PURBECK FORMATIONS. 41 



anterior preraolars and of a canine. The molars exemplify the generic didelphid modifica- 

 tion of the multicuspid mechanism for insect-crushing above defined. They sHghtly 

 increase in size from the first to the sixth ; the third {m 3) is a little displaced, exposing 

 the two roots. The molars are close-set, and the anterior oblique lobe of one overlies the 

 posterior talon of the tooth in advance ; but it is much shorter, and more obtuse and pro- 

 cumbent than in Acltyrodon (fig. 7). 



The last premolar (jo 4), with the well-preserved crown, shows distinctly the front and 

 hind basal talon, the main cone rises higher, and the whole tooth is larger than the three 

 succeeding molars. It is divided by the socket of the small first molar from the second 

 one. Impressions of the crowns of two or three anterior premolars, and of that of a longer, 

 sharp, recurved canine, are preserved in the slab of the matrix. 



% XL Genus — Peramus/ Oiven. Plate II, figs. 10 — 13. 



In this genus the teeth, implanted in a jaw characteristically long and slender, are 

 large in proportion, and that in fore-and-aft extent rather than in height. The main outer 

 cone in the true molars has a base almost equal to its height, and offers a marked contrast 

 with the proportions of that oi Acht/rodon. There is an indication of an accessory internal 

 cusp {n, fig. 10, c, PI. II), from the main cone (ib., 0), distinct from the antero-internal basal 

 cusp (ib., e), which, with the general proportions of the main cone, approximate Peramus 

 to Peraspalax. But the five molars retained in the mandible (fig. 10) decrease in size in a 

 greater degree than do those, m 4 — 7 (PI. II, fig. 9 a), in the type of Peraspalax. Such 

 indications have the greater value in the endeavour to interpret materials so minute and 

 fragmentary and often obscure. It is certain that there is not sufficient evidence for the deter- 

 mination of generic conformity of the specimens above cited and compared in their dental 

 system ; and the marked difference in the form and proportions of the mandible, as exemplified 

 in PI. II, fig. 9 {Peraspalax) and fig. 10 {Peramus), sways with me in offering the present 

 evidence for comparison with future possible acquisitions under the above generic name. 



Species 1. — Peramus tentjirostris, Ow. Plate II, figs. 10, 10 a, 10 b, 10 c. 



This species is represented by three fossils. The first to be described is a left 

 mandibular ramus (PI. II, fig. 10, nat. size, and 10 a and b, magn. 3 diam.), wanting 

 the symphysial extremity (probably supplied by fig. 11), with the last four molars 

 in situ, and sockets of five antecedent teeth. In the last three molars the main cone 

 is high and boldly convex across, as in Styhdon, but is broader at the base, and lower 

 than in that genus. The front border is shorter and more nearly vertical than the hind 

 one, at the base of which is a short thick talon (fig. 10c): the anterior basal talon (ib., e) 

 is much smaller : the sharp apex of a small cusp (») rising from the fore and inner part 



' TTTjoa, pouch ; fivs, mouse. 



6 



