82 



this, but it should be named. Blackburn mentions an example 

 of undulata, Don., illustrating the above tendency (ante, 

 1897, p. 32). 



It is thus quite possible that varicollis, Cart., is a variety 

 of yarelli, C. and G., though this appears to be a good example 

 of geographical variation, in which every specimen has a 

 more or less constant pattern that varies from the type found 

 elsewhere. This remark is also true of rufipes, Macl., a red- 

 legged North Queensland form of octospilota, C. and G. This 

 kind of variation, or species in the making, obviously intensi- 

 fies the difficulty of sharp definition in the tabulation in a 

 genus in which colour is so prominent a feature. In a few 

 species the colour of the underside is variable. This will be 

 treated especially under sex variation, but in the much- 

 described punctiventris, Saund. ( = binotata, Saund. 

 = guttata, Blackb., etc.) the underside is either yellow 

 or coppery, while in species having more or less yellow on 

 the under-surface the amount of such colouration is very 

 variable. S. decipiens, Westw., is also variable in pattern. 



(3) Structural variation (including sculpture and cloth- 

 ing). — The chief variations here will be noted under sexual 

 variation; otherwise such variations are rare, and the best 

 diagnosis of species can be made on structural characters. 

 There are, however, two obvious cases that deserve mention. 



(a) Width, or expansion of the sides of prothorax. 



(b) Structure of the apices of the elytra. 



In a few cases there are some considerable variations 

 under fa) — e.g., klugi, C. and G., often shows wide differ- 

 ences in the form of the prothorax (a fact to which Mr. G. S. 

 Bryant called my attention when collecting in Sydney). This 

 is true also of the species parryi, Hope, which in a long series 

 I cannot distinguish from parvicollis, Saund. (b) The apices 

 of the elytra form, in general, one of the best characters for 

 the separation of allied species, and some authors believe 

 this to be a constant character in the same species. There is, 

 however, sufficient variation here to give cause for trouble 

 (vide Blackburn, ante, 1900, p. 48, on yarelli). The examina- 

 tion of long series of bicincta, Boisd. ; octospilota, C. and G. ; 

 cupripollis, Saund., and others, will show enough variation 

 to make exact description difficult, the spines sometimes being 

 subobsolete ; where the type has distinct spines. Also many 

 of the larger species, whose apex is more or less truncate with 

 a short external spine (e.g., thoracica, Hope; variabilis, Don.) 

 show a variation towards the simply-rounded apex. 



(4) Sexual variation. — The most marked sexual distinc- 

 tion lies in the form of the last abdominal segment, the male 

 having this segment more or less excised (e.g., reichei, C. and 



