Usefulness of the morphological characters used for the generic taxonomy of “ictitheres” [11] was confirmed by examination of new finds and formerly inaccessible material. The latter includes the type specimens from China in the Lagrelius Collection, described by Zdansky [16], and the numerous cranial remains in the Frick Collection. In general, fossil and recent cranial material of hyaenids, viverrids and herpestids in the following institutions were examined: •

American Museum of Natural History, New York (AMNH);

In this contribution, the “ictitheres” are considered as belonging to two different groups. The first one is the Ictitheriinae sensu stricto or true ictitheres [11]. It includes the following genera: •

the nominative genus Ictitherium Roth and Wagner, 1854 – including I. viverrinum (= I. ibericum Meladze, 1967; I. gaudryi Zdansky, 1924; I. sinense Zdansky, 1924; I. intuberculatum), I. pannonicum Kretzoi, 1952 and I. spelaeum Semenov, 1988;

The second group is the tribe Hyaenotheriini, namely: •

the nominative genus Hyaenotherium Semenov, 1989 – including H. magnum Semenov, 1989, H. wongi (Zdansky, 1924) and Hyaenictitherium hyaenoides orlovi Semenov, 1989;

In the present article, the focus is placed on the most common large species that are comparable in size to the Lothagam hyaenids. Smaller members of both groups (such as the genera Plioviverrops Kretzoi, 1938, Protictitherium Kretzoi, 1938 and Tungurictis Colbert, 1939, as well as Hyaenictitherium minimum de Bonis et al., 2005) are almost excluded from the consideration.

The main characters that differentiate the tribes Hyaenotheriini and subfamily Ictitheriinae are presented in Table 1 (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6 and Fig. 7). It should be noted that body size is almost of no value for taxonomical identification above the species level. For example, I. pannonicum, being closely related to I. viverrinum and other relatively small species, is much larger than typical “hyaena-like” Hyaenictitherium minimum (the m1 length is 19.3–21.0 mm in the former [11] and 15.0–16.5 mm in the latter [1]). In addition, the relative sizes of the second molars and the length of the m1 talonid vary greatly both within the genus and even species in ictitheriines.

Judging from its description, measurements and drawings [14], “Ictitherium” ebu is clearly a member of the tribe Hyaenotheriini and hence cannot belong to Ictitherium. The specimen has evenly inflated auditory bullae; lacks an alisphenoid canal; as evidenced by the short and backwardly extended sagittal crest, the temporal lines are presumably long and converge behind the level of the glenoid fossa; occipital crest is relatively small (Fig. 8). Mandible has a single mental foramen, and the lower contour of the mandibular body is nearly straight. The P4 protocone does not protrude anteriorly to the parastyle; the m1 talonid is narrow (35.8% of the crown length) and moderately short (25.1% of the crown length); M2 is not “somewhat reduced” [14] because its length averages only 17.8% of the P4 length. Finally, the limbs of “Ictitherium” ebu are essentially longer and much more slender than even those of H. wongi (Fig. 7). These features reliably disprove the ictitheriine affinity of this hyaenid. The unusual length of its limbs would, on its own, be sufficient to regard this carnivore as a member of the tribe Hyaenotheriini. Considering the slightly inflated auditory bullae and short postorbital and jugular processes of the species, it belongs most probably to the genus Hyaenotherium. Such a conclusion can explain why, in its main features the reconstruction of the Kenyan “Ictitherium” ebu is so similar to that of Ukrainian Hyaenotherium magnum (Fig. 9).

A mandibular fragment of another hyaenid from Lothagam was originally referred [14] to Hyaenictitherium cf. H. parvum (Fig. 10A). Although this specimen is incomplete, its degree of preservation allows reliable generic identification. First of all, it seems questionable that a hyaenid with very narrow and extremely high-crowned premolars (especially p3) could belong to the genus Hyaenictitherium. Dentally, the specimen more closely resembles Hyaenictis. This is also suggested by a single mental foramen and rounded mental part of the mandible in lateral view. Moreover, another fragment of the lower jaw was found in Lothagam, that of a senile individual. It was originally referred [14] to Hyaenictis sp. (Fig. 10B). This specimen differs from the “parvum” specimen in the age-related characters only, namely in the deepening of the mandibular body, particularly in its symphyseal region. This suggests a conspecific status of the two hyaenids, rather than their belonging to the different genera. The original assignment of the specimen to Lycyaena parva Khomenko, 1929 (described from Taraklia, Moldova [7]) seems unjustified. The only feature shared by these two hyaenines is their small size (Fig. 10).

Importantly, the application of the generic name Palhyaena to any “ictitheres” and the name Thalassictis to any hyaenotheriins is a mistake [11]. The misuse of Palhyaena stems from the publications by Gaudry [2] and [3] who erroneously used this name for two non-congeneric specimens. Fig. 11 shows that Palhyaena hipparionum sensu Gervais from Mont Léberon [4] lacks M2 and its postcanine teeth differ appreciably from those of I. hipparionum sensu Gaudry from Pikermi [3]. This indicates that the former specimen having such a powerful P3, belongs neither to Ictitheriinae nor to Hyaenotheriini.

The generic name Thalassictis is often applied to the members of Hyaenotheriini, mainly with additions such as “Gervais, ex Nordmann”, “Gervais, non Nordmann” or “Nordmann, in Gervais”. Passing over the history and details of this evident nomenclatural problem, it could be said that in such a case this name can be applied only to Mustela incerta Gervais ex Lartet or become a synonym of Ictitherium. Originally the name Thalassictis was proposed by Nordmann [8] for the material from the Sarmatian deposits of Kishinev (Moldova). But this name is also quite suitable to the nearly complete skull from the same locality, later described by Pavlow [10] as I. sarmaticum, since many craniodental traits suggest the conspecific status of these finds [11]. Judging by the morphology of the auditory region, cranial vault and occipital bone and in spite of considerable reduction of the second molars, P4 protocone and m1 talonid (Fig. 12), the species from Kishinev is undoubtedly related to the true Ictitheriinae, but not to any Hyaeninae. At the same time, the unique shape of its auditory bulla, very short postorbital process, slight reduction of the fourth root of the P4 (Fig. 2 and Fig. 13), protocone of the P4 not protruding (Fig. 6 and Fig. 12) and very tall m1 protoconid (Fig. 12D) make T. robusta easily distinguishable from all species of the genus Ictitherium and other ictitheriines.

The genera Hyaenotherium, Miohyaenotherium and Thalassictis are valid and morphologically well defined taxa, of which the former two belong to the tribe Hyaenotheriini and the latter one belongs to the subfamily Ictitheriinae.

The genus Hyaenotherium, represented by H. ebu Werdelin, 2005, from Lothagam, can be reliably added to the list of fossil African Carnivora. More cautiously, this species may belong to a new genus of Hyaeninae, but in any case, it cannot be a member of the Ictitheriinae and, even less so, of the genus Ictitherium.

The other small hyaenid remains from Lothagam, originally referred to Hyaenictitherium cf. H. parvum and Hyaenictis sp. [14], most probably belong to the same unidentified species of the genus Hyaenictis.

Palhyaena hipparionum from Mont Léberon belongs to the subfamily Hyaeninae, though not to the tribe Hyaenotheriini. Its generic name cannot be used for any “ictithere”.

The subfamily Ictitheriinae in its common usage is a collective group which needs a comprehensive revision.