Matskási István (szerk.): A Magyar Természettudományi Múzeum évkönyve 87. (Budapest 1995)

Farkas, B.: Fossil trionychid turtle types in Hungarian collections - a preliminary review (Reptilia, Testudines)

Trionyx pseudovindobonensis SZÁLAI, 1934 (Fig. 2) The holotype (HGS 3145) consists of a left femur fragment, instead of that of the right humerus, as stated by both SZÁLAI (1934) and MLYNARSKI (1966). A pencilled inscription on the type label indicates that this error has already been discovered by a previous worker. Nevertheless, no correction has ever been published. As OGUSHI (1911) rightly pointed out, the trionychid femur "ähnelt im großen und ganzen sowohl bezüglich Größe als Gestalt dem Humerus ganz erheblich, so daß man bei flüchtiger Betrachtung (...) die beiden kaum voneinander zu unterscheiden vermag". In aquatic turtles, the femur is subequal in length to the humerus or slightly longer. Nevertheless, it is immediately recognizable by its more elongated head (caput femoris), and, more importantly, by the different degree of development of the tuberositas magnus (trochanter major) versus the lateral process (p. radialis) of the humerus, and the tuberositas internus (tr. minor) versus the medial process (p. ulnaris) of the humerus, i.e. the direction of the longitudinal axis of the caput versus that of the femorotibialis and the triceps brachii, respectively. The medial (ulnar) process of the humerus is much more pronounced than the lateral (radial) process, whereas the tuberositas magnus (trochanter major) is distinctly larger than the t. internus (tr. minor) and extends proximally to or beyond the end of the femoral head in good swimmers such as trionychids (OGUSHI 1911, WALKER 1973). Based on comparisons of femur length and disk length in Recent soft-shelled turtles, the specimen in question must have belonged to a trionychid of at least 40 cm bony disk length. The type material of T. pseudovindobonensis originates from the Miocene as well (Upper Pannonian), from Rákos, "Komitat Pest, Südungarn", according to MLYNARSKI (1966). The town of Rákos itself is located in southern Hungary (Makó municipality, Csongrád County), but SZÁLAI (op. cit.) clearly stated that the specimen was found in Pest County, and actually referred to Rákoshegy in Budapest. Interestingly, MLYNARSKI (1966) plotted this locality almost correctly on his map, but on the Buda side of the city! The main reason for distingvishing this taxon from T. vindobonensis PETERS has been the allegedly different angle between the longitudinal axes of the corpus (femorotibialis) and the caput femoris ("c. humeri" in the original description). Since this angle is highly variable even within individuals (Fig. 3) and the diagnosis of the species is based entirely on the non-existing humerus, it appears unjustified to accept T. pseudovindobonensis as a valid taxon. On the other hand, it is not unlikely that if this angle had been correctly measured in T. vindobonensis (on the femur instead of on the humerus), no significant differences would have been found between them. Unfortunately, I have not been able to examine any material of that species. Thus at present, it seems most acceptable to declare HGS 3145 a "Trionychidae indet.". Trionyx harmati SZÁLAI, 1934 This species is based on a "left humerus" with lateral (radial) and medial (ulnar) processes broken off, of Middle Oligocène (Ruppelian) age found in the brick-laying works of the company Nagybátonyi Újlaki Iparművek, Budapest. The holotype, as far as it can be judged from the single, rather poor photograph existing, is a composite of two individual elements glued together. It has never been deposited in a public collection, and is assumed lost with the death of TIBOR SZÁLAI. As I already pointed out, it is a highly questionable procedure to base a taxon solely on subtle and subjective differences in the shape of the caput humerus, and again on the practically immeasurable but individually variable angle between the length axes of the head and the corpus. I am not convinced that this is a single specimen or it is a humerus either, because there is seemingly no trace of a canalis ectepicondyloideus, so I hesitantly