158 The American Naturalist. 
support this division, and it breaks down utterly when we come to take 
into consideration the osteology of the various species. ao 
The skeleton of Geotrygon has not been examined by me; but I am 
of the opinion that it will not militate against the classification sug- 
gested below, judging as I do from its external anatomy. ee 
My studies of the osteology of the group convince me that our 
“United States pigeons naturally make a very good suborder, containing | 
the family Columbide. Now, if we take the characters presented on 
the part of the skeleton of such a species as Lctopistes migratorius, We 
find that they are essentially repeated by all the other genera save 
Starncenas. When we come to osteologically compare Starncenas wè 
find that it differs very materially and in a number of points, as in 
the general pattern of its sternum, the number and arrangement of its 
vertebree and ribs, some of its cranial characters, and in the characters” 
of its pelvic limbs. os 
From osteological premises, then, our family Columbide divides 
naturally into two subfamilies: the Columbine, containing the genera 
Columba, Ectopistes, Engyptila, Zenaidura, Zenaida, peli 
Columbigallina, Scardafella, and Geotrygon ; and the subfamily Star- 
nænadinæ, containing the genus Starncenas. | ae 
In another connection it is my intention to present these. osteolo 
cal characters of the Columbide in detail.—R. W. SHUFELDT, + 
sonian Institution, January 22d, 1891. fz 
Description of Two New Species of Rodents from Mexico. 
—While recently classifying and arranging the collections of naa 
belonging to the Natural History Section of the Comision 
Exploradora of Mexico, I found two species apparently nê 
characters I now give : a 
S , Sp. nov——Apparently quite similar gil 
Merriam’s recently described S. cryptospilotus, which 1 T i 
from the description. Above, head and body fawn color, with : 
indications of spots. Individual hairs with extreme bases lack, 
lowed by a narrow ring of straw-yellow, subterminally oadl Lag, 
with walnut-brown (which color occupies more than co 
covered by the two preceding colors), and tipped with cre 
Something like one per cent. of the hairs have the walnut-DrO™ 
placed by black ; but these are so relatively few in number 
~ sensibly affect the general tone, Color gradually shading 
_ sides, where it meets, in a sharp line, the white of under parts: 
_ of decidedly lighter fawn color, 3 mm. in width, comm 
