MAMMALS OF NORTHERN COLOMBIA—-HERSH KOVITZ 5 
agree that Geoffroy’s description of variabilis fitted the Santa Marta 
squirrels, and he thought that the type specimens ‘‘are much more 
likely to have come from Western Colombia than from the coast 
district about Santa Marta.’’ Accordingly, he advanced saltuensis to 
specific rank with bondae as a subspecies. 
5. Bangs (1900, p. 91) held his ground and insisted that until it 
could be proved by comparison with the types that he was wrong, his 
identification of variabilis and restriction of its type locality must 
stand ‘‘according to all ruling.” 
6. Allen (1904, p. 434) summed up the arguments and felt it ‘safe 
to assume that the real type locality of S. variabilis is the Magdalena 
River of Colombia, at some point remote from the coast, in the region 
inhabited by Ateles hybridus,’”’ the type of which was also collected 
by Plée. Hence, he argued, bondae, having originated in an entirely 
different region, must differ from variabilis. 
7. Allen (1914, p. 593) described S. saltuwensis magdalenae from 
El Banco, Rio Magdalena, at the mouth of the Cesar. This locality is 
within the area that Allen (1904, p. 435) had thought almost certain 
to be the “‘real type locality of S. variabilis.”” He did not return to 
this question, however, in the description of magdalenae. 
8. Thomas (1928, p. 590) showed that magdalenae was identical with 
splendidus Gray (1842). He identified a specimen from the ‘Rio 
Cesar, Santa Marta,” as perfectly representative of splendidus, and 
referred bondae and saltuensis to it as subspecies. 
The present author has examined in detail the types (two specimens) 
of S. variabilis in the Paris Museum. In addition, he has seen some of 
the specimens studied by Bangs and Allen, including the type ot bondae, 
as well as a considerable amount of new material. In the light of this, 
the following opinions may be offered: (1) Bangs correctly identified 
his squirrels as S. variabilis (=granatensis), and Allen was not justi- 
fied in renaming them without recourse to comparison with the types; 
(2) Allen’s objection to the restriction of the type locality to Bonda 
may be sustained and his assumption of its location somewhere up 
the Magdalena distant from the coast appears valid. 
Thus, one might revert to Alston and adopt the specific name 
varvabilis were it not for the fact that granatensis is the older valid 
name for the medium-sized squirrels of Colombia. However, the 
contrastingly colored reddish and black squirrels from La Gloria, Rio 
Magdalena, agree completely with the lectotype of S. variabilis and 
are sufficiently distinguishable from the nearly uniformly orange-colored 
granatensis to warrant recognition of the name variabilis .as a sub- 
species. The still paler form of the Santa Marta region may retain 
the name bondae. Furthermore, the squirrels of the Rio Cesar, from its 
head to its mouth at the Magdalena, show small but consistent 
differences in color from those higher up the Magdalena, at La Gloria, 
