FOSSIL EEMAINS. 61 



" It was partially imbedded in a layer of blue clay a foot in thick- 

 ness, overlaid by a layer of lighter clay two feet two inches thick, 

 containing- casts and shells of My a arenaria, Macoma subulosa, 

 Mytylus cdulis, Cardium (Serripes) grcenlandicum, Astarte trun- 

 cata, Saxicava distorta. Nucula antiqua, Leda tenuisulcata, L. 

 truncata, Natica clausa and pusilla, and Balanus. The skeleton 

 is in the Museum of the Portland Society of Natural History."* 

 In Europe, Walrus remains were reported by Cuvier t as found 

 at Angers, France, but Gervaisf found later that the only por- 

 tion of those remains accessible to him belonged not to the 

 Walrus, but to the Halitlierium. 



In 1858, however, a part of a cranium was described by Gra- 

 tiolet, from the diluvial deposits of Montrouge, near Paris. He, 

 however, considered it as distinct from the existing species, even 

 generically, and gave it the name Odobenotlierium lartetianum. 

 In 1874, a nearly entire skull was described by Defrance, from 

 similar deposits near the village of Sainte-Menehould, Marne, 

 which he not only considered as identical with the living species, 

 but also referred the fragment previously described by Gratiolet 

 to the same species. Eespecting these specimens he says : 



" En comparant entre elles les tetes du Triclieclms rosmarus 

 de nos iners, de V Odobenotlierium Lartetianum et du Triclieclms 

 de Sainte-Menehould, on leur trouve une resseinblance aussi com- 

 plete que possible, sauf en ce qui concerne la forme et le vo- 

 lume de I'apophyse niastoi'de, point qui presents des differences 

 assez sensible. On sait que dans le T. rosmarus cette apophyse 

 est tres-grande, presque verticale, et saillante la partie infe- 

 rieure du crane; celle de 1' Odobenotlierium, e'galement tres- 

 volumineuse, se prolonge presque horizontalernent en arriere, 

 sans de"passer le crane inf6rieurement ; celle du Triclieclms de 

 Sainte-Menehould pr6sente un volume plus consid6rable encore 

 que dans les deux autres, sans se prolonger en arriere coinme 

 dans 1' Odobenotlierium , rnais inferieurement comme dans le 

 Triclieclms actuel. Ces nuances 16geres indiquent evideminent 

 une e"troite parent^ entre ces trois individus ; aussi est-il diffi- 

 cile de conrprendre que Gratiolet ait voulu 6tablir un nouveau 

 genre sur des particularites pen accentue"e que celles que lui 

 pre"sentait la portion de crane dont il etait possesseur, et qui ne 



" American Naturalist, vol. xii, p. G33, Sept. , 1878 ; see also Portland (Maine) 

 Argus, of July , 1878. 

 tOssem. Foss. 



tZool. et Pal6out. Frangaises, 1859, p. 88. 

 {Bull. Soc. Ge"ol. de France, 2 s6r., xv, 1858, p. 624. 



