tubercle upon the lower sectorial. The superior sectorial is, moreover, less perfect in generally lacking the third lobe or anterior cutting blade, while the premolar formula as a rule is less reduced. To these must be added other characters derived from the skeleton (Fig. 5, B). It is to be borne in mind that these generalized Nimravine forms made their appearance in the oldest Miocene deposits of this country and probably at an earlier date in Europe, if we consider the Phosphorites of France to belong to the upper Eocene, as is done by some palæontologists, so that their ancestry must be sought for in formations older than this horizon. The fact that no representatives of the Viverridæ have yet been found in this country militates against the view of their origin in America, at least, from this source, although of course it is not impossible that migration may have taken place. The same argument applies to the Creodont Miacis, which has thus far been found only in the American Eocene. RELATIONS OF THE CREODONTS TO THE FELIDE.-Schlosser has maintained that the feline phylum has been derived independently from the Creodonts.' Cope formerly held a similar view,' deriving the Felidæ directly from Oxyana. Later he has changed his mind and now considers all the Carnivora as direct decendants of the Miacidæ.' Scott has expressed a similar view' and derives the Nimravida, which in turn gave rise to the Felidæ, directly from the Miacida. With reference to the direct origin of the Nimravidæ from any known Creodont, the evidence heretofore has, however, been so meagre as to compel both Cope and Scott to reject the hypothesis as altogether conjectural. Scott, in his excellent memoir already cited, observes: "No known group of Creodonts can be selected as having any close relations to the cats. The Oxyænas, it is true, do exhibit surprising analogies with this recent family, but the analogy is con Ueber die Beziehungen ausgestorbenen Säugetierfaunen und ihr Verhaltnisse zur Säugetierfauna der Gegenwart. Biologisches Centralblatt, Bd. VIII. Tertiary Vertebrata, p. 264. The Creodonta. American Naturalist, March, 1884, p. 261. Notes on the Osteology and Systematic Position of Dinictis felina, Leidy. Proc. Acad. Nat. Sci., Phila., July 30, 1889, p. 242. fined to the teeth, and is only superficial, as the teeth, which in the two groups look so much alike, are not homologous, and are developed in quite a different way." While this objection holds true of Oxyana it does not necessarily apply to all the Creodonta, and if it can be shown that there is an Eocene genus of Creodonts, which fulfills the necessary requirements, then in our judgment it becomes equally, if not more probable that the ancestry of the Nimravidæ is to be referred to it, rather than to any group which has not been shown as yet to precede it in time. Judging from the tooth and skull structure of the Nimravida, what characters would one be led to look for in their ancestors? We would say that in the upper jaw there should be a short muzzle, well-developed canines, a wide palate posteriorly, rapidly diverging tooth lines, marked tendency to molar reduction, and a large infraorbital foramen. In the lower jaw there should be a moderately deep ramus, straight upon its lower border, a flattened, truncated symphysis, an inferior dental foramen placed well below the tooth line, a distinct scroll-like condylar facet and a reduced second true molar. PALÆONICTIS AND THE FELIDE.—These conditions we have fulfilled completely in the genus Palæonictis; the muzzle is short and the infraorbital foramen large, reminding one at once of Dinictis. The palate is relatively short and broad posteriorly; the second upper molar ( m 2), although present, is reduced quite as much as is the first true molar of the existing cats, having but a single root and a degenerate crown. The fourth superior premolar, while it does not display the perfect blades of the welldeveloped sectorial, nevertheless possesses the required primitive elements from which this tooth has been developed. The blades of the inferior sectorial (pra and pad) produce a distinct 'shear' against the posterior and outer median cusps, as is demonstrated by the increased wear at this point. (See Fig. 5 F, and Plate IV.) In the lower jaw we note the deep ramus with the comparatively straight inferior border; the inferior dental foramen is situated well below the tooth line, the symphyseal region is relatively broad and truncated, and the condylar facet has the distinctively scroll-like pattern seen in the cats. The second true molar is reduced in size, smaller than the sectorial, while the sectorial itself has a relatively large talon and internal tubercle. The two blades of the sectorial are comparatively little developed, and they occupy a position at a considerable angle to the long axis of the jaw. The internal and external tubercles of the anterior triangle form a subsidiary blade which shears against the anterior edge of the internal tubercle of the first upper true molar, as is demonstrated by the wear exhibited in this situation. Fig. 4. Patriofelis ulta, type specimen, National Museum. ramus. Alveoli of teeth from above. Three-fifths natural size. External view of left lower Patriofelis ulta Leidy. As will be noted, Palæonictis has been found thus far only in the lower Eocene (Wahsatch and Suessonian), and before we are. able to establish a connection with the Miocene forms it is necessary to know something of its successors in the Wind River and Bridger formations which lie intermediate. The only specimens known to us from the Bridger which can be related to Palæonictis are the type of Patriofelis Leidy, and an undescribed jaw in the Princeton Collection which Professor Scott has kindly placed at our disposal. We may first consider the former. As shown in Fig. 4, and described by Leidy, the inferior dentition of P. ulta is probably p=3, m=2. The fourth premolar is the best preserved tooth; it is relatively larger than the first true molar, judging by the proportionate measurement of the fangs. The outer face of the crown and the characters of the deuterocone, or posterior basal [September, 1892.] 7 cusp, are closely similar to the corresponding parts in Palæonictis and Ambloctonus. The first true molar of the Patriofelis type is badly broken. The second true molar is proportionately larger than in Palæonictis. ? Patriofelis leidyanus, sp. nov. The Princeton specimen (Fig. 5, C,) may be provisionally referred to the same genus, and distinguished as P. leidyanus. The specific character is that the fourth lower premolar is smaller than the first true molar. It carries the third and fourth premolars together with the first molar or sectorial. It also bears distinct traces of the canine alveolus, which serves to demonstrate that the complete jaw contained but three premolars, the first of which was probably small and single rooted. The length of the tooth line indicates that the jaw was short, resembling in this respect some of the modern cats. The principal interest, however, centers in the first true molar, or sectorial, in which we observe all the elements of the corresponding tooth in the most generalized forms of the Nimravidæ, but also just such an advance over the sectorial of Palæonictis as the sectorial of Felis advances beyond that of Dinietis. In this specimen we note that the two blades are much better developed than in Palæonictis, and occupy a position almost if not quite parallel with the long axis of the jaw. The internal tubercle and talon are much reduced. The posterior faces of the external and internal tubercles are rounded, not flattened as in Palæonictis, and there is no evidence of a secondary shear between this part of the tooth and the anterior edge of the internal lobe of the first upper true molar, which would have undoubtedly been the case if this part of the upper tooth had been well developed. This evidence leads indirectly to the supposition that the first upper molar was considerably reduced, which in turn would. seem to justify the inference that the last upper molar had completely disappeared. While this view is of course inferential, it is nevertheless strongly suggested by what we know of other forms, as, for example, Amblyctonus. The points of wear in one series and a knowledge of the lobes and cusps which produce it |