1874.] 



ON FOSSIL ARVICOLID.E. 



465 



brick-earth of the Drift period at Fisherton, near Salisbury, asso- 

 ciated with Myodes torquatus, Spermophilus erythrogenoides, &c. ; 

 and these we find to be certainly the same as the Somersetshire 

 specimens. A skull from the Bromberg cavern (from the Soem- 

 mering collection) is in the British Museum. 



On comparing the dental pattern of these fossils with Blasius's 

 figures of A. ratticeps (Saugeth. Deutschl. p. 366) we were struck 

 with the great variety of form in the anterior part of the first lower 

 molar, and more especially in the posterior part of the third upper 

 molar. We were at one time inclined to believe that the fossil might 

 be regarded as a distinct species, and are still of opinion that it may 

 represent a race distinct from the recent A. ratticeps ; but the varia- 

 tions are so great, both in the recent and the fossil skulls, that we 

 have sought in vain for any constant characters for specific distinction. 



In the first lower molar many specimens agree exactly with A. 

 ratticeps, but in others the anterior extremity is produced beyond 

 the first inner angle, so as to give the tooth the appearance of 

 having eight cemental spaces instead of seven : this type is some- 

 what exaggerated in Mr. Sanford's fig. 1 d. The outer margin of 

 the first two blended spaces is often less regularly convex than is 

 usually the case in recent A. ratticeps, so that the whole tooth 

 rather resembles Middendorff's illustration of A. obscurus ('Sib. 

 Reise,' ii. pi. xi. fig. 4), and Hensel's of his A. ambiguus ( = A. 

 brecciensis, see p. 466). 



But the most remarkable variation occurs in the third upper 

 molar, which differs so much that it is only the numerous interme- 

 diate forms that convince us that all belong to the same species. A 

 few, like that figured by Mr. Sanford, agree with typical A. ratticeps 

 in having four external and four internal angles ; but in many there 

 are only three external angles, and the whole form of the tooth more 

 resembles that of A. arvalis, while others present a type peculiar to 

 themselves. All we have yet examined differ from A. brecciensis in 

 having more than three internal angles (fig. 2, a-f). 



Fig. 2. 



Teeth of A. ratticeps and A. nivalis. 



