Amphicyonidae TROUESSART , 1885

Morales, Jorge, Fejfar, Oldřich, Heizmann, Elmar, Wagner, Jan, Valenciano, Alberto & Abella, Juan, 2019, A New Thaumastocyoninae (Amphicyonidae, Carnivora) From The Early Miocene Of Tuchořice, The Czech Republic, Fossil Imprint 75 (3 - 4), pp. 397-411 : 399-400

Family Amphicyonidae TROUESSART, 1885

Subfamily Thaumasthocyoninae HÜRZELER, 1940

D i a g n o s i s. Amphicyonidae presenting hypercarnivorous dentition, with reduced M2/m2–m3 and highly developed and sectorial carnassial teeth. Specialized genera with a reduction of the premolars and a large mandibular diastema, well-developed P4 parastyle and reduced M1 lingual area.

I n c l u d e d g e n e r a. Thaumastocyon STEHLIN et HELBING, 1925 , Crassidia HEIZMANN et KORDIKOVA, 2000 , Ysengrinia GINSBURG, 1966 , Tomocyon VIRET, 1929b and Peignecyon n. gen.

R e m a r k s. The taxonomy and systematics of the Thaumastocyoninae amphicyonid have been considered to be quite complex since Hürzeler (1940), who described them as aberrant canids with feline tendencies. The latter author correctly characterized the subfamily based on the complete suppression of m1 metaconid, reduction of the premolars, except the p4, which is reinforced, and the oblique abrasion of the teeth. He included in the group Thaumastocyon bourgeoisi , Agnotherium antiquum KAUP, 1833 , Tomocyon grivensis VIRET, 1929b together with the amphicyonid remains from Frohnstteten, figured by Jäger (1835), unpublished materials housed in the Basel Museum ( Switzerland) from Baigneaux-en-Beauce ( France) and Chairmoille (Switzeland), and eventually, a mandible fragment from Steinheim ( Germany), which was determined by Helbing (1929) as Pseudocyon sansaniensis LARTET, 1851 , and subsequently chosen as the holotype of a new subspecies, Pseudocyon sansaniensis serus KUSS, 1965 ( Kuss 1965: 133) .

Kuss (1962) performed the first in-depth review of the group, in which he proposed the inclusion of Tomocyon grivensis in Agnotherium . He considers that A. antiquum likely had upper molars with a non-reduced lingual area, an argument which favoured the determination of Frohnstetten fossils as A. antiquus , reconstructing the molar morphology with a strong development of the lingual area, using as a model the molars of T. grivensis ( Kuss 1962: fig. 6). For some fossils from Baigneaux-en-Beauce, Kuss (1962: fig. 7) proposes their classification as Agnotherium grivensis (right m1; NMB S.O. 4459), aff. Agnotherium grivensis (for three P4; NMB S.O. 1274, 874 and 4460) and?aff. Agnotherium grivensis (left M2; NMB S.O. 1559).

Ginsburg (1999) seconds this proposal, pointing out the presence of Agnotherium grivensis in Simorre ( France) and Agnotherium antiquum in Pedregueras ( Spain), and the presence of a second Thaumastocyon species , Th. dirus from Los Valles de Fuentidueña ( Spain) ( Ginsburg et al. 1981).

Heizmann and Kordikova (2000) erected the tribe Ysengrini to contain Crassidia , a new genus described in the same research, together with Ysengrinia and Amphicyonopsis VIRET, 1929b . For these authors, the Ysengrini occupies an intermediate position between Amphicyoninae and Thaumastocyoninae , but since the defining features of Thaumastocyoninae , loss of the m1 metaconid and shortening of the entire facial region, are not present in the Ysengrini , they consider them to constitute a specialized branch of the Amphicyoninae . But they did not take into account that both groups share other derived characters, which indicate a close relationship between them.

Amphicyonopsis was described by Viret (1929b) on the basis of m1 (MHNL LGR 1136) from La Grive ( France), but no species was included in the new genus. Kuss (1965) included it, together with other teeth from La Grive as Pseudocyon sansaniensis aff. serus n. subsp. Heizmann (1973) considers that the holotype of this subspecies, the mandible fragment from Steinheim included by Hürzeler (1940) within the Thaumastocyoninae , should be distinguished from Pseudocyon and attributed to Amphicyonopsis ? serus ( KUSS, 1965) .

Mein and Ginsburg (2002), partially in accordance with Kuss (1965), transferred the fossils of La Grive to Pseudocyon sansaniensis and therefore excluded them from the Thaumastocyoninae .

Identification of Agnotherium and Ysengrinia in Africa is problematic, as has recently been discussed by Morales et al. (2016).For these authors, theAfrican amphicyonids previously identified in these genera acquired their hypercarnivorous adaptation independent from the European taxa. This could be the case of Ysengrinia americana , as will subsequently be discussed. The Tuchořice Thaumastocyoninae are important, since they can help to identify the beginning of the more specialized Thaumastocyoninae , and because they are a significant link between these genera and the more primitive species included in the Ysengrini tribe by Heizmann and Kordikova (2000).

Thus, our paper accepts the presence of two tribes of Thaumastocyoninae in the late Oligocene/Miocene of Europe – Thaumastocyonini HÜRZELER, 1940 which included the genera Thaumastocyon STEHLIN et HELBING, 1925 , Tomocyon VIRET, 1929b and Peignecyon n. gen., and Ysengrini HEIZMANNN et KORDIKOVA, 2000 which included the genera Ysengrinia GINSBURG, 1966 and Crassidia HEIZMANNN et KORDIKOVA, 2000 .

We did not consider some taxa, such as Amphicyonopsis serus ( KUSS, 1965) , which we believe to be a species of dubious validity, or Agnotherium antiquum KAUP, 1833 , defined in Eppelsheim (MN 9), Germany, and only represented by one sectorial m1. This tooth displays an advanced degree of dental wear, which has partially erased its morphology. The paraconid mesial cristid is displaced lingually and is well marked, connecting to the basal lingual cingulum; the talonid is short and narrow with a high hypoconid and a reduced entoconid. These characters are associated with the hypercarnivorous tendency, and might likely have been acquired convergently in different Thaumastocyoninae taxa. The genera Thaumastocyon and Tomocyon present an m1 close to A. antiquus , but differ in their associated upper and lower teeth. We do not avail of any criteria to relate the Eppelsheim genus more closely to either one of these genera. Given our current knowledge, this m1 is insufficient with regard to defining a genus, or even a species. Most of the Agnotherium antiquus published determinations have been indirect, based upon tooth size and sectorial morphology of specimens. This has given rise to multiple errors, as Kurtén (1976) initially pointed out. The determination of Agnotherium antiquus in Pedregueras 2A and Charmoille exemplify this circumstance; in both localities the material comprises a lower canine, p4 and m1; the m1s are morphologically close to the tooth of Eppelsheim, although the mesial paraconid-cristid is less marked. The canine and p4 present a similar morphology and robustness to those of La Grive homologous teeth ( Viret 1929a: pl. 19, figs 3, 7). The conclusion is evident; there is a suite of features such as the advanced degree of m1 sectorialitation (loss of metaconid, talonid reduction, sharp cristid, etc.), the robustness of the lower canine and the p4, and others, which enable the Thaumastocyonini clade to be defined, but which by themselves do not permit an unequivocal determination within this taxon.