212 DK. C. J. FORSYTH MAJOR ON MIOCENE SQUIRRELS. [Feb. 28, 



opinion allows only for a movement in the longitudinal direction), 

 this mechanical action of the jaw would be followed by the wearing 

 away first of all of the middle row of the upper molars, which rubs 

 against the inner and outer side respectively of the outer and inner 

 row of the opposing tooth. Compared with the types of " Tri- 

 priodon ccelatus " ' and " caperatus " ^, Dipriodon rohusUis shows its 

 inner cusps nnproportionally extending outwards, and this would 

 justify the supposition that the remnauts of two cusps of the 

 middle row have united with what remains of the enamel-folds of the 

 two inner cusps, a mode of coalescence which we often see realized 

 in worn molars. I therefore fail to find in Marsh's previous publi- 

 cations the proofs of the statement made by him ^ that one series 

 of upper molar teeth of Cretaceous Allotheria has but two rows ; 

 although from certain analogies — with the molars of Mus on the 

 one side, those of Cricetodon on the other — analogies which may 

 hereafter turn out to be homologies, I am quite willing to admit 

 that Prof. Marsh may have in his hands the means for proving it. 

 I cannot admit Osborn to have satisfactorily shown that in this 

 group of Cretaceous Allotheria there are lower molars with three 

 rows of cusps. If there were three rows of cusps with two longi- 

 tudinal grooves in the lower molars, we would have to urge four 

 rows of cusps ^^■ith three longitudinal grooves in the upper molars ; 

 these have not been forthcoming up to the present date. Thus there 

 seems to be no reason for denving Marsh's statement that "the lower 

 molars .... although differing widely in the form and structure 

 of their crowns, have only two parallel series of crescents or 

 tubercles, an outer and inner row, with a groove or valley between 

 them"\ 



For my present purpose it may be sufficient to point out, that 

 both authors agree in stating that there are several forms with 

 upper molars composed of three longitudinal rows of cusps \A-itli 

 two grooves between them, to which correspond Uvo lougitudinal 

 rows with one groove in the lower molars ; but, as stated before, 

 they have not furnished sufficient proof for their opinion that 

 these molars belong to an aberrant Order of Mammalia. In 

 comparing the teeth in question with those of the Mnridae and 

 more particularly of Mus, it becomes evident that the main 

 differences between them consist in this, that whilst in the 

 Cretaceous molars the prevailing division is effected by longi- 

 tudinal grooves, in Mus, on the contrary, the molars are deeply 

 divided by transverse grooves : the longitudinal grooves in this 

 genus, of which there are two in the upper, and one in the 

 low er molars, being relatively shallow. For further particulars on 

 this argument, as far as it relates to Muridse, I refer the reader to 

 a paper by Hensel on Mus orihodon from the Ossiferous Breccias 



1 L. c. part I. pi. ii. figs. 19, 20. 



2 L.c. partIII.pl. v. fig. 2. 



3 L. c. part III. p. 253. 

 * lb. p. 253. 



