22 PROCEEDINGS OF THE NATIONAL MUSEUM VOL. 97 
Measurements of holotype (in millimeters).—Head and body, 245; 
tail, 233; hind foot, 62; ear, 28; condylobasal length, 52.9; zygomatic 
breadth, 33.9; length of nasals, 18.5; supraorbital constriction, 18.5; 
postorbital width, 18.5; width of braincase, 22.6; alveolar length of 
molar row, 9.1. 
Coloration of the paratopotypes (4 males, 5 females).—All individuals 
of the series agree rather closely with holotype; general pattern of the 
pale agouti type (3a) with comparatively weak contrast in shoulder 
regions. Three specimens have a narrow, faintly outlined median 
dorsal band on lower half of back. Tails reddish on upper surface 
with the black hairs from beneath tending to show through; patterns 
on undersides of tails, bicolor, tricolor, and in one specimen, quadri- 
color. In latter, disposition of colors from outer border to midline 
is reddish, black, ochraceous, mixed black and ochraceous; individual 
hairs of middle portion of tail 10-banded, the buffy band at base 
succeeded by 4 bands of black, each of which alternates with each of 
4 bands which become successively redder toward the finely pointed 
black tip. Underparts uniformly reddish. 
Rio San Pepro (6 males, 4 females): Agree generally, with the type 
series but average darker due to the greater tendency toward forming 
a dark median dorsal band. In each of two individuals there is a small 
narrow patch of white on the chest. 
Remarks.—The type locality of the “hoffmanni’’-like norosiensis is 
but a few kilometers from those of splendidus and variabilis, both on 
the opposite side of the Rio Magdalena. Nevertheless, norosiensis is 
sharply distinguished from those other two races. Apparently the 
lower Magdalena is a complete barrier to intergradation between the 
squirrels of the opposing banks. To show gradation from norosiensis 
to variabilis it is necessary to pursue a path along the Central Andes 
to the upper Magdalena (chrysurus), thence across the river and re- 
turning northward along the Cordillera Oriental to meet the range of 
variabilis through that of zuliae. 
Of four specimens taken by H. M. Curran in May 1916, at ‘‘Puerto 
Estrella,” Rio Magdalena, at the mouth of the Cesar, three have been 
considered representative of typical splendidus. The fourth differs 
widely and agrees in every respect with norosiensis. It must be con- 
cluded that this example was taken on the opposite shore of the river. 
In this connection it may be mentioned here that among other indi- 
cations of the importance of the lower Magdalena as a zoogeographic 
barrier, the author found the marmoset, Marikina leucopus, abundant 
from the Rio San Pedro to the very edge of the Magdalena, but com- 
pletely absent on the opposite shore. 
Specimens examined.—Twenty-one. Norosi, 10 (U.S.N.M.); Rio 
San Pedro, above Norosi, 10 (U.S.N.M.); ‘“‘Puerto Estrella,” 1 
(U.S.N.M.). 
