478 Dr. T. A. Chapman's Notes on 



lu common with the whole group, Trocldlium has 

 the appendages fixed to first and second abdominal seg- 

 ments, and also almost as firmly to third ; but this connec- 

 tion dissolves on dehiscence and the suture 1-2 abdominal 

 also opens. Segments 3, 4, 5 and 6 (abdominal) carry 

 two (an anterior and posterior) rows of spines directed 

 backwards, V also does so in the $ ; in the % like 8 and 9 

 it only has the forward row. These spines are stronger on 

 each segment than on the preceding one. 



Specific characters are probably confined to the relative 

 strengths and extensions of these rows, to the development 

 of the nose-horn (beak) and to the exact details of the 

 spines on anal segments, possibly also to the forms of the 

 spines. Even so, it seems difficult to be sure in any case 

 how far the variations observed are specific and how far 

 individual. 



Sphcciformc, scol'uTformc and cvliciforme differ slightly 

 in build, the others taper regularly (or nearly so) ; these 

 taper much more in the last few segments. 



Andrcmforniis differs in the spines in all the rows being 

 finer, smaller and paler than in the other species examined, 

 just as the pupa itself is so, though probably the paleness 

 of the spines affords a good part of the impression that 

 the pupa as a whole is pale. 



The nose-horn (beak) is in accord with the general weak 

 structure in being very small and short, a mere fine nodule 

 on the face, that one almost overlooks — it is nevertheless 

 sharp. (In many species this is large, sharp and 

 formidable.) 



The above memoranda seemed necessary in order to 

 define its generic and subgeneric position, and if possible 

 to suggest points by which it might be distinguished from 

 most nearly allied species. The individual variations 

 are however so considerable and the distinction so slight, 

 that I doubt whether any species of TrocJiilium could be 

 named with certainty from its pupa. Having only one 

 example of andrcniformis, its range of variation cannot 

 be estimated, and it is therefore so far fortunate that its 

 general more delicate structure gives some points of 

 difference with the other species examined. The following 

 description is largely generic, or even family, rather than 

 specific. The arrangement of the terminal crown of spines 

 differs somewhat in most species. 



