Mesosaurus, Gervais, 1865, Gervais, 1865

COULD THE REVERSED EMBOLOMEROUS CONDITION HAVE BEEN PRESENT (ALTHOUGH MASKED) IN OTHER EARLY AMNIOTES THAN MESOSAURUS ?

Despite the reverse embolomerous condition described herein for the mesosaur caudal vertebrae, the architectural con - struction of the tail does not differ substantially from that observed in most groups of early amniote taxa. For instance, the caudal region, when preserved, is composed of vertebrae possessing an intercentrum positioned anterior to the pleurocentrum, as also occurs in the dorsal vertebral segment. Moreover, the intercentrum of vertebrae positioned posterior to the proximal string (pygal) with a variable number of elements, carries the haemal arches even near the tip of the tail ( Heaton & Reisz 1980). However, a rapid revision of the available previous literature (old and more recent) shows that the mesosaur configuration could have been present in the caudal vertebrae of other basal tetrapod taxa. Starting with the study of Everett Olson (1936) on the axial musculature in early tetrapods, we can see some clues: caudal vertebrae assigned to Diadectes Cope, 1878 and Dimetrodon Cope, 1878 are drawn in figure 8 of Olson (1936: 280) and their haemal arches are fused to the posterior end of the centrum. The same condition is apparently present in the synapsid Varanosaurus acutirostris Broili, 1904 , according to Sumida (1989: 155, fig. 4) and could have been developed in earlier taxa such as the seymouriamorphs, as documented by Laurin (1996: 658, fig. 5) for Ariekanerpeton sigalovi ( Tatarinov, 1968) , where four haemal arches are positioned (although not articulated) close to the posterior end of the respective centra. The fact that all four arches display the exactly same position with respect to the centra, makes it improbable that it results from a taphonomic process.

In basal diapsids, there seems to be an anterior “pygal segment”, although it is shorter than that observed in Mesosaurus . In Petrolacosaurus Lane, 1945 , however, intercentra are described in this anteriormost caudal segment, as well as the corresponding haemal arches associated to the third or fourth vertebrae ( Peabody 1952; Reisz 1981). Furthermore, the possibility of development of autotomy was suggested for Araeoscelis Williston, 1910 , although it was only based on the presence of an isolated ossified, roughly conical structure of uncertain identity, bearing strongly marked ribs as a kind of superficial ornamentation, a morphology very similar to what is observed in regenerated tails of some squamates (see Vaughn 1955: pl. 2).

Another interesting case is that of Dolabrosaurus aquatilis Berman & Reisz, 1992 , a drepanosaurid diapsid from the Upper Triassic of North America, where the haemal arches can be fused to both the posterior and the anterior end of the caudal vertebrae ( Berman & Reisz 1992), This suggests a reverse embolomerous condition in the more anterior vertebrae, but a return to the normal configuration at the posterior region. More recently, Hypuronector limnaios Colbert & Olsen, 2001 , another related drepanosaurid, was shown to display the reverse embolomerous condition. In this last taxon, the fusion of the haemal arches to a trapezoid-like intercentrum is more evident, and a putative fracture plane can be also observed at the middle of the centrum in some of the preserved caudal vertebrae (see Colbert & Olsen 2001: fig. 9C). The hypothesis of autotomy suggested for captorhinids (i.e., LeBlanc et al. 2018) is difficult to ascertain because of the fragmentary nature of the analyzed specimens, as explained above. But taking into account the obvious presence of fracture-suture lines in most of the caudal vertebrae shown by LeBlanc et al. 2018) and the articulation of the haemal arches to the posterior end of the centrum observed in the string of caudal vertebrae belonging to a juvenile individual shown ( LeBlanc et al. 2018: fig. 4b), it may be not surprising that capthorinids, diadectids, seymouriamorphs and some synapsids may have had the reverse embolomerous condition, but it was not previously detected.

THE INTENSE CONTROVERSY ABOUT MESOSAUR RELATIONSHIPS IS NOT OVER

Although mesosaurs are most frequently considered to be specialized aquatic amniotes ( Araújo 1977; Oelofsen 1981; Modesto 1996, 1999; among others), they display a combination of primitive and derived characters, perhaps associated to a progressive although incomplete adaptation of an aquatic lifestyle ( Núñez Demarco et al. 2018). However, the extent of such an adaptation in mesosaurs is not completely clear. It is true that some of the mesosaur features reveal limited capability for terrestrial locomotion, a trait also observed in numerous Carboniferous stegocephalians, such as Gephyrostegus Jaekel, 1903 and Westlothiana Smithson & Rolfe, 1990 ( Smithson 1989; Piñeiro et al. 2016; Herbst & Hutchinson 2018). The last two taxa, despite being outside Amniota, exhibit terrestrial specializations like the development of an amniotic tarsus with precursor bones arranged in the normal position of the astragalus and calcaneum but showing different stages of fusion with the intermedium and the fibulare (see figure 10 in Piñeiro et al. 2016). But mesosaurs may not have developed capabilities to an exclusively fully aquatic lifestyle, and they presumably inhabited shallow, plausibly hypersaline water. This is suggested by the the presence of pachyosteosteosclerosis in this taxon, a condition which was suggested to be developed in animals adapted to shallow aquatic environments ( Houssaye 2009; Canoville & Laurin 2010). This hypothesis could be linked to the progressive draught of the Mangrullo and Irati seas, as was suggested in previous papers based on solid evidence suggesting an increasing of the water salinity and the deposition of evaporitic gypsum crystals ( Piñeiro et al. 2012b, 2025; Petri et al. 2022).

Transitional features in mesosaurs have been interpreted as suggestive of possible affinities with other aquatic to semiaquatic taxa, such as the recumbitrostran microsaur lepospondyls ( Piñeiro et al. 2016; Núñez Demarco et al. 2022), which were recently considered by some authors as being part of the stem of Amniota ( Pardo et al. 2017). However, affinities between mesosaurs and recumbirostran microsaurs may be unlikely given that extensive phylogenetic analyses recovered the former as amniotes, close to the base of Sauropsida or at the base of Parareptilia. Furthermore, the position of recumbirostran microsaurs has been far more controversial; they are stem-amphibians according toMarjanović & Laurin (2019), but crown-amniotes according to Pardo et al. (2017) and Mann et al. (2023), among others.

However, mesosaurs have retained many ancestral characters, including: 1) the presence of five phalanges in the fifth pedal toe, whereas just four or three are observed in most tetrapods and basal amniotes ( Hylonomus Dawson, 1860 pes, for instance, is reconstructed with three phalanges at the V toe; Clack et al. 2022); 2) a longer metatarsal V. Among basal stegocephalians, a long toe V, which is the longest of the pedal series, is a condition only present in the anthracosaur Silvanerpeton miripedes Clack, 1994 ( Clack 1994; Ruta & Clack 2006) and the embolomerous Archeria Case 1915 (see Clack et al. 2022). While this character can be autapomorphic for mesosaurs or it can be an adaptation to an aquatic (or semi-aquatic) lifestyle, it is only shared with basal taxa; and 3) mesosaurs displayed an isometric growth pattern, a condition that could be considered of uncertain polarity, given that it is documented in few taxa, but that currently is only observed in basal or stem amniote taxa. The isometry is particularly marked at the level of limbs, as observed in microsaurs and other lepospondyls, sharply contrasting with the allometric pattern characterizing other early amniotes ( Núñez Demarco et al. 2022).

According to a recent taxonomic review of Mesosauridae , Mesosaurus tenuidens is the only valid species ( Piñeiro et al. 2021; but see also Verrière & Fröbisch 2022), reducing even more the already low diversity in early tetrapods observed during the lifespan of mesosaurs in Southern Gondwana. Whereas mesosaur phylogenetic affinities remain controversial, a putative relationship between mesosaurs and the basalmost amniote groups seems possible, even at the level of the amniote stem group ( Piñeiro et al. 2016; Núñez Demarco et al. 2022, but see also Pardo et al. 2017).

Other hypotheses have suggested a position on the reptilian stem ( Laurin & Reisz 1995; Laurin & Piñeiro 2017) or at the base of Parareptilia ( Modesto 2006), but they would have to be revised by including recently published new data about mesosaur anatomy, ontogeny, taxonomy and physiology.

Indeed, mesosaurs have always appeared as a very basal group in the main known phylogenetic trees on amniote relationships, either as basal sauropsids or as basal parareptiles (regarding that the later are also basal sauropsids), and that signal should be taken into account in future studies.