1876.] MR. E, R. ALSTON ON THE ORDER GLIRES. G9 



Of the seven well-marked families into which this section may be 

 divided, the typical one, the Muridce, comprises a great number of 

 genera. The best classification of these with which I am acquainted 

 is that of Dr. Peters, which is here adopted with some little modifi- 

 cation, his groups being ranked as subfamilies, and a slight alteration 

 being made in their arrangement*. M. A. Milne-Edwards having 

 clearly proved that the genera £'//o6im5 and Siphneus really belong to 

 this familyt, the subfamily Siphneince is now placed alongside of the 

 Arvicolince, with which it is so nearly allied. 



Of the other families, the Myoxidce bear a very strong outward 

 resemblance to the Sciuromorpha, which, however, is not mark- 

 edlj' confirmed by their anatomy. Dr. Peters having shown that 

 Platacantliomys must be removed to the MuridceX, the remaining 

 genera of Dormice are all very closely allied, and are isolated from all 

 other known rodents by the complete absence of the caecum. The 

 next family, Lophiomyidce, contains a single form differing in struc- 

 ture not only from all the rest of the order, but even from all tlie 

 known members of the mammalian class§. Nevertheless, if the ex- 

 traordinary development of the temporal and malar regions be over- 

 looked, the whole skull of Lophiomys is truly murine in type ; and 

 this is confirmed by all the other more important points in its 

 anatomy. It is strange that, although its habits appear to be at least 

 partly arboreal, Lophiowys should differ from all the rest of the 

 section in the incomplete development of its clavicles. 



The Spalacida, even when disencumbered by the removal of 

 Siphneus and Ellobius, are still divisible into two subfamilies — the 

 typical Spalacince, which have the normal mandible of the section, 

 and the Bathyergince, in which are found the hystricine characters 

 alreadi' mentioned {supra p. 68), and which were hence named Spa- 

 laces subhystriciformes by Prof. Brandt. The next family, which in- 

 cludes the American rodents with cheek-pouches which open outside 

 the mouth, was founded by Mr. Waterhouse under the name ofSacco- 

 myidfs, and subsequently divided by Prof. Baird into two subfamilies, 

 GeomyincB and Saccomyince. Dr. E. Coues, in a recent valuable 

 memoir, has contended that these latter divisions should rank as 

 separate though allied families|| — a view in which I cannot agree. 

 The diversity in their outward form may be paralleled by that in the 

 Squirrels and Marmots ; and the differences in their cranial structure 

 are, as Dr. Coues himself observes, of a superficial nature. In any 

 case, Mr. Walerhouse's name must be changed ; for Dr. Peters has 

 shown^ that the genus Saccomys of Frederic Cuvier is in all proba- 

 bility, a synonym of Desmarest's Heteromys. The oldest and best- 



* The names only of these divisions appeared in the Monatsb. Ak. Berlin, 

 1866, pp. 13, 14. ; for access to Dr. Peters's hitherto unpublished characters I 

 am indebted to the courtesy of the author. 



t Eecherches pour servir a I'Hist. JMat. des Mammiferes, pp. 71-129. 



\ P. Z. S. 1865, pp. 397-399. 



§ Cf. A. Milne-Edwards, Nouv. Arch, du Mus. iii. pp. 81-118. 



II Rep. Explor. Colorado River, p. 215; Bull. U.S. Survey of Territories, 

 2nd ser. no. 2, pp. 81-90. 



^ Monatsb. Ak. Berlin, 1874, p. 354-359. 



