 North America, South America, and Madagascar (Figs 6, 7). 20 21 Palaeolatitudinal sampling patterns 22 23 24 Although crocodylomorph fossils have been discovered at palaeolatitudes of 76° N (early 25 Eocene of the Canadian Arctic; Estes and Hutchison 1980) and 79° S (late Eocene of 26 Antarctica; Willis and Stilwell 2000), remains that can be assigned to species are restricted 27 to 60° palaeolatitude either side of the Equator (Fig. 8). As with the palaeocontinental data, 28 no palaeolatitudinal belt has a continuous record throughout the evolutionary history of 29 crocodylomorphs (Fig. 9), and for many time intervals we have little or no record of the 30 group in the palaeotropics (a band of approximately 30° either side of the Equator) or 31 32 subpolar regions (although the latter might be a genuine absence outside of greenhouse- 33 dominated intervals). 34 In general, completeness is greater in the Northern Hemisphere (Figs 8, 9). There are high 35 values for 50–60° N, 40–50° N, and 20–30° N, with a moderately high value for 30–40° N, 36 and completeness in more tropical palaeolatitudes is only slightly lower. Aside from 37 palaeolatitudinal bins for which there is only one species, the highest completeness value 38 for a Southern Hemisphere palaeolatitudinal band is 0–10° S. Completeness is also relatively 39 40 high at 20–30° S, whereas values are lower at 30–40° S, 40–50° S and, especially, at 10–20° 41 S. 42 There is no significant correlation between the palaeolatitudinal distribution of 43 crocodylomorph diversity and completeness in any of our analyses (Table 2). A weak, 44 though statistically non-significant correlation is recovered between diversity and numbers 45 of CBCs, although only when generalized differencing is applied (Table 2). In contrast, a 46 strong correlation between completeness and numbers of CBCs is recovered, using both the 47 48 raw data and following the application of generalized differencing (Table 2). Below we 49 describe how the palaeolatitudinal distribution of crocodylomorph completeness and 50 taxonomic diversity varied over the last 230 myr. 51 Excluding the middle Miocene–Quaternary, for which sampling of the palaeotropics is 52 generally good, most preceding time intervals have very low species counts (zero or one) for 53 each of their palaeolatitudinal tropical bins (Fig. 9). With the exception of a single species 54 from the Bathonian of Madagascar, the Late Triassic–Middle Jurassic palaeotropical record 55 56 is limited to the Northern Hemisphere. In contrast, the palaeotropics of the Campanian– 57 58 59 60 Palaeontology Palaeontology Page 8 of 35

1 2 3 Danian (inclusive) is only represented in the Southern Hemisphere (Fig. 9). Notable 4 exceptions of a palaeotropical record, based on multiple species, are restricted to: 10–20° N 5 in the Hettangian–Sinemurian and Hauterivian–Barremian; 0–10° N in the Aptian, Albian, 6 and Cenomanian; 20–30° S in the Campanian, Maastrichtian, and Selandian–Thanetian; and 7 0–10° N in the Aquitanian–Burdigalian (Fig. 9). 8 9 The palaeolatitudinal band spanning 30–40° N is sampled for most of crocodylomorph 10 evolutionary history, with the exception of the Carnian, Pliensbachian–Bathonian, and 11 Ypresian, and the 40–50° N band is sampled continuously from the Callovian onwards (Fig. 12 9). Average completeness values based on more than two species are mostly comparable 13 with or exceed the global average, although there are a small number of low values, i.e. 30– 14 40° N in the Turonian–Santonian, and 40–50° in the Maastrichtian and Danian. 15 Crocodylomorph species are only sampled in the 50–60° N palaeolatitudinal band in the 16 17 Hauterivian–Barremian and Turonian–Ypresian intervals, with completeness generally high, 18 but numbers of species mostly low (Fig. 9). 19 Sampling of temperate palaeolatitudes in the Southern Hemisphere (30–60° S) is 20 sporadic, especially in the Late Triassic–Jurassic, and usually comprises two or fewer species 21 per bin (Fig. 9). Where there are more than two species, completeness values are quite 22 variable. Whereas there is high completeness at 40–50° S in the Hettangian–Sinemurian, 23 24 and at 30–40° S in the Aptian, Albian, Cenomanian, and Maastrichtian, low completeness 25 values characterize the 40–50° S band in the Turonian–Santonian, as well as the 30–40° S 26 band in the Oligocene and Langhian–Tortonian (Fig. 9). Only one species is known from 50– 27 60° S (Albian–Cenomanian), although it is an essentially complete skeleton (i.e. Isisfordia 28 duncani; Salisbury et al. 2006). 29 30 Body size and taxonomic variation in completeness 31 32 33 In general, our regression analyses show non-significant or very weak correlations 34 between crocodylomorph fossil completeness and body size (Fig. 10; Table 3). A significant 35 negative correlation (i.e. negative slope) was recovered when analysing all species (with 36 untransformed or log-transformed body size data), and species within body size category A 37 (smaller than 2 m), indicating that smaller taxa tend to be more complete. However, in 38 these cases, the coefficient of determination (R2) was very low (always smaller than 0.03), 39 40 indicating that the correlation is very weak. For species within the other three size 41 categories, no significant correlation was recovered. Comparisons between categories 42 suggests that taxa within category C (between 4 and 6 meters) are the least complete. 43 The average completeness of several well-recognised crocodylomorph subclades does 44 not strongly deviate from the overall pattern (Fig. 11), with Notosuchia (72 species with 45 average completeness of 54%), Neosuchia (256 species: 54%), Crocodylia (172 species: 46 52%), Alligatoroidea (73 species: 56%) and Tomistominae (17 species: 55%), for example, all 47 48 within a few percentage points of the total average (55%), although Tethysuchia (10 species: 49 48% and Crocodylidae (50 species: 49%) are slightly lower, and Goniopholididae (25 species: 50 60%) a little higher. The spread of completeness data is also relatively consistent between 51 groups (Fig. 11), with most species represented by specimens that are approximately 15– 52 80% complete. Goniopholididae is the exception to this pattern, with the majority of 53 goniopholidid species known from specimens that are 50–80% complete (Fig. 11). Although 54 a paraphyletic grade, non-metasuchian crocodylomorphs (58 species) tend to preserve more 55 56 complete skeletons, with an average completeness of 65% (Fig. 11). 57 58 59 60 Palaeontology Page 9 of 35 Palaeontology

1 2 3 4 Crocodylomorph completeness compared to other groups 5 6 Moderate to strong correlations were recovered between temporal fluctuations in the 7 raw completeness values of Mesozoic crocodylomorphs and those of sauropodomorph 8 9 dinosaurs (rs = 0.51) and, especially, pterosaurs (rs = 0.75) (Table 4). However, both of these 10 correlations disappeared when generalized differencing was applied to the time series. 11 12 DISCUSSION 13 14 The completeness of the crocodylomorph fossil record 15

16 17 Whereas the first two studies to utilise the character completeness metric (Mannion and 18 Upchurch 2010; Brocklehurst et al. 2012) examined taxonomic groups (sauropodomorph 19 dinosaurs and Mesozoic birds) for which the skull represented approximately one-third of 20 the total character set, analyses of other groups have all shown that the skull has a much 21 greater character weighting for those taxa, with: (1) parareptiles = 52% (Verrière et al. 22 2016); (2) plesiosaurs = 52% (Tutin and Butler 2017); (3) pterosaurs = 59% (Dean et al. 23 24 2016); (4) pelycosaurs = 70% (Brocklehurst and Fröbisch 2014); (5) Cretaceous–Palaeogene 25 eutherian mammals = 73% (Davies et al. 2017); (6) crocodylomorphs = 81% (this study); and 26 (7) anomodonts = 82% (Walther and Fröbisch 2013). Thus, for groups such as anomodonts 27 and crocodylomorphs, an ‘80% complete skeleton’ might only preserve a skull. A skeletal 28 completeness metric (SCM), in which the skeleton is divided up into percentages based on 29 the amount of bone for each region (Mannion and Upchurch 2010), is thus likely to result in 30 very different completeness values for groups such as crocodylomorphs, although temporal 31 32 fluctuations in CCM and SCM might still be expected to correlate (see Brocklehurst and 33 Fröbisch 2014; Verrière et al. 2016; Tutin and Butler 2017). 34 This greater weighting of the skull is largely consistent across crocodylomorph 35 phylogenetic data sets (e.g. Brochu and Storrs 2012; Pol et al. 2014). Although we do not 36 disagree that the skull is disproportionately rich in character data compared to the 37 postcranial skeleton, we contend that this is accentuated by historical neglect of the 38 postcrania in studies of crocodylomorph anatomy (Godoy et al. 2016; Martin et al. 2016). 39 40 Although postcrania are usually listed in taxon descriptions, they are not always figured or 41 described in detail, and this bias subsequently affects the choice of characters that end up 42 being incorporated into phylogenetic data sets. Fortunately, this is beginning to change, 43 with several studies generating novel postcranial characters, and demonstrating their 44 impact on resolving crocodylomorph relationships (e.g. Pol et al. 2012; Godoy et al. 2016; 45 Martin et al. 2016). 46 The overall consistency in average completeness across taxonomic groups suggests that 47 48 depositional environment does not have any great control on non-marine crocodylomorph 49 fossil preservation, at least at coarse levels, given that our dataset includes several semi- 50 aquatic groups, as well as fully terrestrial notosuchians living in semi-arid climates (Carvalho 51 et al. 2010). The weak negative correlation between completeness and body size might 52 initially appear surprising, given that small-bodied individuals are likely to be most 53 susceptible to the vagaries of preservation (e.g. Hill and Behrensmeyer 1984; Brand et al. 54 2003). Lagerstätten likely account for the high completeness of some of the very smallest 55 56 species, but given that this general pattern can be observed within size classes (albeit 57 58 59 60 Palaeontology Palaeontology Page 10 of 35

1 2 3 statistically non-significantly), it is probable that this does not drive the overall trend. One 4 possibility pertains to the nature of the crocodylomorph skeleton. In contrast to the delicate 5 bones that comprise the skeletons of primarily small-bodied tetrapod groups such as 6 lepidosaurs and lissamphibians, as well as the light skeletons of birds and pterosaurs, the 7 skeletal elements of crocodylomorphs tend to be fairly robust, especially those of the skull. 8 9 Given that the skull comprises most of the character information for crocodylomorphs, this 10 might explain why even small-bodied taxa tend to have relatively ‘complete’ (i.e. ≥80%) 11 skeletons. There might also be an upper limit of body size for good fossil preservation 12 (Cleary et al. 2015); for example, Mannion and Upchurch (2010) demonstrated that the 13 largest species of sauropodomorph dinosaurs tend to be known from less complete remains 14 than their smaller-bodied relatives. However, a bias against the completeness of small- 15 bodied species was recovered in studies of ichthyosaurs (Cleary et al. 2015; Beardmore 16 17 2017) and some dinosaur groups (Brown et al. 2013). Given that average completeness 18 through time of crocodylomorphs correlates with neither pterosaurs nor sauropodomorphs, 19 it seems that perhaps neither group is a suitable analogue. 20 21 Spatiotemporal bias pervades the crocodylomorph fossil record 22 23 24 Although fossil completeness does not appear to be a primary driver of observed 25 crocodylomorph diversity, there are a number of time intervals in which the two covary 26 (Figs 3, 4). Both decline across the Jurassic/Cretaceous boundary, 145 Ma (see below), but 27 the most notable instance occurs at the K/Pg boundary, 66 Ma. A substantial decline in 28 completeness, from 60% (Maastrichtian) to 38% (Danian), coincides with a taxonomic 29 diversity crash from 50 to 13 species. Given that a sampling standardisation approach 30 reveals little change in overall standing diversity for crocodylomorphs across this mass 31 32 extinction (Mannion et al. 2015), we might regard low average specimen completeness as a 33 confounding factor in our ability to identify species in the Danian. However, we caveat this 34 interpretation by noting that whereas the Maastrichtian record comprises a high number of 35 species from several palaeocontinents, eight of the 13 Danian species are from North 36 America, and completeness actually increases across the K/Pg boundary on that 37 palaeocontinent. As such, although poor levels of completeness might contribute to low 38 observed (i.e. uncorrected) diversity in the Danian, the reduction in our spatial sample 39 40 appears to be the primary control (see below). This interpretation is also supported by the 41 subsequent parallel increase in diversity, completeness, and geographic spread of fossil 42 localities in the middle–late Paleocene (Figs 4, 7). 43 As documented above, no palaeocontinent or palaeolatitudinal band has a continuous 44 fossil record throughout the evolutionary history of crocodylomorphs, and often a time 45 interval is dominated by a single geographical region (Figs 5–8). Consequently, there is no 46 meaningful concept of global terrestrial palaeobiodiversity, with our understanding of 47 48 nearly all diversification and extinction ‘events’ essentially driven by regional patterns. This 49 is well illustrated by diversity patterns across the J/K boundary: whereas there is a semi- 50 cosmopolitan Late Jurassic record, the earliest Cretaceous signal predominantly comes from 51 Europe (Tennant et al. 2016). Is the documented diversity decline across the J/K boundary a 52 global event, or was it restricted to Europe? A similar problem is apparent across the K/Pg 53 boundary (see above), as well as at hypothesized times of a number of evolutionary 54 radiations (see below). 55 56 57 58 59 60 Palaeontology Page 11 of 35 Palaeontology

1 2 3 A poor, often entirely absent, fossil record characterizes the palaeotropics for much of 4 pre-Neogene crocodylomorph evolutionary history (Markwick 1998; Brochu and Storrs 5 2012; Scheyer et al. 2013; Mannion et al. 2015; Salas-Gismondi et al. 2015). Although it 6 remains possible that the palaeotropics were too hot to sustain crocodylomorph diversity 7 during some time intervals, this seems unlikely given that there is a fossil record during the 8 9 Late Triassic, early Late Cretaceous, and late Paleocene, all of which represent warmer 10 intervals (Sellwood and Valdes 2006; Zachos et al. 2008). Also, if ‘taphonomic control taxa’ 11 (i.e. taxa with an approximately similar preservation potential as crocodylomorphs; sensu 12 Bottjer and Jablonski 1988) are present, then we might infer that the absence of 13 crocodylomorphs in this area is genuine (Markwick 1994; Matsumoto and Evans 2010). 14 However, an examination of the distribution of the fossil record of Mesozoic dinosaurs, for 15 example, indicates that there are no palaeotropical regions that yield abundant dinosaur, 16 17 but not crocodylomorph, remains (e.g. Mannion et al. 2012). As such, this scarcity of 18 crocodylomorph fossil remains is best explained as an artefact, resulting from a 19 palaeotropical bias. 20 It is well established that the fossil record of nearly all organisms is dominated by 21 occurrences from North America and western Europe (Jackson and Johnson 2001), for a 22 suite of reasons that include: (1) the longer history of collecting fossils on these continents 23 24 than elsewhere; (2) the relatively shorter distances between conurbations and fossil sites 25 compared to other continents; and (3) the long-term prosperity of most fossil-bearing 26 countries in these two continents. However, the northward drift of Pangaea during the 27 Paleozoic and early Mesozoic resulted in these palaeocontinental regions (as well as much 28 of Asia) lying primarily outside of the tropics from the Jurassic onwards (Allison and Briggs 29 1993; Ziegler et al. 2003). As such, our potential sampling of palaeotropical non-marine 30 fossil sites is dominated by Central America and Gondwana for most of crocodylomorph 31 32 evolutionary history, and yet our sampling of these regions is extremely patchy (e.g. Benson 33 et al. 2013; Kemp and Hadly 2016). Furthermore, rates of chemical weathering are generally 34 higher in the tropics than at other latitudes, as a consequence of the higher temperatures 35 and precipitation levels that characterize much of the tropics (Rees et al. 2004). This results 36 in decreased fossil preservation (Behrensmeyer et al. 2000), as well as an increase in 37 sedimentary rock breakdown (Kemp and Hadly 2016), and therefore acts as an additional 38 filter to sampling palaeotropical biodiversity. In the rare instances that we do sample high 39 40 palaeotropical crocodylomorph diversity prior to the Neogene (see Mannion et al. 2015), 41 these localities usually represent semi-arid environments, inhabited by the ecologically 42 unusual (i.e. predominantly terrestrial) notosuchians (Carvalho et al. 2010). Consequently, 43 these geological, anthropogenic, and taphonomic palaeotropical biases likely have a 44 significant effect on our understanding of the evolutionary history of crocodylomorphs, as is 45 also the case for most taxonomic groups (e.g. Allison and Briggs 1993; Jackson and Johnson 46 2001; Kidwell and Holland 2002; Bush and Bambach 2004; Vilhena and Smith 2011; Kemp 47 48 and Hadly 2016). 49 50 Does poor sampling obfuscate timings of crocodylomorph divergences and diversification? 51 52 Clearly, the absence/paucity of sampling from certain time intervals, palaeocontinents, 53 and palaeolatitudes limits our understanding of patterns in biodiversity, both within time 54 bins (e.g. preventing reconstructions of latitudinal biodiversity gradients), and between 55 56 successive intervals (e.g. precluding assessments of temporal fluctuations in biodiversity). In 57 58 59 60 Palaeontology Palaeontology Page 12 of 35

1 2 3 particular, the scarcity of Southern Hemisphere and palaeotropical crocodylomorph 4 occurrences for much of the group’s 230 myr record significantly impacts upon our attempts 5 to elucidate its evolutionary history. Below, we discuss how sampling affects our 6 identification of the timing of several radiations and divergences within Crocodylomorpha. 7 The dearth of sampling of terrestrial crocodylomorphs during the Pliensbachian–Aalenian 8 9 (190.8–170.3 Ma) impedes our understanding of the early evolution of the group, especially 10 the origin of Neosuchia, with an approximately 20 million year gap between the oldest 11 (Calsoyasuchus from the earliest Jurassic [Tykoski et al. 2002]) and second oldest known 12 neosuchian species (Sunosuchus from the Middle Jurassic [Fu et al. 2005]) (note that this 13 excludes the marine clade Thalattosuchia, whose position within Crocodylomorpha is 14 uncertain – see Wilberg 2015). 15 Aside from Chimaerasuchus, from the Aptian–Albian of China (Wu et al. 1995), and 16 17 fragmentary remains (primarily referred to Doratodon) from the Coniacian–Maastrichtian of 18 Europe (Company et al. 2005; Dalla Vecchia and Cau 2011; Rabi and Sebők 2015), 19 notosuchian remains are only known from Gondwana during the Mesozoic. Until recently, 20 no definitive representatives of Notosuchia have been identified from deposits older than 21 the Aptian (Pol and Leardi 2015). However, based on its sister taxon relationship with 22 Neosuchia, which has its oldest known representative in the Early Jurassic (Tykoski et al. 23 24 2002), Notosuchia has an extremely long ghost lineage (~65 myr). If the group was primarily 25 Gondwanan prior to the Aptian, then the limited opportunities to sample crocodylomorphs 26 from the southern continents during this interval (Benson et al. 2013) might be the cause of 27 this ghost lineage. The recent reinterpretation of fragmentary cranial remains (CCM = 7%) 28 from the Bathonian (Middle Jurassic) of Madagascar as the earliest known notosuchian (Dal 29 Sasso et al. 2017) is in keeping with this view. Furthermore, although their ‘sudden’ 30 appearance in the Aptian of Africa (e.g. Sereno and Larsson 2009; O’Connor et al. 2010), 31 32 Asia (Wu et al. 1995), South America (Pol and Leardi 2015), and possibly Australia (Agnolin 33 et al. 2010), might reflect a genuine notosuchian radiation (Pol et al. 2014; Bronzati et al. 34 2015), this seemingly instantaneous widespread distribution is perhaps indicative of an 35 earlier, unsampled diversification. 36 Divergence times within Crocodylia based on molecular data are generally in close 37 agreement with the fossil record. Recent molecular analyses place the split between 38 Alligatoroidea and Crocodyloidea + Gavialoidea at approximately 100–80 Ma (Roos et al. 39 40 2007; Oaks 2011; Green et al. 2014 [though note that most morphological-only analyses 41 place Gavialoidea outside of Crocodyloidea + Alligatoroidea]). Cretaceous crocodylians have 42 currently only been recovered from Laurasia and North Africa: alligatoroids are known from 43 Campanian–Maastrichtian (83.6–66 Ma) deposits (and possibly from the Santonian [86.3– 44 83.6 Ma] too), and crocodyloids and gavialoids both have a Maastrichtian record (e.g. 45 Brochu 2003; Jouve et al. 2008; Martin and Delfino 2010). Given the generally poor early 46 Late Cretaceous terrestrial fossil record (Benson et al. 2013), coupled with low numbers of 47 48 crocodylomorph species and average completeness outside of South America during this 49 interval, possible ghost lineages of <15 myr would not be too surprising. 50 Other splits within Crocodylia have even greater congruence. For example, the first fossil 51 occurrences of alligatorines and caimanines are known from early Paleocene (66–61.6 Ma) 52 deposits (Brochu 2011), and molecular divergence estimates place this alligatoroid split at or 53 close to the K/Pg boundary (e.g. Roos et al. 2007). There are wide error margins for the 54 timing of the Crocodylinae–Tomistominae split (63.7–39.8 Ma; Oaks 2011), but the earliest 55 56 definitive fossils of both clades are from the Ypresian (56–47.8 Ma) (Brochu 2003; Piras et al. 57 58 59 60 Palaeontology Page 13 of 35 Palaeontology

1 2 3 2007). Oaks (2011) inferred the age of the most recent common ancestor of Crocodylus as 4 13.6–8.3 Ma, and the oldest fossil remains attributable to the genus are from the Tortonian 5 (11.6–7.2 Ma) (Brochu 2000; Brochu and Storrs 2012; Delfino and Rossi 2013). Similarly, a 6 9.8–6.7 Ma age was calculated for the most recent common ancestor of Caiman (Oaks 7 2011), and it first appears in the fossil record in the Tortonian (e.g. Salas-Gismondi et al. 8 9 2015). This overall congruence between morphological and molecular data might merely 10 reflect the serendipity of the crocodylian fossil record, i.e. these divergences happened to 11 occur during our short-lived opportunities to sample a particular place in time. 12 The only significant discrepancy between molecular and morphological data pertains to 13 the timing of the Tomistoma–Gavialis split, but in this regard the fossil record of gavialoids 14 (extending back into the Campanian [Brochu 2004]) substantially predates molecular 15 estimates (29–16.1 Ma [Oaks 2011]). Reconciling this discordance is partly dependent on 16 17 resolving whether Gavialis is the sister taxon to all other extant crocodylians (morphological 18 analyses) or to Tomistoma (molecular and combined analyses). However, it would also 19 require the re-interpretation of ‘basal’ gavialoids (e.g. ‘thoracosaurs’) from the latest 20 Cretaceous–Palaeogene as ‘basal’ eusuchians, rather than as close relatives of Gavialidae, 21 for these divergence dates to be correct (e.g. see discussion in Brochu 1997; Harshman et al. 22 2003; Janke et al. 2005; Oaks 2011; Goldman et al. 2014; Lee and Yates 2018). 23 24 25 Extant taxa and the Pull of the Recent 26 27 The Pull of the Recent was originally coined by Raup (1979) to describe the phenomenon 28 whereby the relatively complete sampling of extant taxa extends the stratigraphic ranges of 29 geologically younger taxa to the present day across intervening time intervals in which 30 fossils of those taxa are absent (Jablonski et al. 2003). Although these taxa must have been 31 32 present during these time intervals, and thus would have contributed to diversity, no 33 equivalent extension can be applied to extinct taxa; therefore, range-through data can lead 34 to the artificial appearance of an exponential rise in diversity towards the present day 35 (Foote 2001; Jablonski et al. 2003). Although this particular issue is not pertinent to our 36 study, and most recent studies have suggested that the effect is limited (Jablonski et al. 37 2003; Sahney and Benton 2017), the Pull of the Recent has also been used as a more general 38 term to describe factors that might lead to better sampling of the fossil record in younger 39 40 relative to older rocks. 41 One aspect of this pertains to the identification of extant species in the fossil record. Of 42 the 24 living species of Crocodylia, 15 of these have a putative fossil record, and most of 43 them (13) are present in Quaternary deposits, meaning that there is little influence of the 44 Pull of the Recent on crocodylian diversity in its original meaning. However, the average 45 CCM of fossil remains of these 15 species is 41%, which is considerably lower than the 46 average value for all crocodylomorphs (56%), Cenozoic species (51%), and also Neogene 47 48 taxa (48%). Whereas seven of these 15 species are known from fairly complete fossil 49 skeletons (>70%), the remaining eight species are all known from fossil skeletons that are 50 less than 9% complete (average of <4%). Their low completeness might therefore suggest 51 that these species can be more readily recognized in the fossil record because we have 52 complete skeletons of living members, and thus we only need a small portion of the skull to 53 make an identification. However, an alternative interpretation is that some of these 54 fragmentary fossil remains are attributed to extant taxa primarily because their provenance 55 56 57 58 59 60 Palaeontology Palaeontology Page 14 of 35

1 2 3 closely matches the species’ present-day distribution, rather than through shared 4 autapomorphies. 5 6 CONCLUSIONS 7

8 9 The fossil record of crocodylomorphs is heavily biased through both time and space, 10 especially in the palaeotropics. In general, Mesozoic species are known from more complete 11 fossil specimens than their Cenozoic counterparts, with an overall trend of decreasing 12 average completeness through time. Our understanding of nearly all diversification and 13 extinction ‘events’ in crocodylomorph evolutionary history is essentially driven by regional 14 patterns, with no global sampling signal. These spatiotemporal biases might also explain 15 several long phylogenetic ghost lineages, including the origin and radiation of Notosuchia. 16 17 The close congruence between the fossil record and crocodylian divergence dates based on 18 molecular data might merely be fortuitous, i.e. divergences happened to occur during our 19 ephemeral spatiotemporal sampling windows. 20 21 ACKNOWLEDGEMENTS 22 23 24 We are grateful for the efforts of all those who have generated crocodylomorph fossil data, 25 as well as those who have entered this data into the Paleobiology Database (especially John 26 Alroy, Matthew Carrano, Anna Behrensmeyer, and Mark Uhen). We also thank all those who 27 facilitated collections access for the compilation of crocodylomorph body size data. 28 Jonathan Tennant also provided us with information on the preserved skeletons of several 29 fossil crocodylomorphs. Silhouettes of representative crocodylomorphs in figures are 30 modified from illustrations by Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted 31 32 at Phylopic (http://phylopic.org), where all license information is available. Comments from 33 Mario Bronzati, Sally Thomas, and an anonymous reviewer improved an earlier version of 34 this work. PDM was supported by a Leverhulme Trust Early Career Fellowship (ECF-2014- 35 662) and a Royal Society University Research Fellowship (UF160216). AAC was supported by 36 an Imperial College London Janet Watson Departmental PhD Scholarship. PLG was 37 supported by a University of Birmingham-CAPES Joint PhD Scholarship (3581-14-4). This is 38 Paleobiology Database official publication number XXX. 39 40 41 DATA ARCHIVING STATEMENT 42 43 Data used in this study are available in the Dryad Digital Repository: 44 https://doi.org/10.5061/dryad.668950m. 45 46 REFERENCES 47 48 49 AGNOLIN, F. L., EZCURRA, M. D., PAIS, D. F. and SALISBURY, S. W. 2010. A reappraisal of the 50 Cretaceous non-avian dinosaur faunas from Australia and New Zealand: evidence for their 51 Gondwanan affinities. Journal of Systematic Palaeontology, 8, 257–300. 52 ALI, J. R. and AITCHISON, J. C. 2008. Gondwana to Asia: plate tectonics, paleogeography and 53 the biological connectivity of the Indian sub-continent from the Middle Jurassic through 54 latest Eocene (166–35 Ma). Earth Science Reviews, 88, 145–166. 55 56 57 58 59 60 Palaeontology Page 15 of 35 Palaeontology

1 2 3 ALLISON, P. A. and BRIGGS, D. E. G. 1993. Paleolatitudinal sampling bias, Phanerozoic 4 species diversity, and the end-Permian extinction. Geology, 21, 65–68. 5 ALROY, J. 2010. Geographical, environmental and intrinsic biotic controls on Phanerozoic 6 marine diversification. Palaeontology, 53, 1211–1235. 7 ALROY, J. 2013. Online paleogeographic map generator. http://paleodb.org/?a=mapForm. 8 9 BARNOSKY, A. D., CARRASCO, M. A. and DAVIS, E. B. 2005. The impact of the species-area 10 relationship on estimates of paleodiversity. PLoS Biology, 3, e266. 11 BATEZELLI, A. 2017. Continental systems tracts of the Brazilian Cretaceous Bauru Basin and 12 their relationship with the tectonic and climatic evolution of South America. Basin 13 Research, 29, 1–25. 14 BEARDMORE, S. R. 2017. The effect of body size on skeletal articulation and completeness in 15 Jurassic Ichthyopterygia (Reptilia). Paläontologische Zeitschrift, 91, 577–587. 16 17 BEARDMORE, S. R., ORR, P. J., MANZOCCHI, T. and FURRER, H. 2012a. Float or sink: 18 modelling the taphonomic pathway of marine crocodiles (Mesoeucrocodylia, 19 Thalattosuchia) during the death–burial interval. Palaeobiodiversity and 20 Palaeoenvironments, 92, 83–98. 21 BEARDMORE, S. R., ORR, P. J., MANZOCCHI, T., FURRER, H. and JOHNSON, C. 2012b. Death, 22 decay and disarticulation: Modelling the skeletal taphonomy of marine reptiles 23 24 demonstrated using Serpianosaurus (Reptilia; Sauropterygia). Palaeogeography, 25 Palaeoclimatology, Palaeoecology, 337–338, 1–13. 26 BEHRENSMEYER, A. K., KIDWELL, S. M. and GASTALDO, R. A. 2000. Taphonomy and 27 paleobiology. Paleobiology, 26, 103–147. 28 BELL, M. A. and LLOYD, G. T. 2015. strap: an R package for plotting phylogenies against 29 stratigraphy and assessing their stratigraphic congruence. Palaeontology, 58, 379–389. 30 BENSON, R. B. J., MANNION, P. D., BUTLER, R. J., UPCHURCH, P., GOSWANI, A. and EVANS, S. 31 32 E. 2013. Cretaceous tetrapod fossil record sampling and faunal turnover: implications for 33 biogeography and the rise of modern clades. Palaeogeography, Palaeoclimatology, 34 Palaeoecology, 372, 88–107. 35 BENSON, R. B. J., BUTLER, R. J., ALROY, J., MANNION, P. D., CARRANO, M. T. and LLOYD, G. T. 36 2016. Near-stasis in the long-term diversification of Mesozoic tetrapods. PLoS Biology, 14, 37 e1002359. 38 BENTON, M. J. 2008. How to find a dinosaur, and the role of synonymy in biodiversity 39 40 studies. Paleobiology, 34, 516–533. 41 BOTTJER, D. J. and JABLONSKI, D. 1988. Paleoenvironmental patterns in the evolution of 42 post-Paleozoic benthic marine invertebrates. Palaios, 3, 540–560. 43 BRAND, L. R., HUSSEY, M. and TAYLOR, J. 2003. Decay and disarticulation of small 44 vertebrates in controlled experiments. Journal of Taphonomy, 1, 69–95. 45 BROCHU, C. A. 1997. Morphology, fossils, divergence timing, and the phylogenetic 46 relationships of Gavialis. Systematic Biology, 46, 479–522. 47 48 BROCHU, C. A. 2000. Phylogenetic relationships and divergence timing of Crocodylus based 49 on morphology and the fossil record. Copeia, 2000, 657–673. 50 BROCHU, C. A. 2003. Phylogenetic approaches towards crocodylian history. Annual Review 51 of Earth and Planetary Sciences, 31, 357–397. 52 BROCHU, C. A. 2004. A new Late Cretaceous gavialoid crocodylian from eastern North 53 America and the phylogenetic relationships of thoracosaurs. Journal of Vertebrate 54 Paleontology, 24, 610–633. 55 56 57 58 59 60 Palaeontology Palaeontology Page 16 of 35

1 2 3 BROCHU, C. A. 2011. Phylogenetic relationships of Necrosuchus ionensis Simpson, 1937 and 4 the early history of caimanines. Zoological Journal of the Linnean Society, 163, S228–S256. 5 BROCHU, C. A. and STORRS, G. W. 2012. A giant crocodile from the Plio-Pleistocene of 6 Kenya, the phylogenetic relationships of Neogene African crocodylines, and the antiquity 7 of Crocodylus in Africa. Journal of Vertebrate Paleontology, 32, 587–602. 8 9 BROCKLEHURST, N. and FRÖBISCH, J. 2014. Current and historical perspectives on the 10 completeness of the fossil record of pelycosaurian-grade synapsids. Palaeogeography, 11 Palaeoclimatology, Palaeoecology, 399, 114–126. 12 BROCKLEHURST, N., UPCHURCH, P., MANNION, P. D. and O’CONNOR, J. 2012. The 13 completeness of the fossil record of Mesozoic birds: implications for early avian evolution. 14 PLoS ONE, 7, e39056. 15 BRONZATI, M., MONTEFELTRO, F. C. and LANGER, M. C. 2015. Diversification events and the 16 17 effects of mass extinctions on Crocodyliformes evolutionary history. Royal Society Open 18 Science, 2, 140385. 19 BROWN, C. M., EVANS, D. C., CAMPIONE, N. E., O’BRIEN, L. J. and EBERTH, D. A. 2013. 20 Evidence for Taphonomic Size Bias in a Model Mesozoic Terrestrial Alluvial-Paralic System. 21 Palaeogeography, Palaeoclimatology, Palaeoecology, 372, 108–122. 22 BUSH, A. M. and BAMBACH, R. K. 2004. Did Alpha Diversity Increase during the 23 24 Phanerozoic? Lifting the Veils of Taphonomic, Latitudinal, and Environmental Biases. 25 Journal of Geology, 112, 625–642. 26 BUSH, A. M., MARKEY, M. J. and MARSHALL, C. R. 2004. Removing bias from diversity 27 curves: the effects of spatially organized biodiversity on sampling-standardization. 28 Paleobiology, 30, 666–686. 29 CARVALHO, I. S., GASPARINI, Z. B., SALGADO, L., DE VASCONCELLOS, F. M. and MARINHO, T. 30 S. 2010. Climate’s role in the distribution of the Cretaceous terrestrial Crocodyliformes 31 32 throughout Gondwana. Palaeogeography, Palaeoclimatology, Palaeoecology, 297, 252– 33 262. 34 CLEARY, T. J., MOON, B. C., DUNHILL, A. M. and BENTON, M. J. 2015. The fossil record of 35 ichthyosaurs, completeness metrics and sampling biases. Palaeontology, 58, 521–536. 36 CLOSE, R. A., BENSON, R. B. J., UPCHURCH, P. and BUTLER, R. J. 2017. Controlling for the 37 species-area effect supports constrained long-term Mesozoic terrestrial vertebrate 38 diversification. Nature Communications, 8, 15381. 39 40 COHEN, K. M., FINNEY, S. C., GIBBARD, P. L. and FAN, J.-X. 2013. The ICS International 41 Chronostratigraphic Chart. Episodes, 36, 199–204 [updated here: 42 http://www.stratigraphy.org/ICSchart/ChronostratChart2017-02.pdf]. 43 COMPANY, J., PEREDA SUBERBIOLA, X., RUIZ-OMEÑACA, J. I. and BUSCALIONI A. D. 2005. A 44 new species of Doratodon (Crocodyliformes: Ziphosuchia) from the Late Cretaceous of 45 Spain. Journal of Vertebrate Paleontology, 25, 343–353. 46 DALLA VECCHIA, F. M. and CAU, A. 2011. The first record of a notosuchian crocodyliform 47 48 from Italy. Rivista Italiana di Paleontologia e Stratigrafia, 117, 309–321. 49 DAL SASSO, C., PASINI, G., FLEURY, G. and MAGANUCO S. 2017. Razanandrongobe 50 sakalavae, a gigantic mesoeucrocodylian from the Middle Jurassic of Madagascar, is the 51 oldest known notosuchian. PeerJ, 5, e3481 52 DAVIES, T., BELL, M., GOSWAMI, A. and HALLIDAY, T. J. D. 2017. Completeness of the 53 eutherian mammal fossil record and implications for reconstructing mammal evolution 54 through the Cretaceous-Palaeogene mass extinction. Paleobiology, 43, 521–536. 55 56 57 58 59 60 Palaeontology Page 17 of 35 Palaeontology

1 2 3 DEAN, C. D., MANNION, P. D. and BUTLER, R. J. 2016. Preservational bias controls the fossil 4 record of pterosaurs. Palaeontology, 59, 225–247. 5 DELFINO, M. and ROSSI, M. A. 2013. Fossil crocodylid remains from Scontrone (Tortonian, 6 Southern Italy) and the late Neogene Mediterranean biogeography of crocodylians. 7 Geobios, 46, 25–31. 8 9 DRISCOLL, D. A., DUNHILL, A. M., STUBBS, T. L. and BENTON, M. J. 2018. The mosasaur fossil 10 record through the lens of fossil completeness. Palaeontology (doi: 10.1111/pala.12381). 11 DYKE, G. J., McGOWAN, A., NUDDS, R. and SMITH, D. 2009. The shape of pterosaur 12 evolution: evidence from the fossil record. Journal of Evolutionary Biology, 22, 890–898. 13 ESTES, R. and HUTCHISON, J. H. 1980. Eocene lower vertebrates from Ellesmere Island, 14 Canadian Arctic Archipelago. Palaeogeography, Palaeoclimatology, Palaeoecology, 30, 15 325–347. 16 17 FARLOW, J. O., HURLBURT, G. R., ELSEY, R. M., BRITTON, A. R. and LANGSTON JR, W. 2005. 18 Femoral dimensions and body size of Alligator mississippiensis: estimating the size of 19 extinct mesoeucrocodylians. Journal of Vertebrate Paleontology, 25, 354–369. 20 FOOTE, M. 2001. Inferring temporal patterns of preservation, origination, and extinction 21 from taxonomic survivorship analysis. Paleobiology, 27, 602–630. 22 FOUNTAINE, T. M. R., BENTON, M. J., NUDDS, R. L. and DYKE, G. J. 2005. The quality of the 23 24 fossil record of Mesozoic birds. Proceedings of the Royal Society of London B, 272, 289– 25 294. 26 FU, Q.-M., MING, S.-Y. and PENG, G.-Z. 2005. A new species of Sunosuchus from Zigong, 27 Sichuan, China. Vertebrata PalAsiatica, 43, 76–83. 28 GODOY, P. L., BRONZATI, M., ELTINK, E., MARSOLA, J. C. A., CIDADE, G. M., LANGER, M. C. 29 and MONTEFELTRO, F. C. 2016. Postcranial anatomy of Pissarrachampsa sera 30 (Crocodyliformes, Baurusuchidae) from the Late Cretaceous of Brazil: insights on lifestyle 31 32 and phylogenetic significance. PeerJ, 4, e2075. 33 GOLD, M. E. L., BROCHU, C. A. and NORELL, M. A. 2014. An Expanded Combined Evidence 34 Approach to the Gavialis Problem Using Geometric Morphometric Data from Crocodylian 35 Braincases and Eustachian Systems. PLoS ONE, 9, e105793. 36 GREEN, R. E., BRAUN, E. L., ARMSTRONG, J., EARL, D., NGUYEN, N., HICKEY, G., et al. 2014. 37 Three crocodilian genomes reveal ancestral patterns of evolution among archosaurs. 38 Science, 346, 1254449. 39 40 HARSHMAN, J., HUDDLESTON, C. J., BOLLBACK, J. P., PARSONS, T. J. and BRAUN, M. J. 2003. 41 True and false gharials: a nuclear gene phylogeny of Crocodylia. Systematic Biology, 52, 42 386–402. 43 HILL, A. and BEHRENSMEYER, A. K. 1984, Disarticulation patterns of some modern East 44 African mammals. Paleobiology, 10, 366–376. 45 HURLBURT, G. R., HECKERT, A. B. and FARLOW, J. O. 2003. Body mass estimates of 46 phytosaurs (Archosauria: Parasuchidae) from the Petrified Forest Formation (Chinle Group: 47 48 Revueltian) based on skull and limb bone measurements. New Mexico Museum of Natural 49 History and Science Bulletin, 24, 105–113. 50 IRMIS, R. B., NESBITT, S. J. and SUES, H-D. 2013. Early Crocodylomorpha. Geological Society, 51 London, Special Publications, 379, 275–302. 52 JABLONSKI, D., ROY, K., VALENTINE, J. W., PRICE, R. M. and ANDERSON, P. S. 2003. The 53 Impact of the Pull of the Recent on the History of Marine Diversity. Science, 300, 1133– 54 1135. 55 56 57 58 59 60 Palaeontology Palaeontology Page 18 of 35

1 2 3 JACKSON, J. B. C. and JOHNSON, K. G. 2001. Measuring Past Biodiversity. Science, 293, 2401– 4 2404. 5 JANKE, A., GULLBERG, A., HUGHES, S., AGGARWAL, R. K. and ARNASON, U. 2005. 6 Mitogenomic analyses place the gharial (Gavialis gangeticus) on the crocodile tree and 7 provide pre-K/T divergence times for most crocodilians. Journal of Molecular Evolution, 61, 8 9 620–626. 10 JOUVE, S., BARDET, N., JALIL, N.-E., PEREDA SUBERBIOLA, X., BOUYA, B. and AMAGHZAZ, M. 11 2008. The oldest African crocodylian: phylogeny, paleobiogeography, and differential 12 survivorship of marine reptiles through the Cretaceous–Tertiary Boundary. Journal of 13 Vertebrate Paleontology, 28, 409–421. 14 KEMP, M. E. and HADLY, E. A. 2016. Rocking Earth’s biodiversity cradle: challenges, 15 advances, and prospects for conservation paleontology in the tropics. Journal of 16 17 Vertebrate Paleontology, 36, e1179640. 18 KIDWELL, S. M. and HOLLAND, S. M. 2002. The Quality of the Fossil Record: Implications for 19 Evolutionary Analyses. Annual Review of Ecology and Systematics, 33, 561–588. 20 LEE, M. S. Y. and YATES, A. M. 2018. Tip-dating and homoplasy: reconciling the shallow 21 molecular divergences of modern gharials with their long fossil record. Proceedings of the 22 Royal Society of London, Series B, 285, 20181071. 23 24 MANNION, P. D. and UPCHURCH, P. 2010. Completeness metrics and the quality of the 25 sauropodomorph fossil record through geological and historical time. Paleobiology, 36, 26 283–302. 27 MANNION, P. D., BENSON, R. B. J., UPCHURCH, P., BUTLER, R. J., CARRANO, M. T. and 28 BARRETT, P. M. 2012. A temperate palaeodiversity peak in Mesozoic dinosaurs and 29 evidence for Late Cretaceous geographical partitioning. Global Ecology and Biogeography, 30 21, 898–908. 31 32 MANNION, P. D., BENSON, R. B. J., CARRANO, M. T., TENNANT, J. P., JUDD, J. and BUTLER, R. 33 J. 2015. Climate constrains the history and biodiversity of crocodylians. Nature 34 Communications, 6, 8438. 35 MANNION, P. D., CHIARENZA, A. A., GODOY, P. L. and CHEAH, Y. N. 2018a. Data from: 36 Spatiotemporal sampling patterns in the 230 million year fossil record of terrestrial 37 crocodylomorphs and their impact on diversity. Dryad Digital Repository. 38 https://doi.org/10.5061/dryad.668950m. 39 40 MANNION, P. D., CARRANO, M. T., ALROY, J. and BEHRENSMEYER, A. K. 2018b. Taxonomic 41 occurrences of Crocodylomorpha recorded in the Paleobiology Database. Fossilworks. 42 http://fossilworks.org. 43 MARCOT, J. D., FOX, D. L. and NIEBUHR, S. R. 2016. Late Cenozoic onset of the latitudinal 44 diversity gradient of North American mammals. Proceedings of the National Academy of 45 Sciences, USA, 113, 7189–7194. 46 MARKWICK, P. J. 1994. “Equability”, continentality, and Tertiary “climate”: The crocodilian 47 48 perspective. Geology, 22, 613–616. 49 MARKWICK, P. J. 1998. Crocodilian diversity in space and time: the role of climate in 50 paleoecology and its implication for understanding K/T extinctions. Paleobiology, 24, 470– 51 497. 52 MARTIN, J. E. and DELFINO, M. 2010. Recent advances in the comprehension of the 53 biogeography of Cretaceous European eusuchians. Palaeogeography, Palaeoclimatology, 54 Palaeoecology, 293, 406–418. 55 56 57 58 59 60 Palaeontology Page 19 of 35 Palaeontology

1 2 3 MARTIN, J. E., AMIOT, R., LÉCUYER, C. and BENTON, M. J. 2014. Sea surface temperature 4 contributes to marine crocodylomorph evolution. Nature Communications, 5, 4658. 5 MARTIN, J. E., DELFINO, M. and SMITH, T. 2016. Osteology and affinities of Dollo's 6 goniopholidid (Mesoeucrocodylia) from the Early Cretaceous of Bernissart, Belgium. 7 Journal of Vertebrate Paleontology, 36, e1222534. 8 9 MATSUMOTO, R. and EVANS, S. E. 2010. Choristoderes and the freshwater assemblages of 10 Laurasia. Journal of Iberian Geology, 36, 253–274. 11 MONTEFELTRO, F. C., LARSSON, H. C. E., DE FRANCA, M. A. G. and LANGER, M. C. 2013. A 12 new neosuchian with Asian affinities from the Jurassic of northeastern Brazil. 13 Naturwissenschaften, 100, 835–841. 14 NOTO, C. 2011. Hierarchical Control of Terrestrial Vertebrate Taphonomy Over Space and 15 Time: Discussion of Mechanisms and Implications for Vertebrate Paleobiology. In ALLISON, 16 nd 17 P. A. and BOTTJER, D. J. (eds). Taphonomy: Process and Bias Through Time (2 edition). 18 Springer Netherlands, 287–336. 19 OAKS, J. R. 2011. A time-calibrated species tree of Crocodylia reveals a recent radiation of 20 true crocodiles. Evolution, 65, 3285–3297. 21 O’CONNOR, P. M., SERTICH, J. J. W., STEVENS, N. J., ROBERTS. E. M., GOTTFRIED, M. D., 22 HIERONYMUS, T. L., JINNAH, Z. A., RIDGELY, R., NGASALA, S. E. and TEMBA, J. 2010. The 23 24 evolution of mammal-like crocodyliforms in the Cretaceous Period of Gondwana. Nature, 25 466, 748–751. 26 PIRAS, P., DELFINO, M., DEL FAVERO, L. and KOTSAKIS, T. 2007. Phylogenetic position of the 27 crocodylian Megadontosuchus arduini and tomistomine palaeobiogeography. Acta 28 Palaeontologica Polonica, 52, 315–328. 29 POL, D. and LEARDI, J. M. 2015. Diversity patterns of Notosuchia (Crocodyliformes, 30 Mesoeucrocodylia) during the Cretaceous of Gondwana. Publicación Electrónica de la 31 32 Asociación Paleontológica Argentina, 15, 172–186. 33 POL, D., LEARDI, J. M., LECUONA, A. and KRAUSE, M. 2012. Postcranial anatomy of Sebecus 34 icaeorhinus (Crocodyliformes, Sebecidae) from the Eocene of Patagonia. Journal of 35 Vertebrate Paleontology, 32, 328–354. 36 POL, D., NASCIMENTO, P., CARVALHO, M., RICCOMINI, A. B., PIRES-DOMINGUES, R. A. and 37 ZAHER, H. 2014. A new notosuchian from the Late Cretaceous of Brazil and the phylogeny 38 of advanced notosuchians. PLoS ONE, 9, e93105. 39 40 RABI, M. and SEBŐK, N. 2015. A revised Eurogondwana model: Late Cretaceous notosuchian 41 crocodyliforms and other vertebrate taxa suggest the retention of episodic faunal links 42 between Europe and Gondwana during most of the Cretaceous. Gondwana Research, 28, 43 1197–1211. 44 RAUP, D. M. 1972. Taxonomic diversity during the Phanerozoic. Science, 177, 1065–1071. 45 RAUP, D. M. 1979. Biases in the fossil record of species and genera. Bulletin of Carnegie 46 Museum of Natural History, 13, 85–91. 47 48 R CORE TEAM. 2017. R: A language and environment for statistical computing, v. 3.4.3. R 49 Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/ 50 REES, P. M., NOTO, C. R., PARRISH, J. M. and PARRISH, J. T. 2004. Late Jurassic climates, 51 vegetation, and dinosaur distributions. Journal of Geology, 112, 643–653. 52 ROOS, J., AGGARWAL, R. K. and JANKE, A. 2007. Extended mitogenomic phylogenetic 53 analyses yield new insight into crocodylian evolution and their survival of the Cretaceous- 54 Tertiary boundary. Molecular Phylogenetics and Evolution, 45, 663–673. 55 56 57 58 59 60 Palaeontology Palaeontology Page 20 of 35

1 2 3 SAHNEY, S. and BENTON, M. J. 2017. The impact of the Pull of the Recent on the fossil 4 record of tetrapods. Evolutionary Ecology Research, 18, 7–23. 5 SALAS-GISMONDI, R., FLYNN, J. J., BABY, P., TEJADA-LARA, J. V., WESSELINGH, F. P. and 6 ANTOINE, P.-O. 2015. A Miocene hyperdiverse crocodylian community reveals peculiar 7 trophic dynamics in proto-Amazonian mega-wetlands. Proceedings of the Royal Society of 8 9 London, Series B, 282, 20142490. 10 SALISBURY, S. W., MOLNAR, R. E., FREY, E. and WILLIS, P. M. A. 2006. The origin of modern 11 crocodyliforms: new evidence from the Cretaceous of Australia. Proceedings of the Royal 12 Society of London, Series B, 273, 2439–2448. 13 SCHEYER, T. M., AGUILERA, O. A., DELFINO, M., FORTIER, D. C., CARLINI, A. A., SÁNCHEZ, R., 14 CARRILLO-BRICEÑO, J. D., QUIROZ, L. and SÁNCHEZ-VILLAGRA, M. R. 2013 Crocodylian 15 diversity peak and extinction in the late Cenozoic of the northern Neotropics. Nature 16 17 Communications, 4, 1907. 18 SELLWOOD, B. W. and VALDES, P. J. 2006. Mesozoic climates: general circulation models and 19 the rock Record. Sedimentary Geology, 190, 269–287. 20 SERENO, P. C. and LARSSON, H. C. 2009. Cretaceous Crocodyliforms from the Sahara. 21 Zookeys, 28, 1–143. 22 SMITH, A. B. and MCGOWAN, A. J. 2007. The shape of the Phanerozoic marine 23 24 palaeodiversity curve: how much can be predicted from the sedimentary rock record of 25 western Europe? Palaeontology, 50, 765–774. 26 TENNANT, J. P., MANNION, P. D. and UPCHURCH, P. 2016. Environmental drivers of 27 crocodyliform extinction across the Jurassic/Cretaceous interval. Proceedings of the Royal 28 Society of London, Series B, 283, 20152840. 29 TUTIN, S. L. and BUTLER, R. J. 2017. The completeness of the fossil record of plesiosaurs, 30 marine reptiles from the Mesozoic. Acta Palaeontologica Polonica, 62, 563–573. 31 32 TYKOSKY, R. S., ROWE, T. B., KETCHAM, R. A. and COLBERT, M. W. 2002. Calsoyasuchus 33 valliceps, a new crocodyliform from the early Jurassic Kayenta Formation of Arizona. 34 Journal of Vertebrate Paleontology, 22, 593–611. 35 VERRIÈRE, A., BROCKLEHURST, N. and FRÖBISCH, J. 2016. Assessing the completeness of the 36 fossil record: comparison of different methods applied to parareptilian tetrapods 37 (Vertebrata: Sauropsida). Paleobiology, 42, 680–695. 38 VILHENA, D. A. and SMITH, A. B. 2013. Spatial Bias in the Marine Fossil Record. PLoS ONE, 8, 39 40 e74470. 41 WALTHER, M. and FRÖBISCH, J. 2013. The quality of the fossil record of anomodonts 42 (Synapsida, Therapsida). Comptes Rendus Paleovol, 12, 495–504. 43 WILBERG, E. W. 2015. What’s in an Outgroup? The Impact of Outgroup Choice on the 44 Phylogenetic Position of Thalattosuchia (Crocodylomorpha) and the Origin of 45 Crocodyliformes. Systematic Biology, 64, 621–637. 46 WILBERG, E. W. 2017. Investigating patterns of crocodyliform cranial disparity through the 47 48 Mesozoic and Cenozoic. Zoological Journal of the Linnean Society, 181, 189–208. 49 WILLIS, P. M. A. and STILWELL, J. D. 2000. A probable piscivorous crocodile from Eocene 50 deposits of McMurdo Sound, East Antarctica. Antarctic Research Series, 76, 355–358. 51 WU, X.-C., SUES, H.-D. and SUN, A. 1995. A plant-eating crocodyliform reptile from the 52 Cretaceous of China. Nature, 376, 678–680. 53 ZACHOS, J. C., DICKENS, G. R. and ZEEBE, R. E. 2008. An early Cenozoic perspective on 54 greenhouse warming and carbon-cycle dynamics. Nature, 451, 279–283. 55 56 57 58 59 60 Palaeontology Page 21 of 35 Palaeontology

1 2 3 ZIEGLER, A. M., ESHEL, G., REES, P. M., ROTHFUS, T. A., ROWLEY, D. B. and SUNDERLIN, D. 4 2003. Tracing the tropics across land and sea: Permian to present. Lethaia, 36, 227–254. 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Palaeontology Palaeontology Page 22 of 35

1 2 3 FIGURES 4 5 FIGURE 1. Skeletal outline of a representative crocodylomorph (Deinosuchus), showing the 6 body regions used to partition the skeleton. Illustration by Scott Hartman, used with 7 permission. 8 9 10 FIGURE 2. Box-and-whisker plots showing the distribution of crocodylomorph completeness 11 for different time intervals. 12 13 FIGURE 3. Global patterns of crocodylomorph completeness, number of species, and 14 number of crocodylomorph-bearing collections (CBCs) plotted through time (using stage- 15 level time bins). Silhouettes of representative crocodylomorphs are modified from 16 17 illustrations by Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at Phylopic 18 (http://phylopic.org), where all license information is available. 19 20 FIGURE 4. Global patterns of crocodylomorph completeness, number of species, and 21 number of crocodylomorph-bearing collections (CBCs) plotted through time (using ~9 myr 22 time bins). Silhouettes of representative crocodylomorphs are modified from illustrations by 23 24 Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at Phylopic 25 (http://phylopic.org), where all license information is available. 26 27 FIGURE 5. Box-and-whisker plots showing the distribution of crocodylomorph completeness 28 for each palaeocontinent. 29 30 FIGURE 6. Global palaeogeographic reconstructions showing the distribution of non-marine 31 32 crocodylomorph fossil occurrences through time for the: (A) Late Triassic (plate 33 reconstruction age = 220 Ma); (B) Early Jurassic (200 Ma); (C) Middle Jurassic (170 Ma); (D) 34 Late Jurassic (150 Ma); (E) Early Cretaceous (120 Ma); (F) Late Cretaceous (80 Ma); (G) 35 Paleocene (60 Ma); (H) Eocene (45 Ma); (I) Oligocene (30 Ma); (J) Miocene (15 Ma); (K) 36 Pliocene (4 Ma); (L) Quaternary (1 Ma). Global palaeogeographic reconstructions from 37 Fossilworks (http://fossilworks.org/ [Alroy 2013]) based on data in the Paleobiology 38 Database (https://paleobiodb.org/) (Mannion et al. 2018b). Ages in parentheses refer to the 39 40 palaeogeographic reconstruction date. 41 42 FIGURE 7. Palaeocontinental patterns of crocodylomorph completeness, number of species, 43 and number of crocodylomorph-bearing collections (CBCs) plotted through time (using ~9 44 myr time bins) for: (A) Africa; (B) Asia; (C) Australasia; (D) Europe; (E) North America; (F) 45 South America. Silhouettes of representative crocodylomorphs are modified from 46 illustrations by Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at Phylopic 47 48 (http://phylopic.org), where all license information is available. 49 50 FIGURE 8. Crocodylomorph completeness, number of crocodylomorph species, and number 51 of crocodylomorph-bearing collections plotted through space, using 10° palaeolatitudinal 52 bands. Silhouettes of representative crocodylomorphs are modified from illustrations by 53 Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at Phylopic 54 (http://phylopic.org), where all license information is available. 55 56 57 58 59 60 Palaeontology Page 23 of 35 Palaeontology

1 2 3 FIGURE 9. Palaeolatitudinal patterns of crocodylomorph completeness, number of species, 4 and number of crocodylomorph-bearing collections (CBCs) plotted through time (using ~9 5 myr time bins) for: (A) 50–60° N; (B) 40–50° N; (C) 30–40° N; (D) 20–30° N; (E) 10–20° N; (F) 6 0–10° N; (G) 0–10° S; (H) 10–20° S; (I) 20–30° S; (J) 30–40° S; (K) 40–50° S. Silhouettes of 7 representative crocodylomorphs from selected palaeolatitudinal regions are modified from 8 9 illustrations by Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at Phylopic 10 (http://phylopic.org), where all license information is available. 11 12 FIGURE 10. Global patterns in total length (TL) of species plotted against completeness for: 13 (A) all species (TL untransformed); (B) all species; (C) species smaller than 2 m (size class A); 14 (D) species between 2–4 m (size class B); (E) species between 4–6 m (size class C); (F) species 15 larger than 6 m (size class D). TL is logged in plots B–F. 16 17 18 FIGURE 11. Box-and-whisker plots showing the distribution of completeness for selected 19 crocodylomorph subclades. Silhouettes of representative crocodylomorphs are modified 20 from illustrations by Evan Boucher, Scott Hartman, and Nobumichi Tamura, hosted at 21 Phylopic (http://phylopic.org), where all license information is available. 22 23 24 TABLES 25 26 TABLE 1. Correlation test results for time series comparisons. 27 28 Comparison Raw data Generalised differencing 29 CCM vs TDE (stage; global) rs = -0.30 p = 0.06 rs = 0.04 p = 0.82 30 31 CCM vs TDE (9 myr; global) rs = -0.14 p = 0.53 rs = 0.34 p = 0.11 32 CCM vs CBCs (stage; global) rs = -0.24 p = 0.13 rs = 0.04 p = 0.81 33 CCM vs CBCs (9 myr; global) rs = -0.04 p = 0.84 rs = 0.47 p = 0.03* 34 TDE vs CBCs (stage; global) rs = 0.86 p = 9.06e-13* rs = 0.78 p = 7.73e-8* 35 TDE vs CBCs (9 myr; global) rs = 0.84 p = 3.26e-7* rs = 0.72 p = 0.0002* 36 CCM vs TDE (9 myr; SAm) rs = 0.55 p = 0.01* rs = 0.69 p = 0.002* 37 CCM vs CBCs (9 myr; SAm) rs = 0.21 p = 0.37 rs = 0.50 p = 0.03* 38 39 40 Statistical comparisons between temporal fluctuations in crocodylomorph completeness 41 (CCM), diversity (TDE), and numbers of crocodylomorph-bearing collections (CBCs) for 42 stage-level and 9 myr bin-level analyses. Results presented for global and South American 43 (SAm) analyses. Spearman's rank correlation coefficients (rs) were calculated for each 44 pairwise comparison based on the raw data and after generalized differencing. *Significant 45 46 at alpha (p-value) = 0.05. 47 48 49 TABLE 2. Correlation test results for palaeolatitudinal data series comparisons. 50 51 Comparison Raw data Generalised differencing 52 53 Completeness vs TDE rs = 0.04 p = 0.90 rs = 0.19 p = 0.61 54 Completeness vs CBCs rs = 0.88 p = 0.0004* rs = 0.77 p = 0.01* 55 TDE vs CBCs rs = 0.25 p = 0.45 rs = 0.56 p = 0.10 56 57 58 59 60 Palaeontology Palaeontology Page 24 of 35

1 2 3 Statistical comparisons between the palaeolatitudinal distribution of crocodylomorph 4 completeness, diversity (TDE), and numbers of crocodylomorph-bearing collections (CBCs) 5 for 10° palaeolatitudinal bands. Spearman's rank correlation coefficients (rs) were calculated 6 for each pairwise comparison based on the raw data and after generalized differencing. 7 *Significant at alpha (p-value) = 0.05. 8 9 10 11 TABLE 3. Regression results for completeness and total length of species. 12 13 Dataset N Intercept Slope p-value Adjusted R2 14 15 All species (TL untransformed) 363 3.055 -0.006 0.034* 0.009 16 17 All species (log10 TL) 363 0.388 -0.001 0.003* 0.021 18 19 Category A (log10 TL) 167 0.05 -0.001 0.038* 0.019 20 Category B (log10 TL) 131 0.48 -0.0003 0.075 0.016 21 22 Category C (log10 TL) 39 0.701 -0.0003 0.097 0.047 23 24 Category D (log10 TL) 26 0.927 -0.0004 0.406 -0.011 25 26 27 Results of OLS regressions of average crocodylomorph completeness on total length (TL) of 28 species, using different datasets: including all species and untransformed TL values; all 29 species and log-transformed TL values; species within size category A (smaller than 2 30 metres) and log10 TL; species within size category B (2 to 4 metres) and log10 TL; species 31 within size category C (4 to 6 metres) and log10 TL; species within size category D (larger 32 than 6metres) and log TL. *Significant at alpha (p-value) = 0.05. 33 10 34 35 36 TABLE 4. Correlation test results for clade comparisons. 37 38 Comparison Raw data Generalised differencing 39 Crocodylomorphs vs pterosaurs rs = 0.75 p = 5.23e-5* rs = 0.26 p = 0.25 40 41 Crocodylomorphs vs sauropodomorphs rs = 0.51 p = 0.02* rs = -0.10 p = 0.66 42 43 Statistical comparisons between global temporal fluctuations in Mesozoic crocodylomorph, 44 pterosaur, and sauropodomorph dinosaur completeness. Spearman's rank correlation 45 coefficients (rs) were calculated for each pairwise comparison based on the raw data and 46 after generalized differencing. *Significant at alpha (p-value) = 0.05. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Palaeontology Page 25 of 35 Palaeontology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 skull cervicals dorsals pelvic girdle osteoderms

pectoral girdle

caudals forelimbs hindlimbs

Palaeontology Palaeontology Page 26 of 35 100 100 1 2 3 4 5

8080 6 7

s 8 s 9 e n 10 e t 11 e l 60

p 12

m 13 o

c 14 f 15 o

e 16 g 17 a t 060 18 4040 n e

19 Percentage completeness c r 20 e

P 21 22 23

24 2020 25 26 27 28 29 30 0 0 31 32 All Mesozoic Cenozoic Triassic Jurassic Cretaceous Palaeogene Neogene + 33 All Mesozoic Cenozoic Triassic Jurassic Cretaceous Paleogene Quaternary 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Palaeontology Page 27 of 35 Palaeontology 400

1 2 3 300 4 5 6 7 200 8 9 10 Number of CBCs 11 100 12 13 14 15 0 16 100 Albian Aptian Norian Danian Carnian Lutetian Chattian Toarcian Turonian Bajocian Aalenian Ypresian Rupelian Rhaetian Callovian Tithonian Langhian Zanclean Tortonian Oxfordian Bartonian Coniacian Thanetian Selandian Messinian Bathonian Santonian Berriasian Barremian Aquitanian Priabonian Hettangian Piacenzian Burdigalian Quaternary Sinemurian Hauterivian

17 Campanian Valanginian Serravallian Cenomanian Maastrichtian Kimmeridgian 18 Pliensbachian 0 10 20 30 40 50 60 70 80 90 19 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 20 21 22 80 23 24 25 26 27 60 28 29 30 31 32 40 33 34 35 36 37 Completeness (%) / Number of species 20 38 39 40 41 J6 J5 J4 J3 J2 J1

K8 K7 K6 K5 K4 K3 K2 K1 42

Tr5 Tr4 0 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 43 44 45 LT EJ MJ LJ EK LK Pa Eo Ol Mi P 46 Triassic Jurassic Cretaceous Palaeogene Neogene Q 47 48 0 10 20 30 40 50 60 70 80 90 49 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 50 51 52 53 54 55 56 57 Late Triass 58 59 60

Palaeontology Palaeontology Page 28 of 35 400

1 2 3 300 4 5 6 7 200 8 9

10 Number of CBCs 11 100 12 13 14 15 0 16 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 100Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 17 18 0

19 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 20 21 80 22 23 24 25 60 26 27 28 29 40 30 31 32 33 34 20 35 Completeness (%) / Number of species 36 37 38 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 39 40 41 LT EJ MJ LJ EK LK Pa Eo Ol Mi P 42 43 Triassic Jurassic Cretaceous Palaeogene Neogene Q 44 0 10 20 30 40 50 60 70 80 90 45 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Palaeontology Page 29 of 35 Palaeontology 100

1 2 3 4 5 80 6 7 8 9 10 11 60 12 13 14 15 16 17 40 18

19 Percentage completeness 20 21 22 23 20 24 25 26 27 28

04 08 100 80 60 40 20 0 0 29 30 31 Africa Asia Australasia Europe Madagascar N America S America 32 Africa Asia Australasia Madagascar S_America 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Palaeontology A PalaeontologyB Page 30 of 35

1 2 3 4 5 6 7 8 9 C D 10 11 12 13 14 15 16 17 18 19 20 21 E F 22 23 24 25 26 27 28 29 30 31 32 G H 33 34 35 36 37 38 39 40 41 42 43 44 I J 45 46 47 48 49 50 51 52 53 54 55 K L 56 57 58 59 60

Palaeontology Page 31 of 35150 Palaeontology 50 Africa 40 Asia 100 30 1 20 2 50

No. of CBCs 10 3 No. of CBCs 02 04 50 40 30 20 10 0 010150 100 50 0 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 4 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 0 10 20 30 40 50 60 70 80 90 0 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 10 20 30 40 50 60 70 80 90 5 15 100 110 120 15130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 6 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 10

10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 0 10 20 30 40 50 60 70 80 90 7 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 8 5 9 5 No. of species 10 No. of species 015 10 5 0 015 10 5 0 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 11 0 10 20 30 40 50 60 70 80 90 0 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 10 20 30 40 50 60 70 80 90 12 100 100 110 100120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 80

J6 J5 J4 J3 J2 J1 80 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 13 Ng3 Ng2 Ng1 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 14 60 60 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 15 40 40 16 20 20 Completeness (% ) J6 J5 J4 J3 J2 J1 04 08 100 80 60 40 20 0 J6 J5 J4 J3 J2 J1 04 08 100 80 60 40 20 0 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 17 Tr5 Tr4 Tr3 Pg1 Pg0 Pg1 Pg0 Completeness (%) 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 H Lr Lr Up Up Up Md Mio Plei Plio Eoc Oligo 18 Paleo

LT EJ MJ LJ EK LK Pa Eo Ol Mi LT EJ MJ LJ LK Pa Eo Ol Mi 0

EK 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 19 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 20 21 40 Australasia 60 Europe 22 30 23 40 20 24 25 10 20 No. of CBCs No. of CBCs 02 040 30 20 10 0 26 0 70 50 30 10 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 27 Ng3 Ng2 Ng1 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 10 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 28 15 8 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 29 0 0 10 20 30 40 50 60 70 80 90

10 20 30 40 50 60 70 80 1090 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 30 6 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 4 31 5 32 2 No. of species No. of species 10 8 6 4 2 0 33 0 15 10 5 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 34 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 100 35 80 80 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 36 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 37 60 100 60110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 38 40 40 39 20 20 Completeness (%) 04 08 100 80 60 40 20 0 04 08 100 80 60 40 20 0 Completeness (%) J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Pg1 Pg0 0 Pg1 Pg0 0 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 40 H H Lr Lr Lr Lr Up Up Up Up Up Up Md Md Mio Plei Plio Mio Eoc Plei Plio Eoc Oligo Oligo Paleo Paleo 0 0 10 20 30 40 50 60 70 80 90 LT EJ MJ LJ LK Pa Ol Mi 10 20 30 40 50 60 70 80 90 LT MJ Ol Mi Eo 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 41 EK 100 110 120 130 140 150 160 170 180 EJ190 200 210 LJ220 230 240 EK250 260 270 280 LK290 300 310 Pa Eo 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 42 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 43 250 100 44 200 North America 80 South America 45 46 150 60 47 100 40 No. of CBCs 48 50 No. of CBCs 20 010200 100 50 0 49 0 100 80 60 40 20 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 50 15 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 51 30 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 0 52 0 10 20 30 40 50 60 70 80 90 10 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 53 20 54 5 10 No. of species

55 No. of species 02 30 20 10 5 0 015 10 5 0 0 0 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 56 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 57 100 100 110 100120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 58 80 80 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 59 60 60 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 60 40 40 20 20 J6 J5 J4 J3 J2 J1 04 08 100 80 60 40 20 0 04 08 100 80 60 40 20 0 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 Completeness (%) K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg1 Pg0 Completeness (%) 0 Pg1 Pg0 0 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 Pg5 Pg4 Pg3 Pg2 Pg1 Pg0 Ng3 Ng2 Ng1 H H Lr Lr Lr Lr Up Up Up Up Up Up Md Md Mio Plei Plio Mio Eoc Plei Plio Eoc Oligo Oligo Paleo Paleo 0 LT EJ MJ LJ LK Pa Ol Mi 0 EK Eo 10 20 30 40 50 60 70 80 90 LT EJ MJ LJ EK LK Pa Eo Ol Mi 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 Palaeontology 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 Palaeontology Page 32 of 35 Number of CBCs 0 100 200 300 400 1 2 80–90N 3 4 5 70–80N 6 7 60–70N 8 9 50–60N 10 11 40–50N 12 13 30–40N 14 15 20–30N 16 17 10–20N 18 19 20 0–10N Completeness 21 22 0–10S No. of species

23 Palaeolatitude No. of CBCs 24 10–20S 25 26 20–30S 27 28 30–40S 29 30 31 40–50S 32 33 50–60S 34 35 60–70S 36 37 70–80S 38 39 80–90S 40 41 42 0 20 40 60 80 100 43 Completeness (%) / Number of species 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Palaeontology Page 3350–60 of 35 N 40–50 NPalaeontology 30–40 N 200 5 100 12 100 14 12 4 80 10 80 1150 10 8 2 3 60 60 8 3100 6 2 40 40 6 4 4 4 No. of species No. of species 5 50 No. of species 1 20 20 6 2 2 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3

7 Tr5 Tr4 Tr3 J6 J5 J4 J3 J2 J1 Pg1 Pg0 Pg1 Pg0 Completeness / No. of CBCs K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 0 0 0 Completeness / No. of CBCs 0 Pg1 Pg0 Completeness / No. of CBCs 0 0 8 LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi 0 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310

9 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 10 11 20–30 N 10–20 N 0–10 N 12100 100 100 12 13 10 80 80 15 80 14 10 6 1560 60 60 8 10 16 4 6 1740 40 40 18 5 4

2 No. of species

20 No. of species 20 20 19 No. of species 2 20 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Pg1 Pg0 Pg1 Pg0 Pg1 Pg0 Completeness / No. of CBCs 0 0 0 0 0 Completeness / No. of CBCs

Completeness / No. of CBCs 0 21 LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi 0 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90

22 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 23 24 0–10 S 10–20 S 20–30 S 25100 8 100 10 100 26 80 80 8 80 20 27 6 2860 60 6 60 15 29 4 3040 40 4 40 10 No. of species 31 2 No. of species 20 20 2 20 5 32 No. of species 33 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 Tr5 Tr4 Tr3 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Completeness / No. of CBCs Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Pg1 Pg0 Completeness / No. of CBCs Pg1 Pg0 Pg1 Pg0 0 0 0 0 0 Completeness / No. of CBCs 0 34 LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi 0 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 35 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 36 37 30–40 S 40–50 S 38100 7 100 10 39 6 80 80 40 5 8 Completeness 4160 60 4 6 42 No. of species 4340 3 40 4 No. of CBCs 44 2 No. of species

20 20 No. of species 45 1 2 46 J6 J5 J4 J3 J2 J1 J6 J5 J4 J3 J2 J1 K8 K7 K6 K5 K4 K3 K2 K1 K8 K7 K6 K5 K4 K3 K2 K1 Tr5 Tr4 Tr3 Tr5 Tr4 Tr3 Completeness / No. of CBCs Pg1 Pg0 Pg1 Pg0 0 Completeness / No. of CBCs 0 0 0 47 LT LJEJ EK LK P Eo O Mi LT LJEJ EK LK P Eo O Mi 0 0 10 20 30 40 50 60 70 80 90 48 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 49 50 51 52 53 54 55 56 57 58 59 60

Palaeontology A B 12.5 Palaeontology Page 34 of 35 1.0

1 2 10.0 3 4 0.5 5 7.5 6

7 10 8 5.0 0.0 9 Log TL 10 11 12 2.5 Untransformed TL (m) TL Untransformed −0.5 13 14 15 0.0 16 17 0 25 50 75 100 0 25 50 75 100 18 Completeness (%) Completeness (%) 19 C20 D 21 0.3 22 0.6 23 24 25 26 0.0 27 0.5 28 29 30 31 −0.3 32 33 0.4 34 35 −0.6 36 37 38 0.3 39 40 0 25 50 75 100 0 25 50 75 100 41 Completeness (%) Completeness (%) 42 43 E44 F 45 1.1 46 47 0.75 48 49 50 1.0 51 52 0.70 53 54 55 0.9 56 57 0.65 58 59 60 0.8 0.60 Palaeontology 0 25 50 75 100 0 25 50 75 100 Completeness (%) Completeness (%) 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Page 35of

Percentage completeness 100 80 40

20 60

0

Non-Metasuchia

Notosuchia Neosuchia

Palaeontology Palaeontology

Goniopholidae

Crocodylia

Alligatoroidea

Crocodylidae Tomistominae