Morphological and biometrical description of skulls, teeth, and limb bones of extant and fossil Old World herniones (including E. hydruntinus) and of New World 'stilt-Iegged' and other slender species from Blancan to Holocene. An Appendix presents ways in which the approximate size of some missing bones or dimensions may be deduced from available ones.

The discussed and/or illustrated fossils were found in Bolivia (Tarija), Canada (Yukon), China (Choukoutien, Gulongshan, Jiling, Loufangzi), Ecuador (Oil Fields), Ethiopia (Melka Kunturé), France (Lunel-Viel), Germany (Süssenborn), Greece (Agios Georgios, Petralona), Hungary (Dorog), Italy (Romanelli), Mexico (Cedazo, San Josecito), Mongolia (Sjara-osso-gol), Spain (Venta Micena), ex-Soviet Union (Akhalkalaki, Binagady, Chokurcha, Chukochya, Kabazi, Kolyma, Krestovka, Kurtak, Staroselie, Tologoj), USA (Alaska, Arkalon, Cedar Meadow, Channing, Conkling, Dry Mountains, Hay Springs, Leisey Shell Pit A, Lissie Formation, Natural Trap, Pool Branch, Powers Ranch, Rock Creek, San Diego, Santo Domingo, Seymour Formation, Shelter, Slaton, Trinity River). Numerous raw or statistically elaborated data are given in Tables.

There is no evidence for the existence of Old World hemiones in the New World nor of 'stilt-Iegged' equids in the Old World. The first 'stilt-Iegged' equid was found at Santo Domingo, New Mexico, and is believed to be Late Blancan. It was probably at the origin of E. calobatus (Arkalon, Rock Creek) and of the smaller E. semiplicatus (Channing, Rock Creek). Slender, but not 'stilt-Iegged', equids found at Natural Trap, Wyoming, ca. 12 ky ago belong to Amerhippus. AlI these species share with Oid World Sussemiones (and some hemiones) peculiar patterns on the lower cheek teeth.

The slender Equus sp. B of Leisey Pit A, Florida, ca. 1.2 Ma, as weIl as Amerhippus francisci and E. tau (probably a senior synonym of E. quinni) share conventional lower cheek teeth patterns. The skulls of A. francisci and E. tau, however, are quite different.

Paleontological data suggest a common origin of Amerhippus, Sussemiones, and 'stilt-Iegged' equids during the late Blancan. Old World hemiones seem to have differentiated later. 

The Zanclian marls from Papiol provided otoliths belonging to 53 teleost taxa; nine of those are new for the Mediterranean Pliocene. The association reflects a bathymetry between 150 and 350 m, but it is likely that such depths only existed at the initial stage of flooding of the Llobregat bay. A compilation of the available data for the whole Mediterranean realm at Zanclian time provided a list of 163 taxa, of which 105 could be identified at species level. This last group provides the most useful data for evaluating the composition and affinities of the Mediterranean Zanclian fauna, which appears to be significantly different from the Recent Mediterranean fauna. The Zanclian fauna counts 68 % of Recent species: 12,5 % were already represented in the Mediterranean Miocene, 37 % appeared there at the Zanclian, and 18 % are extra-Mediterranean today. This last group is essentially composed by oceanic fishes, living in the warm parts of the Atlantic and Indo-Pacific Oceans. It is evident that a non neglectable number of Recent species only penetrated very recently in the Mediterranean, which stresses the differences between the more oceanic Zanclian Mediterranean fauna and the present one.


  

A new Tertiary megadermatid is described from Rackham's Roost Site, a Pliocene limestone cave deposit on Riversleigh Station, northwestern Queensland, Australia. It appears to represent the first Australian record of Megaderma GEOFFROY, 1810, a genus otherwise known from Tertiary African and European taxa and the living Asian species M. spasma (LINNAEUS, 1758) and M. (Lyroderma) lyra PETERS, 1872. Megademza richardsi n. sp. is one of the smallest megademiatids known. It exhibits a mixture of plesiomorphic and autapomorphic features, the latter appearing to exclude it from being ancestral to any living megadermatid. The new species is one of eight megadermatids identified from the Australian fossil record, most of which are referable to Macroderma MILLER, 1906. 

New Dendromus species (Rodentia, Muroídea) from Langebaanweg (Pliocene, South Africa). Stratigraphical and paleoecological consequences.

Two new species of Dendromus are described from the Langebaanweg site which precises the evolutionary trend among this genus in South Africa and gives further paleoenvironmental indications. Two evolutionary stages are described: D. darti nov. sp. shows low-crowned molars with bunodont cusps and its more closest relative would be D. melanozis from the Cape region. On the contrary, D. averyi nov. sp. is more lophodont and is better related with the modem D. melanotis. Both species are at a less evolved stage than the Dendromus sp. from Laetolil Beds at Laetoli. The Langebaanweg deposits cannot still be dated by biostratigraphy but they clearly cannot be older than the basis of Pliocene times. The association of Dendromus and Mystromys in the same levels indicates a grassland environment with woodland patches as well as probable swamps. 

A well-preserved anterior cranium of Paranourosorex gigas RZEBIK-KOWALSKA, 1975 from Podlesice, Poland, provides additional knowledge of this rare shrew, including almost complete maxillary dentition. The systematic status of this species as a relative of recent Anourosorex MILNE-
EDWARDS, 1870 is confirmed 

Three new fossiliferous localities, two of karstic origin, Castelnou 3 and Font Estramar, respectively Late Upper Miocene and Lower Pliocene, and one of lacustrine origin, Thuir, Lower Pliocene, add data about the transition between Miocene and Pliocene faunas of rodents in southern France. An unexpected association of taxa was present in the late Upper Miocene, including between others, Myocricetodon, Hispanomys, Ruscinomys, Cricetus barrierei, Promimomys and a new species of Stephanomys, S. dubari nov. sp. Myocricetodon is still known in the Lower Pliocene. It is shown that the large field-mice known since the Late Upper Miocene belong to two different lineages, on one side, A. jeanteti, on the other side, A. gudrunae followed by A. gorafensis. Biochronological and biogeographical implications are discussed. 

The late Pliocene site (MN 16) of Balaruc II (Hérault, France) has provided remains of the following amphibians and reptiles: Chelotriton pliocenicus nov. sp. and Triturus marmoratus (Salamandridae), cf. Rana (Ranidae), cf. Blanus (Amphisbaenidae), cf. Agama (Agamidae),  Gekkonidae indet., Lacerta s.l. (Lacertidae), "Ophisaurus" sp. (Anguidae), Michauxophis occitanus (Aniliidae), Erycinae indet. (Boidae), Elaphe cf. E. longissima and Malpolon sp. (Colubridae), cf. Naja (Elapidae) and Vipera sp. (Viperidae). The salamandrid Chelotriton pliocenicus and the aniliid Michauxophis occitanus constitute, up to now, the only records of these groups in the European Pliocene. The fauna is indicative of a warm, dry
subtropical climate. 

The Mont-Hélène's fauna [Pyrénées-Orientales, France], includes 15 species of rodents with a new one, Occitanomys montheleni n. sp. among the 9 species of the Murids which are listed. The uncommon Cricetid, Blancomys neglectus, is well represented in the fauna. Peculiarities of the population referred to Slephanomys cf. donnezaniare discussed. The locality a fissure filling may be the oldest one of Tabianian age known in Southern France. The diversity of the Murids gives evidence of a subtropical climate and of a diversified environment which may be linked to the spreading of the coastal plain following the filling up of the Roussillon Neogene Basin. 

The first lists of Insectivores (Erinaceidae, Talpidae and Soricidae) from the Pliocene beds of Southern France and North-East Spain are given in this paper. The material from twelve localities is studied. These localities are geographically situated in Languedoc (Celleneuve, Vendargues, Nîmes, Sète, Balaruc 2 and Seynes), in Roussillon (Terrats, Serrats-d'en-Vacquer, Château d'eau and Mont-Hélène) and in North-East Spain (Layna, Medas Islands and Puebla de Valverde). These faunas correspond to the Early, Middle and Late Pliocene. 1 to 8 taxa are identified in these localities and 14 specific taxa are presently listed for this period in this area. Two new specific taxa are described as Galerix depereti nov. sp. from all the Early Pliocene localities in the North-Pyrenean area and as Desmanella gardiolensis nov. sp. from Balaruc 2. For this small mammals, two faunal assemblages are recognized. The first one is dated from the Early Pliocene (F 1, 2 and 3 zones in Aguilar et Michaux) and is characterized by Galerix depereti and rare and little diversified Soricids. The second one is Late Pliocene in age (zones G 2 and G 3). The fossils of the genus Talpa are relatively abundant and the Soricids are diversified and very abundant. The Middle Pliocene (zone G 1) is a transitional period. ln these faunas, most of the insectivore genera are known from the European Late Miocene beds (8 on 10). This fact demonstrates a relative continuity between the invectivore faunas from the Late Miocene to the Early Pliocene. In conclusion, somme paleoecological considerations are suggested.
  

A fragmentary mandible and maxilla of a small sized Agriotherium of a young individual discovered from a Pliocene fissure filling in Xiaoxian county (Anhuei Province, China) are described. Judging from the morphology of the dentition and its dimensions the new material can be identified as Agriotherium inlermedium (STACH l957). Hendey's proposition (1980) that the Agriotherium species are derived from Indarctos is reconsídered on the basis of the new documents. As a result of a more general phylogenetic discussion it can be stated, that: 1. the supposed size increase as well as other trends, leading from Indarctos to Agriotherium are untenable ; 2. there are no positive indications to assume a phylogenetic transition of these two genera. 3. there are no real arguments in favor of an adaptational reversal in the evolution of Agriotherium. Hence, many features of that genus supposed by Hendey to be derived are plesiomorphic ; 4. regardless of the previous points it is methodologícally impossible to establish direct ancestor - descendant relationships between Indarctos and Agriotherium species, as Hendey did. Based on the data available and especially on the characters of the new material from China it is more likely that Agriotherium and Indarctos are two genera which developed independently. While advanced Agriotherium species, e.g. A. africanum, resemble in some respects Indarctos by adaptational analogies, more primitive species, e.g. Agriotherium intermedium, are quite dissimilar to lndarctos. While Indarctos might be derived from an Ursavus like forerunner, Agriotherium has its roots more likely somewhere in between Ursavus and the Hemicyon-group. 

A skull of Parabos cordieri, undamaged by fossilisation, allows us to make an endocranial cast, which is compared to those of Bos taurus, Boselaphus tragocamelus and Hippotragus equinus. These are studied by general form (shape, volume, « rolling-up›› and inclination), and by detailed morphology (gyri and sulci patterns, rete mirabile, importance of rhinencephalon and cerebellum).
A preliminary study is made in order to recognize convolution topography, gyri, sulci, and the various encephalic regions as clearly on an endocranial cast as on a bare encephalon; an intermediate stage is the study of an encephalon covered with «soft» meninges.
The comparative study reveals great similarity between the endocranial cast of Parabos cordieri and that of Boselaphus tragocamelus, both belonging to the tribe Boselaphini. On the other hand, a marked differences separates the cast of Parabos cordieri from that of Bos taurus, representing the tribe Bovini to which Parabos cordieri had been referred. The endocranial cast of Hippotragus equinus, which had long been considered the descendant of Parabos cordieri, is also quite different from that of Parabos. 

The intensive exploration of the Pliocene Hadar Formation, rich in hominid remains, led us to the discovery of several micromammals levels. ln some of them, rodents are very abundant. The stratigraphic repartition of these levels do not cover the whole fossiliferous series of the formation but takes place only in the sedimentary members from Sidi Hakoma and Denen-Dora (rancing from 3.1 - 3.2 MY to 2.8 - 2.9 MY, according to the recent geochronological data). During this gap of time, the species do not show morphological changes, what allowed us to gather, in the same taxa, forms of slighty different ages.
Two striking facts, giving a lot of indications, characterize these small rodents'faunas. First, we notice the domination of the Muridae, as well on a qualitative way (number of species) as on a quantitative one (number of individuals). Then, it appears that, until now, two genera of these murids were known only in the south-western asiatic regions. So, we can suppose continuous biotops between Africa and Indian Subcontinent before 3 MY. In this hypothesis, the hominids had already the possibility to leave their african « cradle ››. Finally, almost all studied genera are still represented at the present time. This fact, previously observed in Laetolil, Omo, Olduvai contributes to remove hope of establishing a biochronological scale based on rodents, in tropical zone. Nethertheless, that allows to try a reconstruction of the palaeoenvironnement, by using the principle of actualism.
  

The study of that locality allowed the description of a new Muridae : Paraethomys athmeniae n. sp. It reveals the existence of new rodent for Algeria : first, a Sciuridae, Atlantoxerus cf. rhodius, and second, a Gliridae, Eliomys truci. So, that work shows the presence of the genus Eliomys in North Africa before the middle of Pleistocene. Lastly, Paraethomys cf. anomalus gives an exact datation of that bed. 

Two steps can be distinguished in the history of the first invasion of western and south western Europe by the arvicolines. The first step corresponds to the installation of these rodents with the immigration of Promimonys inxuliferus Kowalski, then of Mimomys stehlini Kormos and of Mimomys gracilis (Kretzoi). The second is characterized by the establishment of a geographic differentiation in the arvicoline fauna between the south of France and Spain, from where are described new species of Mimomys (Mimamys cappettai, Mimomys septimanus, Mimonys medasensis), and the rest of France, where are found only elements already known from central Europe or England (Mimomys polonicus Kowalski, Mimomys pliocaenicus F. Major, Mimomys reidi Hinton, or forms very close to the latter). This geographic differentiation, which is very certainly the consequence of the division of Europe into distinct climatic provinces, one of them being the southern province comprising at least Spain and southern France, could result from a cladogenetic evolution of Mimomys stehlini and Mimomys gracilis after their immigration. The present work is also a contribution to the search for correlations between the diverse micromammal localities of the latest Pliocene (or early Villafranchian) and of the early Quaternary of Europe. 

The murid fauna of the terminal Pliocene of southwest Europe is rich in at least eight genera and ten species. With the species belonging to the genera Apodemus, Rhagapodemus, and Stephanomys not being studied here, the study of the other murids resulted for one thing in the description of three new genera and three new species: Castillomys crusafonti n. g., n. sp., Occitanomys brailloni n. g., n. sp., Anthracomys meini n. sp., Valerymys ellenbergerí (THALER) n. g., and for another thing in the recognition of a form hitherto unknown in this region, Micromys praeminutus KRETZOI. Systematic study has shown that certain species of the terminal Pliocene fauna had their ancestors in the Turolian fauna presently known in Spain. The evolutionary lineages thereby recognized have been studied more in detail and a list of the evolutionary tendencies of the dendal characters has been given. A chart of the probable phyletic relationships between the different murids of the Pliocene faunas of southwest Europe (With the genus Rhagapodemus and Apodemus dominans being excluded) is given in conclusion of this work.