1-18 ERNEST HILL AND L. G. HAYDON. 



written ^Yitll a then limited acquaintance with larva?, for if, 

 from the above table of thirty-two species, we endeavour to 

 pick out certain characteristics forming natural groups cor- 

 responding with Theobald's generic classification the difficulties 

 are even greater. Considering, in the first instance, the 

 antenna3 we see that a multibranched hair is found starting 

 from the intero-anterior aspect of five, its presence or absence 

 not specified in two, and that it is definitely absent from 

 twenty-five. Of the five, two belong to genus Anopheles, 

 three to Myzorhynchus. Even if it were possible to find 

 affinities between these two species of Anopheles and the 

 genus Myzorhynchus, which would justify their inclusion 

 in the latter genus, the difficulty would not be surmounted, 

 because the prominent dendriform external frontal hair 

 ( PI. XXII, fig. d) forms at least an equally distinctive 

 feature, and is found in Anopheles maculij)ennis, but 

 not in lindesajM, and again is absent from natalensis. In 

 six species of the six genera represented in the table, the 

 external frontal hair is dendriform (Pis. XX and XXII d). 

 One of these is an Anopheles, one a Cellia, and four 

 belong to genus Myzorhynchus; but of all the genera that 

 which has the most pronounced characteristics in the imago 

 is Cellia. Again, we find that of four mounted specimens of 

 Cellia squamosa, which we have kept, the external frontal 

 hair is penniform in one, dendriform in three. There can be 

 no doubt of the identity of the former, seeing tliat it differs 

 in no other shade from the latter. The distinction between 

 the two types is barely appreciable with a lower magnification 

 than X 100, and might readily be overlooked. 



The most striking feature which we have seen in any larva 

 is a large dendriform hair found on the first abdominal segment 

 of a species of Myzorhynchus, which we take, despite the 

 difference as to the amount of white in the second hind tarsus 

 of imago, to be Theobald's paludis (PI. XXII, fig. /). 

 This hair certainly engendered expectation of its being of 

 some specific value, but we are forced to regard it as merely 

 an individual peculiarity. We have only seen it in three 



