Batropetes Carroll & Gaskill 1971

Description

BONE MICROANATOMY, PROPORTIONS AND LIFESTYLE OF BATROPETES

Micro-CT data from MB.Am.1232, a postcranial skeleton of an adult Batropetes palatinus, reveal a thin, solid cortex throughout the proximal and distal limb bones, the girdles and the vertebrae (Fig. 3). In the humerus, the cortex makes up less than half of the diameter at mid-diaphysis; elsewhere in the humerus, and everywhere in the femur, it is much less. All ribs are split throughout their length, which is visible both on the outside (Fig. 2) and in the scan images (Fig. 3); this indicates collapse of an extensive marrow cavity under diagenetic pressure. These observations confirm (e.g. Buffrénil & Rage 1993; Laurin et al. 2004, 2011; Germain & Laurin 2005; Kriloff et al. 2008; Canoville & Laurin 2009, 2010; Buffrénil et al. 2010; Cooper et al. 2011; Dumont et al. 2013; Quémeneur et al. 2013) previous interpretations of Batropetes as terrestrial (Glienke 2013, 2015; contra Carroll 1991; Mann & Maddin 2019), even though the resolution of the scan does not permit us to distinguish spongiosa from the infill of the marrow cavity.

The µCT data allow us to reconstruct the humerus of MB.Am. 1232 in three dimensions (Fig. 3 C-E). We find a dorsal process (accentuated by breakage) as reported in various lissamphibians, “microsaurs” and amphibamiforms, and a triangular deltopectoral crest that is not deflected medially as it is in salamanders (e.g. Ambystoma Tschudi, 1838: Sigurdsen et al. 2012: fig. 3A) or to a lesser degree in Eocaecilia Jenkins & Walsh, 1993 (Jenkins et al. 2007: fig. 42; Sigurdsen et al. 2012), but slightly laterally, producing a shallow concavity lateral of it (Fig. 3D), similar to the less extreme cases among salientians (Sigurdsen et al. 2012).

COMPARATIVE LIMB PROPORTIONS AND LIFESTYLES

The morphometric variability of the limbs of the sampled taxa, both extant and extinct, reflects different locomotor functions, which we categorize for the extant species following Lires et al. (2016), Oliveira et al. (2017a, b), and references therein. In our LDA (Figs 4-6; Table 3; Appendices 2-4), the fossil individuals mostly plot with caudates and squamates (which retain much of the ancestral tetrapod body shape) in a wider cluster including the LU cluster of extant species and the separately categorized cluster of extant individuals known to routinely engage in digging behavior (LUD).

In the LDA, the LU and LUD clusters do not separate well in most comparisons (Figs 4-6; Appendices 2-4). Indeed, the right side of MB.Am.1232 is classified as LU, the left side as LUD (Table 3). Only the comparison of canonical variant 1 to canonical variant 4 (Fig. 5; Appendix 2) shifts the digging individuals further away from all other locomotor categories, but they still retain a large overlap. This is in part due to the wide definition of “digging” in the analysis, and in part to the facts that LU is the plesiomorphic state and that LUD is directly derived from it (while e.g. Sw is evolutionarily derived from J, not directly from LU).Nonetheless, MANOVA finds all five locomotor categories to be clearly distinct (F = 50.037, df = 16 and p -value = 9.28 × 10–109, well below the detection threshold of 2.2 × 10 –16).

The LDA prediction of the added extinct taxa using Bayesian posterior probability (Table 3) recovers most of them as digging and plots them outside the overlap area of LU and LUD (Fig. 5; compare Fig. 4), but classifies one of the Batropetes specimens (the only one included of B. fritschi (Geinitz & Deichmüller, 1882)) as a toad-like HW. The other Batropetes specimens are classified as LUD, except for the right side of MB.Am.1232 as mentioned.

A direct comparison of the ranges of the four used limb measurements reveals that Batropetes generally falls within the range recovered as LU/LUD. The relative lengths of radius and ulna, however, also overlap with the HW category (Fig. 6), revealing a more elongate distal forelimb.

Triadobatrachus also still falls within the LU/LUD cluster, as it did in Lires et al. (2016). Specifically, Triadobatrachus is classified as LU (Table 3), agreeing with the idea that limb morphology is generally plesiomorphic for most taxa falling within LU and LUD.

Doleserpeton Bolt, 1969 is the only taxon that does not cluster with any of the defined groups representing locomotor categories in Figures 4 and 5. It plots as a distant outlier in the LDA (Figs 4-6), because once the measurements are divided by the geometric mean, the femur length appears to be smaller than in all other specimens used in this analysis, while the radius-ulna length appears to be greater. Because sufficiently articulated or associated skeletons are not known (Bolt 1969; Sigurdsen & Bolt 2010; Gee et al. 2020), the measurements were taken from different specimens, corrected for size, as well as from the skeletal reconstruction by Sigurdsen & Bolt (2010), and both linear measurements (from the figured bones as well as from the reconstruction) show the same relation once they are divided by the geometric mean. However, we cannot exclude a measurement error in the literature at this point. Nor can we exclude the possibility that some of the measured material comes from other amphibamiform taxa, of which two are known from skulls found at the same site (Fröbisch & Reisz 2008; Anderson & Bolt 2013; Atkins et al. 2020), as discussed in detail by Gee et al. (2020).

Of the other two amphibamiform temnospondyls that we were able to sample, Platyrhinops Steen, 1931 is classified as a lateral undulator as expected, with absence of digging behavior (LU) weakly favored (BPP = 59 %) over its presence (LUD; BPP= 41%), while Micropholis Huxley, 1859, with its particularly short trunk and long limbs (Schoch & Rubidge 2005), emerges unambiguously as a hopper/walker (90% and 95% for the two specimens) – more froglike in this respect than Triadobatrachus (BPP = 71% for LU, <0.1% for HW). The LDA reveals that Micropholis is particularly close to Bufo bufo (Linnaeus, 1758) in linear discriminants 1 and 2, though widely separated by linear discriminant 4 (Figs 4; 5; Appendix 2).

The three “microsaurs” other than Batropetes are classified as lateral undulators, in agreement with their interpretations as terrestrial in the literature. For Tuditanus Cope, 1871, with its particularly lizardlike proportions (very similar to those of contemporary early amniotes of the same size), LU is favored (64 %) over LUD (36 %), while the opposite is the case for the early brachystelechid Diabloroter Mann & Maddin, 2019 (34 % vs 66%), and for the particularly stocky Pantylus Cope, 1881 (20 % vs 80%).

The two specimens of the albanerpetid near-lissamphibian Celtedens ibericus are classified as LU (78% and 81 % respectively) over LUD (22 % and 19 %). While this is evidence against limb-based digging (see alsoDaza et al. 2020), it may not contradict head-based digging in leaf litter (Wiechmann 2000; Gardner 2001; and references therein).

It is noteworthy that Triadobatrachus, which has a considerably longer tarsus than all non-salientians in our sample, remains in LU even though we ignore its tarsus, and does not join HW. As in Lires et al. (2016), no other salientian is found in LU or LUD.

Authors

• Jansen, Maren
• Marjanović, David

DOI: 10.5281/zenodo.20762805

Publication Date: 2022-07-06

Back to publications list

About