608 ARTHUR WILLEY. 



tified with the primitive ventral flexure of the embryo of my 

 Peripatus (cf. Fig. 2, a and b), with the difference that in the 

 Myriapods the flexure goes much deeper and is much more 

 pronounced than in Peripatus (cf. Metschnikoff [9], and 

 Korschelt and Heider [7]).^ 



Thus in the Myriapods (as in Peripatus) the ventral flexure 

 of the embryo may be said to be a product of invagination, but 

 the orifice of invagination does not narrow down at all, remain- 

 ing freely open on all sides. 



In Lepisma, as shown by the observations of Heymons, we 

 have a primary ventral flexure of the embryo, strictly com- 

 parable with the ventral flexure of the Myriapod embryo, and 

 the invagination which produces the ventral flexure is accom- 

 panied by a narrowing of the orifice of invagination, thus 

 giving rise to an amniotic cavity opening by an amniotic pore. 



The preceding remarks may be resumed in tabular form as 

 follows : 



P. NOViE-BRITANNI^. MyRIAPODA. LePISMA. 



Shallow ventral flexure." Deep ventral flexure. Amniotic cavity and 



Amniotic pore. 



Thus although the amnion itself first appears within the 

 group of the Hexapoda, it does not owe its origin to purely 

 mechanical causes, as has been so often supposed, but can be 

 traced back, through the link supplied by Lepisma, to the 

 primitive ventral flexure of ancestral forms. 



I think I have now said enough to explain my theory of the 

 embryonic membranes of insects, and I give it for what it is 

 worth. I have endeavoured to trace the serosa back to a 

 primitive trophoblast, and the amniotic cavity back to a primi- 

 tive ventral flexure. 



New Museums, Cambridge ; 



September 22nd, 1898. 



1 In the embryos of P. novse-britannise the primary ven\ral flexure 

 does not involve the trophic vesicle; whereas in Myriapods it does involve 

 the vitellus. 



' That is to say, shallow in its primary relations to the trophic vesicle. 



