188 Mr. H. Scott 07i 



lateral process is bare, and has at its apex a sole-like area, 

 the surface of which appears different to the general 

 surface of the organ ; in a balsam-mount under a high 

 power this area has an exceedingly finely spotted appear- 

 ance, due to the presence of an extremely dense felt-work of 

 very minute, short, curved setae (possibly not in any way 

 homologous with ordinary setae or spines, but formed by 

 some kind of breaking up of the general chitinous surface). 

 A sole-like area, apparently closely similar, is present in 

 the lateral processes of the larva of the allied Kelloggina 

 barnardi. 



Brandnal filaments (PI. IX. fig. 2) arranged in tufts : 

 in the full-grown lar\a there are ten in each tuft, this being 

 apparently the largest number yet recorded in the family ; 

 almost always three are directed forwards, three backwards, 

 and the remaining four ventrally and outwards. There is 

 among the material one much younger larva, only 325 mm. 

 long ; it agrees closely with the larger larvae, so that I haye 

 very little doubt it belongs to the same species, but the 

 number of branchial filaments in each tuft is much smaller, 

 being at most six. and sometimes apparently only five. It 

 therefore appears that the number of filaments in a tuft 

 increases with the growth of the larva. A bi'anchial tuft of 

 a full-grown larva, stained with para-carmine and mounted in 

 balsam, shows distinctly that the ten filaments arise in five 

 pairs, the bases of the two members of each pair being con- 

 tiguous. It is therefore possible that the increase in number 

 of the filaments occurs by a process of fission during the 

 development of the larva. 



Additional support is given to the above conclusion by 

 two Liponeura-lnvvee preserved in the Cambridge Museum. 

 They were collected in the Rhone Valley in July 1897 by 

 W. Bateson, F.R.S., and are probably Li/joneura cinerascens, 

 Loew, as they agree with Bezzi's description [op. cit. 1913, 

 p. 77) of that species in all points (except, possibly, that 

 of colour). One of these larvae is only about half tlie 

 size of the other ; and, while the laiger one has in each tuft 

 the number of brancliial filaments characteristic of the 

 genus, namely seven, the half-grown larva has only four 

 filaments in each tnft. The number of filaments in a tuft 

 is a very important systematic character ; evidently, then, 

 one must beware of being led to false conclusions by the 

 examination only of partly-grown specimens. 



Terminal (i. e., Sid'th) Segment (PI. IX. figs. 1,2). — The 

 anterior portion, bearing the lateral processes, is almost 

 identical in form with segments 2-5 ; its branchial filaments 



