TKOPHOBLAST AND SEllOSA. 603 



5. We may, however, be quite certain that yolk is uot the 

 most primitive kind of trophic organ for the intra-uterine 

 nutrition of the embryo, for the simple reason that the accumu- 

 lation of yolk, such as occurs in the eggs of the Australian 

 and New Zealand species of Peripatus, is a step towards 

 secondary oviposition on terra firm a, a condition which is 

 actually realised in the case of the Victorian species, P. 

 oviparus Dendy. 



6. The lecithality and deposition of the eggs of insects are 

 both secondary. 



7. With oviparity the trophoblast necessarily ceased to act 

 as an absorbent mucous membrane. 



8. The trophoblast therefore became transformed into the 

 blastoderm and its derivative the serosa, by substitution, this 

 being apparently the prevailing method by which an epithelial 

 transformation or metamorphosis is efl'ected. 



9. The serosal dorsal organ of insects is the product of the 

 ontogenetic involution of the serosa. 



10. The phylogenetic involution of the trophoblast has also 

 to be accounted for, it being understood that the serosa was 

 not directly derived from a primitive trophoblast, but by sub- 

 stitution. Although phyletically related, the trophoblast and 

 serosa are therefore not identical structures. Accordingly we 

 may well expect to find some vestiges of the true primitive 

 trophoblast cropping up here and there. 



11. 1 see in the dorsal organ of the Poduridae and in the 

 indusium of the Locustidse, vestiges of the true trophoblast, 

 botii of these structures {)resenting, in the earlier stages of 

 their development, many of the characters of a mucous mem- 

 brane. This interpretation of facts at least accounts for the 

 otherwise meaningless dorsal organ of the Poduridae (Fig. 5). 



12. The dorsal organs of Crustacean embryos and the primi- 

 tive cumulus of Arachnids must require special explanation, 



of tlie embryo shows that this is a special case, and not one from which 

 phylogenetic conclusions may safely be drawn. An inverted development 

 must necessarily be secondary, as compared with a development whicli is not 

 inverted. 



VOL. 41, PART 4. — NEW SERIES. T T 



