Avec ce volume se clôture l'étude géologique et paléontologique des sites à hominidés de Koobi Fora. Il fait suite aux deux précédents ouvrages, parus respectivement en 1978 et 1983, consacrés également à l'étude des faunes recueillies sur ces gisements dans le cadre du Koobi Fora Research Project. 

The study of dental variations in Landenodon from Dormaal (early Eocene, Belgium) shows that only one
species is present in this locality : L. woutersi QUINET, 1968 (Condylarthra, Arctocyonidae, Arctocyoninae). A high morphological variability is described, especially concerning the metaconid of P/4, the paraconid of M/2, and some characters of M1/ which have been found to be independant of each other. This variability is discussed. 

La découverte de nouveaux gisements à micromammifères dans le Miocène du Sud de la France et d'Espagne montre que, s'il n'est plus possible d'accepter dans son intégralité le schéma phylétique de Fahlbusch du genre Megacricetodon, il existe toutefois des segments de lignées à l'intérieur de provinces géographiques différentes (Péninsule Ibérique, France, Europe Centrale), avec lesquels on peut établir une biochronologie locale, segments qu'il est aussi possible de placer les uns par rapport aux autres, en utilisant les corrélations que l'on peut reconnaître entre certains niveaux repères des mammalogistes et l'échelle stratigraphique marine.
En utilisant ces données nouvelles, et d'après l'étude de la première molaire inférieure, nous proposons le
schéma évolutif suivant, pour les espèces de grande taille :

1 - au début du Burdigalien supérieur. une espèce commune, M. collongensis, issue de la forme ancestrale M. primitivus, est présente, de la Péninsule Ibérique à l'Europe Centrale. Son évolution est mal connue en Espagne, par contre cette lignée s'éteindrait en France au Burdigalien terminal avec le stade M. gersii nov. sp. de Sansan et de Luc-sur-Orbieu.
2 - au cours de la régression post-burdigalienne. reconnue en Languedoc et dans la Péninsule Ibérique apparaîtrait une espèce immigrante, M. crusafonti qui, dans cette dernière province, évolue pour donner M. ibericus de Nombrevilla. L'espèce M. gregarius de La Grive M et de La Grenatière dériverait de cette forme immigrants, en ayant une évolution différente.
3 - en Suisse, Bavière et Europe centrale, la lignée M. primitivus - M. collongensis va s'individualiser, dès le Burdigalien supérieur ou terminal, et donner naissance à la lignée M. bavaricus qui s'éteindra avec le stade M. germanicus nov. sp. du gisement anté - Hipparion de Anwil. Cette dernière espèce a été trouvée pour la première fois en France, dans le gisement de Collet Redon. 

Rodents from the new localities of Châteauredon (Alpes-de-Haute-Provence) and La Denise (Bouches-du-Rhône) show more similarities with the rodents known from the Czecho-Slovak locality of Franzenbad than with the species found in contemporaneous Lower Miocene faunas located westward of the present Rhône Valley. This is another data wich support the existance of several biogeographical provinces in Southern France during Lower Miocene. The correlations which have been settled down allow new datations of several formations in Provence and give new information on the Aquitanian and Burdigalian paleogeographies of this region. More precise stratigraphical allocations are given to some other localities of Central Europe. 

Until now Aplodontidae of the European Oligocene have been documented by four species only. The phylogenetic relations remained obscure. as the distribution of only one species has been known in some detail. New material made it possible to define the stratigraphic range of two of the already existing species (Plesispermophilus angustidens, Sciurodon cadurcense) and to follow their development during the Oligocene beginning with the event of the « Grande Coupure ››. Sciurodon remained nearly without change until the end of the Middle Oligocene. Plesispermophilus angustidens split into two distinct phyletic lines, one of which (P. macrodon n. sp.) reaching considerable size, is represented till the beginning of the Upper Oligocene (Pech de Fraysse, Gaimersheim). The other line leads to Plesispermophilus ernii (basal Upper Oligocene of Burgmagerbein 1. terminal Upper Oligocene of Coderet). Besides the already known forms a new small-sized species (P. atavus n. sp.) is described, which by its primitive features closely resembles the genus Plesispermophilus. Two other small-sized species already known from the Upper Oligocene (? P. argoviensis) and Lower Miocene (? P. descedens) seem to be closely related to the new species. It cannot be decided whether they are descendents of this line or have developed independently, because of their poor fossil record.
Comparison of the evolutionary modalities in the different phylogenetic lines reveals general trends. the most striking of which is the complication of the tooth pattern by the development of additional crests. In the lower molars the cusps diminuate in size and are more and more transformed into ridges. ln addition new connection between the crests appear. in the upper molars, the « selenodont » shape of the teeth becomes more and more dominant, and in the two main evolutionary lines of Plesispermophilus the metaconulus becomes duplicated. A further evolutionary trend is the size increase of the premolars compared to the molars, which is even more pronounced in the Miocene Aplodontidae.
Phylogenetic relations between the primitive Plesispermophilus and certain « prociurines ›› of Northern America as well as between Plesispermophilus (P. angustidens) and more progressive forms of the Upper Oligocene (P. ernii, P. macrodon n. sp.) can be documented. In this light, the taxonomic distinction between Prosciurinae (bunodont) and Allomyinae (selenodont) sensu Rensberger 1976 can be shown to be artificial, because it separates forms from each other, which are evidently closely related. Consequently the separation into two subfamilies has been abolished. 

The geomyoid genus Dikkomys is well represented in a sample from the Black Bear Quarry Il local fauna of Early Hemingfordian age in Bennett County, South Dakota. Isolated unworn P/4's of Dikkomys matthewi WOOD have a prominent median cristid (sagicristid) with a connection to the metaconid and the hypolophid. With wear, P/4 does not become as molariform as P/4 because of this cristid.
A large sample of the Whitneyan beteromyid Proheteromys nebraskensis WOOD contains variants of the P/4 with on incipient sagicristid in approximately 18 percent of the population. The upper dentition and lower molars of Proheteromys nebraskensis are sufficiently generalized to indicate probable ancestry to Dikkomys. 

A sample of 231 isolated teeth of lagomorphs from the upper Miocene of Montredon (southern France), identified as the ochotonid Prolagus crusafonii DOPEZ, 1975, is studied, comparing it with other populations of the same species as well as with its closest species P. oeningensis (KÖNIG, 1825).
The maintenance dining 5 m.y. of a high morphological variability in a group of non-selected features is interpreted as an evolutionary response. This may be related with the heterogeneity and probably too with the regression of the environment. 

As a result of the enrichment of eotheriodont material by one of us (P.K.T.), these specimens (essentially Biarmosuchur and Eotitanosuchur) are reexamined and refigured. A reevaluation of their particularities supports the distinction of two families, for which new diagnoses are proposed. This leads us to discuss the affinities of these families, with respect to the sphenacodonts on one hand, and to the South African primitive theriodonts on the other (gorgonopsids and ictidorhinids). This study contains inherent paleogeographic consequences which are considered in conclusion. 

The study of another faunal assemblage (mostly macromammals) from Torralba de Ribota (Calatayud, Zara-
goza Prov.) demonstrates the Middle "Burdigalian" age of the deposit, MEIN zone 4a. Some ten genera have been recognized. a.o. Anchitherium, Aceratherium and Lagopsis. 

The rodents from the spanish locality of Can Ponsic 1 bring new data about some rodents species of the beginning of the Upper Miocene in South-West Europe. The criticims made by Mein and Freudenthal about the validity of the species Hispanomys thaleri from Can Llobateres are not justiíied. The study of the anatomy of the skull of Rotundomys from Can Ponsic 1 gives accurate information about the affinity of this genus with Cricetulus, and shows that the hypothesis, according to which Rotundomys is an ancestral form of the Arvicolids, is unlikely. The systematics of Heteroxerus and the phylogeny of the mio-pliocene Muscardinus species are also discussed. The Can Ponsic 1 locality is a little older than Can Llobateres.

  

A reconstruction of a multituberculate endocranial cast is made on the basis of a complete natural cast prepared from the skull of Chulsanbaatar vulgaris, and other less complete endocasts, all from the Late Cretaceous of Mongolia. The multituberculate endocast is of mammalian pattern but it has retained a therapsid-like lateral profile with a deep rhombencephalon and a shallow telencephalon. It is characterized by: a heart-shaped cross-section of the telencephalon; an extensive lissencephalic neocortex; a very prominent pons placed far anteriorly; a lack of cerebellar hemispheres, and very large paraflocculi. Its structure, very different from the brains of other mammals, suggests thats the Multituberculata branched very early from the main mammalian stock. This supports Simpson's (1945) idea that the Multituberculata should be placed in a subclass of their own: Allotheria MARSH. The endocast and braincase structure show that the Multituberculata had strongly developed senses of smell and hearing. The encephalization quotient of approx. 0.55 evaluated for Ch. vulgaris is relatively high for a Mesozoic mammal. 

The complex taxonomic history of the paramyid rodent genus Ailuravus is reviewed. It has been described as Hyracotherium, as a creodont carnivore and as a lemuroid primate - errors at the ordínal level that are most unusual for a rodent. The genus is a member of the poorly known subfamily Ailuravinae, probably derived from some European Early Eocene species of Paramys. Aíluravus was a large arboreal paramyid with highly rugose cheek teeth, very well developed hypocone, and a remarkably weak lower incisor. It was tropical to subtropical. Three named species are recognized, A. macrurus from the Lutetian of Messel; the genotype, A. picteti, from Egerkingen, Buchsweiler and the Geiseltal, slightly later in the Lutetian; and A. stehlinschaubi, new name, from the Bartonian of Mormont-Eclépens and Robiac. One or more unnamed species are present in the Ypresian of Cuis. The species are close to a phyletic sequence. No later representatives of the genus are known. The late Eocene to earliest Oligocene North American paramyid Mytonomys, whose relationships have been obscure, is tentatively referred to the Ailuravinae. 

The present work is devoted to the study of the Cretaceous and Tertiary fishes (teeth of Selachians, Actinopterygians and Dipnoans) collected during a recent expedition in Niger. The Maestrichtian localities have yielded a new genus and a new subspecies of Selachian: Igdabatis sigmodon nov. gen., nov. sp. and Lamna biauriculata nigeriana nov. subsp. The locality of Sessao, which has been attributed to the Thanetian by means of the study of the fish, has furnished by screen-washing an interesting fauna wherein six new species are described: Raja Iouisi, Dasyatis sessaoensis, D. sudrei, D. russelli, Hypolophites thaleri and Ceratodus casieri. Comparison of these faunas with contemporary faunas of Africa has brought out a certain endemism in the Iullemmeden Basin during the late Cretaceous and the early Tertiary. 

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. 

Based on material from 30 localities, morphologic dental, cranial and biometric analyses have permitted the characterization of two parallel Issiodoromyine lineages, and also the definition of diverse species representing several evolutive stages. Thus it is that new lineages complete the contribution made by the Theridomyinae and Cricetidae and permit, for the Quercy in particular, additional precision in the biochronologic succession of the localities. One of the lineages is limited to the genus Pseudoltinomys LAVOCAT; the other evolves from the genus Elfomys HARTENBERGER to the genus Issiodoromys BRAVARD in GERVAIS. The latter is affected by profound anatomical changes due to a functional modification of the mastication apparatus. These changes seem to be able to be put in relation with the aridification and cooling of the climate at the end of the Eocene. At the end of the middle Oligocene, a new chewing structure is achieved. It is found in diverse living rodents that inhabit a rather arid steppe environment (Cavia, Pedetes, Ctenodactylus). To these supposed nearby ecologic conditions, these rodents have responded in a convergent fashion. It is possible to attribute to the extreme specialization of Issiodoromys its incapacity to adapt to the new climatic crisis of the end of the Oligocene. The arrival of immigrants may be considered as another cause of its disappearance at this time, complementary or not with the first. 

Cranial anatomy of a late Miocene rhizomyid, Brafhyrhizomys cf. B. pilgrimi, provides new evidence on the origin of the dorsal, round infraorbital foramen of living rhizomyines. Primitive rhizomyids retain a myomorphous keyhole foramen with a long ventral slit that retracts upward during the evolutionary history of the Rhizomyidae. The primitive condition of the elongated ventral slit is represented by Kanisamys sivalensis. Among later burrowers the foramen shows progressive dorsal migration, the ventral slit terminating midway up the snout in B.tertracharax and B. choristos ; well above the midline of the snout in Brachyrhizamys cf. B. pilgrimi. Apparently this transformation began earlier among Rhizomyinae than among Tachyoryctinae and continued to a more derived stage in rhizomyines. ln living Rhizomyx the ventral slit is absent and only a high round hole remains at the anterior end of the zygomatic arch. 

The skull remains referred to the genus Palaeotherium are recorded and described. Biometrical tests are made to elucidate intrageneric allometric relationships and to allow comparisons with various other perissodactyls. Apart from the well known shortness of post canine diastems and deepness of the narial opening, the genus is characterized by a great lengthening of the splanchnocranium, owing to a spreading of the post-orbital facial region, by a reduced area of the eye-socket and by the prevalence of the temporal muscle with regard to the masseter; this original shape of the masticatory apparatus needs to be related to the morphology of the jugal teeth and particularly to their  asymmetrical semi-hypsodonty.
    These animals, whose running ability was evidently poor, appear to have been adapted to rather closed environments, feeding on relatively soft vegetable matter; olfactory sense was likely to play a leading part in interindividual and environmental relationships. Such evolutionary trends might explain the disappearance of most of them, as clirnatic conditions deteriorated at the end of the Eocene, before the "Grande Coupure" which affected mammalian faunas at that time.
      Although the present paper is not directly concemed with phylogenetics, it invalidates the supposed ancestor-descendant relationship between P. castrense and P. magnum, and it suggests a possible emergence of the P. medium lineage from a P. siderolithicum stock. Moreover, the structure of the post-orbital facial area allows the  erection of a new sub-genus, Franzenitherium, for the  species lautricense and duvali.

  

The revision of the Lutetian chiropterans from Messel, first described by Revilliod in 1917, is based on the anatomy of the teeth and the skeleton.  A figuration or refiguration of thematerial utilized accompanies the new description, which goes beyond that of the original monograph.
The study shows a certain variability of the dental structure within the genera Palaeochiropteryx Revilliod and Archaeonycteris Revilliod,  as well as a general resemblance of the two forms. The morphology of the teeth permits, however, the verification of the validity of the different species: Palaeochiropteryx tupaiodon Revilliod, P. spiegeli Revilliod, Archaeonycterís trigonodon Revilliod, and Archaeonycteris revilliodi, n. sp.
Some differences of the skeletal and dental anatomy tend to indicate a stage of evolution less advanced for the genus Archaeonycteris.
The comparison of the chiropterans of Messel with the principal groups of living chiropterans, as well as with different Eocene fossíls (notably Cecílionycteris Heller and Icaronycteris Jepsen) leads to a more precise idea of the anatomy of primitive chiropterans. This comparison also permits the proposition that the oid forms so far described by integrated in a superfamily, the Palaeochiropterygoidea and allows   a general phylogenetic hypothesis to be advanced for the order Chiroptera. 

In recent years, the techniques of chemical preparing have permitted a rich paleontologic material to be obtained from the phosporitic sediment of Bouzigues (Hérault, France). The fauna of this locality is comprised of quite varied microvertebrates, amphibians, reptiles, birds, mammals. Twenty five species of the latter, belonging to seven orders, are today known from the site. Among them, the numerous rodents have allowed L. Thaler to chronologically situate this fauna in the Zone of Laugnac (<< late Aquitanian ›> of some authors).
The chiropterans are, with the rodents, the best represented of the locality's mammals. Three families comprise the bat fauna, with nearly complete dominance by one of them (Hippoxideridae) over the two others (Megadermatidae and Vespertílionidae)
Six forms are described, of which three are new species and one a new sub-genus.
Megaderma braillomi n. sp., an animal of rather large size, shows like the Miocene megaderms several evolved dental characters, translating the adaptation of these animals to a partially carnivorous regime. The Bouzigues species seems, however, to represent a particular lineage.
Hipposideros (Brachipposideros n. subgen.) dechaseauxi n. sp. and Hípposideros (Brachipposideros) cf. collongenris Depéret, small sized forms, belong to a group rather well represented in the late Oligocene and early Miocene of Europe, and not distinguished until now within the genus Hipposideros.
Hipposideros (Pseudorhinolophus) bouziguensis n. sp., is the most abundant mammal in the locality and, occuring at the Oligocene-Miocene limit, the last representative known of the subgenus Pseudorhínolophus, common in Europe from the middle Eocene.
However, beyond Neogene and Quaternary times, certain among the numerous living species of Hipposideros are close to Pseudorhinolophus and others to Brachipposíderos. 'This fact would in the future justify a global revision of the genus, on the basis of comparative anatomy of the squeleton and of the teeth.
The bat fauna of Bouzigues is completed by two small Vespertilionidae, rare forms, Myoris sp. I and sp. II.
  

The formic acid preparation of a nodule collected from the Lower Triassic of Spitsbergen, and containing a skull with some associated postcranial elements of Lyrocephaliscus euri (WIMAN), precises and enriches previous descriptions of the reputedly well-known trematosaur. lt permits more detailed descriptions of the occipital region, of the course of the carotid arteries and collateral vessels, and refutes the previous description of a stapedial foramen in trematosaurs. The discovery of pleurocentres associated with the other vertebral components confirms the primitive rhachitomy of trematosaurs.