1882.] AFRICAN MUN GOOSES. til 



to any localitv, tliougli there is a certain tendency for all, or at least 

 most of tlie specimens from one district to be alike in colour ; in fact, 

 if this had not been so, I should not have felt justified in regarding 

 the various forms even as distinct varieties. Again, H. aibicauda 

 varies most peculiarly in the amount of white ])resent on the 

 tail, some s))ecimens having that member nearly entirely wliite and 

 others wholly black, individuals of each type being moreover found 

 in the heart of the district mostly inhabited by those of tlie other. 

 Another remarkable point in connection with variation in colour is 

 the fact tliat in certahi species there seem to be two forms, one with 

 anuulated and the other with unannulated fur. Thus ordinary 

 specimens of H. galera^ Gr., have annulated fur, while others from 

 various localities, among which are the types of //. j^hito, Tennn., 

 have their longer hairs entirely unannulated. In the same way 

 //. mutgUjella, lliipp., seems to be the unannulated form of H. graci- 

 lis, H.iodopriimnus, Heugl., being intermediate, and H.granti, Gr., 

 similarly that of //. hudius. It is of course possible that these two 

 forms represent a seasonal change, though the few dated specimens 

 that I have seen do not, on the whole, lend mucii support to this 

 view, especially when we remember how comparatively little seasonal 

 change of temperature must occur in the region with which we now 

 have to deal. 



In the second place, attention should ije drawn to the apparent 

 frequency with which specimen;- of this group seem to lose the 

 minute first toes on both fore and hind feet. Tlius boih "Cgnictis" 

 melanura and " Gcderella " ochracea ' owe their generic names to the 

 fact of their halluces having been accidentally lost ; and Bdeogale 

 nigripes, Puch., has probably been referred to that genus for the 

 game reason '. In the work already referred to'', Temminck has 

 drawn attention to this same point, and has given several instances 

 which had fallen under his own observation. Tiie minute claw of 

 the first toe is probably very easily knocked off during the skinning 

 and stutting of specimens ; and the digit iiself, without the claw, is so 

 small and inconspicuous that it would not be observable without the 

 closest scrutiny. On the whole it would seem that though some 

 few cases may occur in which the claw has either never been 

 developed or has been knocked off daring life, yet as a rule, so far 

 as I have seen, the loss has probably occurred after death, there being 

 generally distinct traces of the former presence of the claw, thus 

 showing tliat it could not have been lost during life, for then the 

 skin would presumably have healed over the place. 



Thirdly, it is worthy of note that while the dimensions of the 

 teeth are very constant, their number is by no means so : thus, one 

 of our specimens of Crossarchus zehra (p. 89) has an additional 

 true molar above on each side behind the two normal ones, the extra 

 ones not being in any way crowded or rudimentary, but fully 



' I have softened auc! opened the skin ol' the hind foot of tlie type of this 

 species, and found, as I expoclcd, a broken metatarsus, about one tenth of 

 an inch in length, and of the usual thickness, adiiering to the entocuneiform bone. 



8 See below, p. 77. ^ Op. cii. p. 107. 



