Extrated from the American Naturalist, February, 1891. 
Notes on the Classification of the Pigeons.-Quite recently 
the writer has very thoroughly compared the characters presented on 
the part of the skeletons of specimens of nearly all the genera of the 
United States Columbidae. There appears to be a difference of opinion 
as to how these birds should be classified. Coues, in his "Key" 
(second edition), states it as his opinion that " the order Columbae 
may be separated into three groups or suborders : Didi, Pterocletes, 
and Peristerae,-the first two certainly, the last probably, of a single 
family. The Peristerae alone are American. These he divides in the 
following way: 
Subfamilies. 
Suborder. 
Family. 
( i. Columbinae. 
Peristerje. 
Columbidae. 
< 2. Zenaidinae. 
Starnoenadinae. 
In the Columbinae he includes the genera Columba and Ectopistes ; 
in the Zenaidinae, the genera Engyptila, Zenaidura, Zenaida, Melo- 
pelia, Columbigallina, Scardafella, and Geotrygon ; and finally, in the 
Starnoenadinae, the genus Starncenas. 
The American Ornithologists' Union, in its official check-list, pre- 
sents the order Columbae to contain the family Columbidae, and cre- 
ates no subfamilies for the genera just named above. 
Mr. Ridgway, in his " Manual," adopts the same scheme of classi- 
fication. 
Coues primarily bases his division of the Columbidae into sub- 
families upon the following characters : 
Tarsi scutellate, feathered Columbinae. 
Tarsi scutellate, nakedZenaidinae. 
Tarsi reticulate, nakedStarnoenadinae. 
The remaining characters, in so far as we have any knowledge of 
them at present, except in the case of the Starnoenadinae, do not go to 158 
The American Naturalist. 
[February, 
support this division, and it breaks down utterly when we come to take 
into consideration the osteology of the various species. 
The skeleton of Geotrygon has not been examined by mej 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. 
My studies of the osteology of the group convince me that our 
United States pigeons naturally make a very good suborder, containing 
the family Columbidae. Now, if we take the characters presented on 
the part of the skeleton of such a species as Ectopistes migratorius, we 
find that they are essentially repeated by all the other genera save 
Starnoenas. When we come to osteologically compare Starncenas we 
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 
vertebrae and ribs, some of its cranial characters, and in the characters 
of its pelvic limbs. 
From osteological premises, then, our family Columbidae divides 
naturally into two subfamilies: the Columbinae, containing the genera 
Columba, Ectopistes, Engyptila, Zenaidura, Zenaida, Melopelia, 
Columbigallina, Scardafella, and Geotrygon ; and the subfamily Star- 
noenadinae, containing the genus Starncenas. 
In another connection it is my intention to present these osteologi- 
cal characters of the Columbidae in detail.-R. W. Shufeldt, Smith- 
sonian Institution, January 22d, 1891.