94 MR. K. ANDERSEN ON BATS [May 16, 



7-oiijxi, is rather consiclei'able (as is also the variation in the ex- 

 ternal dimensions of this Bat) ; but among 18 skulls of the typical 

 foi'm of roiixi, from localities so many and so distant inter se 

 as to represent practically the whole area covered by this form, 

 I do not find any so small as the largest among 11 skulls of 

 horneensis (and h. sjxcdix) ; in so far there is no difficulty in 

 discriminating them. The tooth-rows, too, in rovxi, are longer. 

 As to the small S. Chinese race of rouxi (described below), the 

 skull has the same length as the largest of horneensis, but the 

 brain-case is decidedly broader, the zygomatic and maxillar width 

 greater. 



Dentition (19 skulls). Pg, most often, quite external (12 skulls); 

 not rarely half in row, or | in row (6 skulls) ; in one aged 

 individual (teeth much woi-n) Pg is wanting, on both sides of the 

 mandible, and the alveoli have disappeared. Oingula of p., and p^, 

 most often, in contact or separated by a very narrow, sometimes 

 almost hairline, interspace (13 skulls); in the remaining (6) 

 individuals, distinctly separated, but the width of the interspace 

 is not always quite the same on both sides of the mandible. 



The upper canine and p'^ are, with rare excejations. distinctly 

 separated, p" completely in the tooth-row (17 skulls, out of 19), 

 as in all the foregoing species. The size of p" and, therefore, the 

 width of the interspace between c and p* vary, however, to a 

 certain extent ; but in no instance is the width of the interspace 

 as broad as (p" as well developed as) in simjjlex : this is a thing of 

 the past. As to the remaining two skulls (Ceylon, Nepal), the 

 interspace is very narrow, 2^' half external. This is the first time 

 we have to note instances of p" not being comjiletely in the 

 tooth-row. 



As a general conclusion : — (1) In Rh. i'oiixi Pg has arrived 

 so far on its way towards disappearance as to be, generally, 

 external ; but still, not rarely, the individual variation falls back 

 to a foi-mer stage : p, partly in the tooth-row ; and in some aged 

 individuals the dentition (p^ disappeared) jwints forioards to sub- 

 sequent stages in the series of evolution : Rh. ferrum-equimmi 

 (p., rather often lost) and Rh. acrotis (p^ always lost). (2) As to 

 p" in rouxi, it is generally in the row, rarely half external ; this 

 latter, again, ])oixits forivards towards subsequent stages : thomasi, 

 ferriim-eqidnum, and acrotis (p" always external, or lost). 



Distribution. From S. China through the Himalayas to the 

 Indian Peninsula and Ceylon. 



Technical name. As Rh. rouxi has for many years been com- 

 pletely confused with Rh. affinis, some i-emai-ks are necessary to 

 prove that the name rouxi belongs to the species hei'e under 

 consideration. The type locality of Rh. rouxi is "Calcutta"*; 

 the types (in the Leiden Museum) were collected by the French 

 naturalist, M. Roux. There is in the Tomes Collection (British 

 Museum) a skin also collected by Roux. The essential points 



* Temminck, loc. infra cit., p. 30 e ; . Jentiiik, ' Catalogue systematique ties 

 M'ammifereSj' Mus. d'hist. nat. Pays-Bas, xii. (1888) p. 161 (under Bh. affinis). 



