1919] N. Annandale : Boynhay Streams Fauna. 115 



taculatus, by a slight swelling. From the more normal species of 

 Discognathus the genus is further separated by the structure of 

 the pectoral fins, which are greatly expanded, entirely horizontal 

 in position and with their rays broad, flat and sparingly branched. 

 The chest is also flat, devoid of scales, highly muscular and, with 

 the pectoral fins, capable of adhering tightly to flat or uneven 

 surfaces The other generic differences are trivial or inconstant : 

 Discognafhus may have two or four barbels or none, Psilorhynchiis 

 two or none ; the two lips are joined together by a fold in Discog- 

 nalhus, separated by a groove in Psilorhynchus. In the Indian 

 species there is no difficulty in separating the genera, but both in 

 India and Africa forms occur that are intermediate in one way or 

 another. Boulenger ' in describing the Nilotic species of Discog- 

 nafhus says : 



" Aberrant Cyprininae adapted for living in torrents and 

 mountain rivers. The mental disk by which they are enabled 

 to fasten themselves to stones varies greatly in its degree of 

 development, and is so reduced in some specimens of the 

 species described below as D. quadrimaculalus as to be almost 

 indistinguishable. Such specimens might be referred to the 

 genus Crossochilus, Van Hasselt^ were it not for the more 

 reduced gill-openings, which do not extend to the lower 

 surface of the head, the isthmus being much broader than in 

 the related genera." 



The resemblance between D. quadrimaculatus and Psilorhyn- 

 chus seems to me to be very close indeed, both in the .structure 

 of the mouth and as regards general facies There is no resemblance, 

 however, in the structure of the pectoral fins, and the lips are 

 continuous in the African fish. 



If we have in this African species a close aoproximation 

 to Psilorhynchus, we have among Indian forms a still closer 

 resemblance m the modification of the pectoral fins and chest in 

 one or possibly two species of Discognathus. This is the case in 

 D. nasutus. Here again, however, we have a difference as well as 

 a resemblance, for the pectoral fins in D. nasutus differ from those 

 of all species of Psilorhynchus in that all the rays except the 

 outermost ones are branched, though sparingly so, whereas in 

 Psilorhynchus several of the outer rays are simple or nearly so. 



Resemblances between the mouth-parts in these fish are perhaps 

 more important genetically than those between the fins, both 

 because they are more exact and because we find similar modifi- 

 cations in the fins of other unrelated genera (such as Homaloptera 

 and even some Siluridae) that also live in mountain streams. The 

 latter resemblances are, therefore, without doubt convergent. It 

 will be as well, however, to inquire a little further, before coming 

 to any conclusion, what are the normal modifications in the 



' The Fishes of the Nile. p. 180(19071. See also the same author's Frex/;- 

 Wafcr Fishes of Africa. I, p. .^43 ^ IQ09). 



