NEW GENERA AND SPECIES OF COELACANTHS FROM THE BEAR GULCH LIMESTONE (LOWER CARBONIFEROUS) OF MONTANA (U.S.A.) by RICHARD LUND * & WENDY LUND * ABSTRACT RI~SUM]~ Four new spegies of coelacanths (Actinistia) from the Namurian E2b marine Bear Gulch Limestone of Montana are described, and a fifth species, originally described as an actinopterygian, is redescribed. Two new suborders are diagnosed. The taxonomy is : Suborder Coelacanthocdei, family Rhabdoderma- tidae, Caridosuctor populosum nov. sp., suborder Hadronectoroidei nov., family Hadronectoridae nov., Hadronector donbairdi nov. sp., Polyos- teorhynchus simplex nov. sp. and Allenypterus mon- tanus MELTON. Lochmocercus aciculiodontus nov. sp. is incertae sedis. Quatre nouvelles esl~ces de Coelacanthes (Actinis- tia) sont d6crites dans le calcaire namurien marin E2b de Bear Gulch (Montana, U.S.A.), et une cinqui~me, d~crite auparavant comme Actinopt&ygien, est red& trite. La taxonomie suivie dans cet article est : Sous- ordre Coelacanthoidei, famille Rhabdodermatidae, Caridosuctorpopulosum nov. sp., sous-ordre Hadro- nectoroidei nov., famille Hadronectoridae nov., Hadronector donbairdi nov. sp., Polyosteorhynchus simplex nov. sp. et Allenypterus montanus MELTON. Lochmocercus aciculiodontus nov. sp. est incertae sedis. KEY-WORDS : ACTIN1STIA (COELACANTHIFORMES), CARBONIFEROUS, NAMURIAN, MONTANA, NEW TAXA. MOTS-CLI~.S : ACTINISTIA (COELACANTHIFORMES), CARBONIFt~RE, NAMURIEN, MONTANA, NOUVEAUX TAXONS. INTRODUCTION The osteichthyan order Coelacanthiformes (Actinis- tia) includes fishes ranging from Late Devonian to the recent. Known fossil coelacanths differ from each other only in proportions or in relatively small details of osteology. The morphology of the Palaeozo~'c coe- lacanths, however is known only for Diplocercides ( = Nesides) (Stensi0, 1922, 1937), Rhabdoderma (Moy- Thomas, 1937 ; Forey, 1981), and Coelacanthus (Schaumberg, 1978), from the Late Devonian, Late Carboniferous and Early Permian respectively. Of these three genera, moreover, the description and dia- gnosis of Rhabdoderma were not based upon the type species of the genus and are seriously in error (Moy- Thomas, 1937 ; Lurid & Lurid, in press). The osteo- logy of the Palaeozoic coelacanths, therefore, is so poorly known that plausible hypotheses of relations- hips are not presently possible. * Department of Biology, Adelphi University, Garden City, N.Y. 11530 U.S.A. Geobios, n ~ 17, fasc. 2 p. 237-244, 5 fig. Lyon, avril 1984