EARLIER STAGES OF CATACLYSTA LEMNATA. 39 



being the larger. They are ranged in some degree in transverse 

 and other lines, but broadly their arrangement is too irregular 

 to be described. They present various circular lacunae, one of 

 which, a little above and behind the spiracle and two to three 

 times its diameter, is conspicuous. The spiracles are very in- 

 conspicuous. 



The head and mouth parts are not very intelligible without 

 elaborate drawings. The second (?) antennal joint is very long, 

 and the jaws have a remarkable form. Where they face each 

 other they have not merely a toothed margin, but have a circular 

 face, hollowed centrally, and with teeth round more than half 

 the margin, so that they are like scoops with toothed edges. 

 This structure seems to be attained by the ordinary five teeth 

 being placed in more crowded disposition than usual, and then 

 continued and supplemented beyond the end with largest (not 

 smallest) teeth by a little row of four smaller accessory teeth, 

 which I do not remember to have observed before in other larvse 

 (of course I have examined really very few). The circle in which 

 the teeth lie is, however, continued right round to the attached 

 margin of the jaw, suggesting that the two jaws form a more 

 than usually closed pocket, possibly to retain sap, &c., in sub- 

 aqueous mastication. 



The larva moults four times. The difficulty of following any 

 individual larva and noting its moults seemed to be so great that 

 I did not attempt it, but I preserved first instar larvae and full- 

 grown ones, as well as a considerable number in intermediate 

 stages, of which those in second instar were the only ones of 

 whose stage I was certain. But, arranging all my specimens by 

 the sizes of their heads, I find that between the second and the 

 last instars two, and only two, sizes occur, and these five sizes 

 range themselves in regular order. This method is of course 

 nevertheless not so sure to be correct as the actual observation 

 of each moult in one individual. 



The pupa is 8-10 mm. in length and 3 mm. in breadth, 

 varying a little in size, and especially the females are the larger 

 and wider ; but there is much latitude in size in both sexes. 

 The apparent size varies also a good deal, owing to the amount 

 of collapse possible in the two free abdominal segments (five 

 and six). In a dead pupa these close up very much by 

 drying, and in an empty pupa-skin they are often completely 

 telescoped. 



The widest part of the pupa is at the end of the wings, some 

 5*5 mm. from the front. Seen laterally, the pupa is of somewhat 

 ordinary oval form, well rounded at each end, and a little flat in 

 front ; but, viewed dorsally (or ventrally), it tapers to either end, 

 and this looks more remarkable forwards, conically rather than 

 by an oval outline. At any rate, it differs from our average idea 

 of a pupa in this direction. 



