426 MR. n. I. rococK on the kxtekxal 



characters which distinguish Dolichotis from Cavia in tlie old 

 sense, and llydrochoerus fi'om both. In the first place, llydro- 

 choerus clearly difiers from DoUchotis and CoAna much more than 

 these two differ from each other. As set forth in the table that 

 follows, DoUchotis also in certain well-defined particulars stands 

 aside from all the genera into which Cavia has been split up. I 

 propose to express these differences by severing Hydrochm-us as 

 the representative of a distinct family * from the Caviidas, and to 

 divide the latter into two subfamilies Dolichotinse and Caviina^. 

 The distinguishing characters of these groups may be briefly 

 stated as follows : — 



a. Last uppei- molar longer tliaii tlie tliree others taken together; palate 

 long, its median line about twice its width, the palatine bones 

 not deeply excised by short mesopterj-goid fossa. Feet webbed 

 up to ill-defined digital pads, which with the claw form a hoof- 

 like termination to the digits ; digits of fore foot unevenly 

 spaced, the fifth set high above the fourth and carrying a 

 reduced and se])arate interdigital element of the plantar pad ; 

 the third digit markedly larger than the second or fourth and 

 submediau ; no additional ridge on ear HYDEOCHCERiDJii. 



h. Last upper molar much shorter than the combined length of the 

 others ; palatine bones so deeply excised by the mesopterygoid 

 fossa that the median length of the palate is at most about equal 

 to the width between the last molars. Feet not webbed ; digital 

 pads well defined and distinct from the compressed claws; digits 

 of fore foot tolerably evenly spaced : third and fourth digits 

 nearly evenly paired, subequal, the third not markedly larger 

 than second and fova-th ; plantar pad with external lobe con- 

 fluent with median lobe ; a strong ridge beneath the supratragus 

 in the ear Caviid,i;. 



«-. Skull with nasals narrowed and strongly convex in front, the 

 distal edge deeply excised ; the interorbital region very wide 

 owing to an expansion, notched anteriorly and defined by a 

 notch behind, overhanging the orbits; upper edge of malar 

 geniculate, paroccipital processes long. Ear long, with simple 

 supratragus. Nostrils expanded. A single pair of widel}^ 

 separated anal glands above the anus. Digits short, legs long, 

 with area behind plantar pads overlapped by hairs ; plantar 

 pads very large, projecting behind; carpal pad very small; 

 a distinct tail BoUchotina. 



Ir. Skull with nasals nearly parallel-sided, not strongly convex in 

 front, and distal edge « ith shallow excision ; no appreciable 

 bony growth roofing the orbit, upper edge of malar not geni- 

 culate ; paroccipital processes comparatively short. Ear short, 

 M'ith valvular supratragus. Nostrils not expanded. Anal 

 glands, where known, opening beneath the anus. Digits more 

 elongated, legs short, naked beneath, plantar pads not project- 

 ing behind; carpal pad large ; no tail Caviince. 



Other differentiating characters might have been added to 

 those given above. But these must suffice. 



* I submit that the characters upon which Hi/drochoerus may be separated as a 

 family from the Caviidee are of higher systematic value than those upon which 

 Castor has been separated as a distinct fam'ily from the Sciuridse (see Miller, Mamm. 

 of Western Europe, p. 947, 1&12). 



