Classification of the Mongooses. 523 
behind that point; (upper lip shallow, 
cleft ; digits 5—5). 
6. Digits 2 to 5 always united by a web 
which projects beyond the margin of 
the plantar pad. 
ce. Either a space or pm' intervening 
between the upper canine and pm*. Mungos *. 
c’. No space between upper canine and 
pm", pm’ suppressed .........++¢ Helogale. 
b'. Digits 2 to 5 separated down to plantar 
pad, owing to suppression of the webs 
present in other genera ..........+: Atilar. 
a’. Dentition more generalized, rather crush- 
ing than sectorial, upper carnassial 
(pm*) set forwards so that its posterior 
angle is well in advance of the root of 
the malar arch, the whole or practically 
the whole of m!' being also in advance 
of that point, the line of the cheek-teeth 
forming a much more even curve at the Ones ee 
junction of pm* and m’. ; 
d. Webs deeper as compared with length 
of digits; ears moderate or small, 
with no pocket behind the antitragal 
ridge; skull elongated, lower; inner 
portions of upper m’, m? thick and 
strong ; last lower molar with median 
external cusp or flat-crowned. 
e. Legs short, semiplantigrade, fore 
paws broad, with very long claws; 
at most the heel of the hind foot 
hairy, ears small and rounded (no 
groove on upper lip). 
f. Snout short as in Mungos, infra- 
narial portion of rhinarium shal- 
low; anal sack complex........ Ariela. 
f'. Snout Jong, infranarial portion of 
rhinarium deep; anal sack simple. Crossarchus. 
e’. Legs long, digitigrade, fore paws 
narrow, claws shorter, metatarsus 
covered with hair almost to the 
. plantar pad, ears longer. 
g. Upper lip grooved, posterior cheek- 
teeth cuspidate; mesopterygoid 
fossa set further back. 
h. Pollex and hallux retained .... Ichnewmia. 
h'. Pollex and hallux suppressed .. Bdeogale. 
g'. Upper lip ungrooved, posterior 
cheek-teeth flat-crowned ; meso- 
pterygoid fossa set more forwards. Rhynchogale, 
d', Webs very shallow as compared with 
length of digits, especially between 
digits 3 and 4, 4 and 5; ears very 
large for the group, with a small 
ee ee es Re es ee 
* I suspect this genus will prove to be divisible into three or more 
genera when better known. 
