150 GLIRES. 
the white-bellied colliwi phase. The rufous-bellied specimens lead us 
directly into 
2. The rigidus type. Upper parts usually marbled with fulvous and black, the 
hairs being glossy brown or black with a median band of yellowish brown. 
Lower parts bright rufous (rigidus 18), or varied with rufous and white, either 
symmetrically (intermedius 22) or asymmetrically (nicoyanus**). In many 
specimens, as observed by Mr. Allen, there is a more or less marked tendency 
to the development of a pale grizzled lateral stripe placed high on the flanks 
and confining the dark dorsal area to a broad median band. From these we 
have every gradation into the rufous phase of 
3. The dorsalis type. Upper parts with the median dorsal area black, flanks and 
lower parts white, fulvous, or rufous. In many individuals the hairs of the 
black dorsal area will be found to have concealed rings of white or fulvous; 
so that we have a further transition into 
4, The colliwi type. Upper parts much as in the rigédus form, but usually less 
rufous ; lower parts white (colliai", adolphei*) or pale fulvous (variegatoides®, 
griseocaudatus 1, pyladei , oculatus*!). Usually the dark coloration of the 
back is extended over the shoulders, hips, and limbs; but in others it is con- 
fined to the middle of the back, as in the last phase, while the rufous-bellied 
phases lead us directly back into the rigidus form. 
5. The melanius type. Upper and lower parts uniform glossy dark brown, often 
with irregular patches of paler brown. ‘This is a very peculiar phase, and I 
have not yet seen any direct intergradation with the other varieties; but 
uniformity of coloration is so abnormal among the Squirrels that it seems 
impossible to doubt that Mr. Allen is right in regarding it as a melanism. 
The colour is exactly similar to that of the back of the white-bellied dorsalis 
varieties, in which also the paler apparently faded patches are sometimes seen. 
It appears to be much rarer than any of the other forms described. 
With regard to the synonymy, I may observe that I have been able to examine the 
types of all the “species” here united, excepting that of S. hypopyrrhus, which, how- 
ever, has been well described by Wagler and Wagner ; it appears to be a dark variety 
without the usual wash of white on the tail. That of S. nigrescens only differs in 
having the fur of the lower parts ringed like that of the back; it agrees well with 
specimens in the Paris and Berlin Museums. S. colliwi and S. adolphei are synonyms 
for the white-bellied form above described, from which I am also unable to distinguish 
S. richardsoni of Gray (not of Bachman), on which both S. boothiw and 8. fuscovarie- 
gatus were founded. S. variegatoides, S. griseocaudatus, S. pyladei, and S. oculatus 
are all intermediate links with more or less fulvous lower parts, leading to the phase 
with a red or red-and-white belly, on which the names S. intermedius, S. regidus, and 
Macroxus nicoyanus have been bestowed. 
