596 
PACIFIC SCIENCE, Vol. XV, October 1961 
Evotomys (= Clethrionomys). Its heavy and ap- 
parently imperfectly rooted teeth more closely 
resemble those of many species of typical Mi- 
crotus than they do the weak, perfectly rooted 
teeth of true Evotomys . The relationship of the 
root of the lower incisor and the posterior 
lower molar, while not typical of either genus 
are clearly suggestive of Microtus rather than 
Evotomys . The palate structure, on the other 
hand, appears to agree with that of Evotomys. 
Also, in a key to the Arctic species of red- 
backed voles Miller (1898^: 359) separated rufo- 
canus by the comment: "Teeth large and heavy 
as in Microtus (never perfectly rooted?). . . 
In characterizing rufocanus, Hinton (1926: 245) 
also emphasized its approach to the species of 
Microtus and on page 215 pointed out that in- 
dividuals look mature "long before the molars 
show the slightest sign of rooting.” To distin- 
guish rufocanus from the other species of Cle- 
thrionomys Miller ( 1900) created the subgenus 
Craseomys with the following characters: 
Skull as strongly angular as in Microtus, the 
postorbital processes well developed; teeth rel- 
atively as large as in Microtus, the molar row 
about equal to the diastema; roots of molars 
developed late in life; root of posterior lower 
molar lying in a distinct capsule on the lingual 
side of incisor root. 
Soon after Thomas ( 1905) placed the Japanese 
forms bedfordiae and andersoni in the subgenus 
Craseomys. Later (1907), he used Craseomys 
as a genus for Craseomys regulus ( = Clethri- 
onomys rufocanus regulus) from Korea and 
noted that not even the oldest in the series of 
18 specimens showed any trace of roots on the 
molars. 
Aschizomys lemminus was described as a new 
genus and species by Miller (1898£) on the 
basis of a single specimen and was characterized 
in these terms: "Palate as in Clethrionomys. 
Molars small and weak, as in Clethrionomys, 
but teeth growing from a persistent pulp as in 
Microtus) which strongly displaces root of 
large posterior lower molar. Plantar tubercles 
six. Number of mammae unknown.” The illus- 
trations accompanying the original description 
indicate other features common to both lemmi- 
nus and rufocanus: the encapsulated roots of the 
upper second molar and the lower third molar 
and the reduced lateral tubercles on the rostrum. 
The upper third molar of lemminus is longer 
than in rufocanus and there is little doubt that 
the two species are separate. The third upper 
molar of Clethrionomys niigatae of northern 
Honshu is long and similar to lemminus. In the 
original description of Aschizomys Miller recom- 
mended that rufocanus was allied to lemminus 
and later Imaizumi (1957: 199) suggested that 
Clethrionomys niigatae might well be placed in 
Aschizomys. Also, Hinton (1926: 43) stated that 
lemminus seemed to be very close to rufocanus. 
However, as Miller suggested ( 1898 ^: 359), the 
molars of rufocanus form roots very late in life; 
it is quite possible that rufocanus and lemminus 
are alike in this respect as well, and that a large 
series of Aschizomys lemminus would probably 
contain a few individuals old enough to have 
partly rooted molars. Granting this supposition, 
a reasonable arrangement would be to place 
rufocanus with lemminus in Aschizomys . 
In this case Craseomys Miller, 1900 will be- 
come a synonym of Aschizomys Miller 1898. 
Recent authors have regarded Craseomys a syn- 
onym of Clethrionomys; perhaps because they 
considered it unnecessary to retain a subgenus 
which contained but a single species. In placing 
rufocanus, niigatae, andersoni, and a new spe- 
cies from Honshu together with lemminus all 
in Aschizomys, one must decide the proper posi- 
tion of this group. Although Miller preferred to 
call it a genus, he did consider that Aschizomys 
could be placed with Eothenomys and Ante- 
liomys. Eothenomys combines characters of Mi- 
crotus and Clethrionomys; and Aschizomys (in- 
cluding rufocanus and andersoni) bridges the 
gap between Eothenomys and Clethrionomys. 
As pointed out by Ellerman and Morrison- 
Scott (1951: 670), Russian authors place As- 
chizomys as a subgenus of Alticola. The two 
groups are certainly very close; but, if I am 
correct in presuming that very old specimens 
of Aschizomys lemminus will tend to develop 
roots on the molars, then Aschizomys is more 
appropriately placed with Clethrionomys. 
Inasmuch as a few old individuals of rufo- 
canus and niigatae have partly rooted molars, 
it seems best to place Aschizomys as a subgenus 
under Clethrionomys, and the forms known in 
Japan can be separated by the key below. 
