BY R. J. TILLYARD. 801 



f] -scribing the colour pattern on p.793, since they occupy the 

 darkisli spots tliere mentioned, and shown in Plate Ixxix., figs. 4, 

 5. Each pinaculum carries from two to four dolichasters, and a 

 single small seta upon a raised papilla. In specially good pre- 

 parations, I have seen the blind ending of a small trachea upon 

 the integument close to the pinaculum; this would appear to 

 indicate the original position of the meso- or metathoracic 

 spiracle, which has evidently become suppressed. 



The Sjnracles. — In all three larval instars, the number of 

 functional spiracles is eighteen, viz.. a pair upon the prothorax, 

 and a pair upon each of the first eight abdominal segments. They 

 are all very small, simple openings, the rim of which is slightly 

 strengthened and darkened. As far as I can see, the}^ are quite 

 devoid of armature or ornamentation of any sort. A single 

 trachea, whose calibre is approximately equal to that of the 

 spiracle itself, passes inwards from it to the main tracheal trunk 

 on each side. The prothoracic spiracles are the largest pair: 

 next in size are those of the eighth segment, with those of the 

 seventh, sixth, etc., in descending order of magnitude up to the 

 first. Text-tig. 8, .<?/;., shows one of the spiracles of the seventh 

 segment in situ, with its corresponding pinaculum (all but one 

 of the surrounding dolichasters removed). Owing to the great 

 number of these latter, it is always difficult to locate the spiracles 

 in this larva. 



The Anal Papilla (Text-fig.9).— The last two segments of the 

 abdomen, viz., the ninth and tenth, are much narrower than the 

 rest, and are more or less retractile within the eighth. They do 

 not carry dolichasters, but only very delicate, slender macro- 

 trichia of reduced size. The tenth segment forms the anal 

 papilla or sucker, by means of which the larva is enabled to move 

 rapidly backwards, or to hold on tight when attacking its prey. 

 This segment is furnished with a pair of dorso-lateral curved 

 processes, probably representing the original cerci, and a pair of 

 ventral processes, somewhat cornute in shape, and curving out- 

 wards to meet the incurving tips of the dorsal processes. These 

 latter are armed with two series of closely set, short, sharp setaj, 

 one set directed outwards and the other inwards. I have found 



