BY R. J. TILLYARD. 417 



in size, yet the larva of the latter is remarkably large, both 

 as compared with that of the former and with its own imago. 

 In both species, the larvae are large for the size of the imago. 

 For instance, Diphicodes huematodes is a larger insect than 

 C. pi/gmcea ; the larva is, however, smaller. In the case of 

 6'. montana, it is hard to resist the conclusion that the imago 

 must, at one time, have been a much larger insect. The 

 reduction in size of the larvae has not kept pace with that of 

 the imagines, because the actual caenogenesis was confined to 

 the vinf/s and abdomen of the imagines only, (note the 

 Ei/rorduliaji character of head, thorax and legs still remain- 

 ing). This may also have been partly due to the existence 

 of the two broods, which I have already explained. If the 

 species were preserved by the complete sacrifice of one of two 

 broods in the larval stage, there did not exist the further 

 necessity for change that affected the imagines. But a certain 

 amount of reduction of larval size must necessarily accom- 

 pany the imaginal reduction ; and this is still proceeding, 

 without affectingr the Eiirnrdulian character of the larvae. 



iv. .1 romparifton of the two known Species of the Genus. — 

 This has already been discussed under headings C, D and E 

 of iii., with a view to shewing that Cordulephya is specialised 

 by reduction. The same argument is, of course, valid for its 

 inclusion in the Eucordulina. It would, however, be greatly 

 strengthened if we could point to one or more species that form 

 intermediate links between Cordulephya and the main body 

 of the Eucordulina. It would be too much to expect a com- 

 plete series of forms linking the two together, but we might 

 expect to find some Australian Corduline, as yet untouched 

 by the main stream of invaders, which would give us an idea 

 of what the ancestor of Cordulephya was like. Such a form, 

 I venture to assert, may be seen in H esperocordulia berthoudi 

 Tillyard*, a rare species found in South-Western Australia. 



* " Further Notes on some rare Australian Corduliince " These Proceed- 

 ingss 191 1, p. 376. 



