4 
Classification of the Mongooses. 517 
expressed his views regarding the true affinities of the genera, 
or whether the arrangement, in its entirety or in part, was 
merely a matter of convenience for the determination of the 
genera, I am not sure. 
Mivart’s classification was published in the same year as 
that of Thomas (P. Z. 8. 1882, p. 185). He pointed out 
that the genera may be arranged in various ways, 7%. eé., 
according to the number of anal glands, the number of digits, 
the number of teeth, and the presence or absence of the sub- 
nasal groove ; and it is quite clear, I think, that Mivart had 
no preference for one category over another. The use he 
made of the anal glands has already been discussed (P. Z. S. 
p. 366, 1916). With.regard to the other groups, by the 
number of toes Suricata is ranged alongside Bdeogule, by the 
character of the upper lip it falls with Rhznogale and Cros- 
sarchus, by the number of premolar teeth it is associated with 
Helogale and Crossarchus. 
Suricata has been selected here as a test of Mivart’s proposed 
classification, because, in my opinion, the simple structure of 
the ear in that genus shows that it cannot be closely affiliated 
with any other genera of mongooses, all of which have com- 
plicated highly specialized ears; and this conclusion further 
suggests that the suppression of the divisional line of the 
upper lip may be an independently acquired resemblance 
between Suricata and Crossarchus or Rhynchogale. 
From a comparison of the genera, admitted in my paper in 
1916, both mutually and with those of the subfamilies of the 
Viverridee, it may be assumed as a working hypothesis that 
the immediate ancestor of the mongooses possessed the 
following characters :— 
1. The snout was of moderate length, and a naked grooved 
strip of skin ( philtrum) extended from the rhinarium, which 
had a deep infranarial portion, to the edge of the upper lip. 
2. The cheek-teeth, consisting of four premolars and two 
molars above and below on each side, were of a crushing and 
cuspidate rather than of a shearing and piercing type, with the 
upper carnassial (pm*) set well in front of the posterior angle 
of the cheek where the inferior edge of the zygoma rises, 
thus leaving space behind for two well-developed molars, the 
last molar of the mandible being also well developed *. 
* If it be claimed, as it may be claimed, that the ancestral form had 
the specialized carnivorous dentition of the kind seen in Mungos, then 
that genus, setting aside the ear, differs but little from the hypothetical 
progenitor of the group, and the more generalized omnivorous dentition 
of such forms as Ichnewmia and Crossarchus has been secondarily acquired. 
A similar argument applied to the subfamilies of Viverridee will involve 
Ann. & Mag. N. Hist. Ser. 9. Vol. iii. 34 
