FORMATION OF MEMBRANES AND PRIMITIl'E BAND. 243 



In the fly-embryo the existence of an amnion and serosa has 

 been denied. I have already figured (Fig. 2) what I believe to 

 be the amnion and serosa when the embryo is in an early 

 stage of development, but a short time later both apparently 

 disappear. 



At a still earlier period, before the primitive band is 

 developed, ventral and lateral involutions of the blastoderm 

 occur. These have been recently described by Graber [114J 

 as ptychoblastic invaginations ; I have given copies (PI. XIII ) 

 of several of his figures of transverse sections of an embryo, 

 exhibiting a stage of extreme interest, which is also described 

 and figured by Voeltzkow [111]. I am at a loss, however, to 

 understand on what grounds Graber supposes these invagina- 

 tions to be mesoblastic. I shall hereafter show that the 

 mesoblast, as distinct from the parablast, has an altogether 

 different origin. It appears to me that Graber's ventral 

 ptychoblast, which Voeltzkow recognised to be the rudi- 

 mentary condition of the whole primitive band, is the 

 amnion, and Graber's lateral ptychoblasts are folds of the 

 serosa. With regard to his dorsal ptychoblast, I regard it as 

 the hypoblast undergoing invagination and converting the 

 blastula into a gastrula. 



I shall return to the subject of the dorsal ptychoblast 

 of Graber, and in this place I shall deal exclusively with the 

 ventral and lateral folds. 



The ventral fold, which I consider the true amnion, is a 

 thick-walled tube consisting of a single layer of large cylindrical 

 epithelial cells, formed by a longitudinal invagination of the. 

 blastoderm, which is eventually closed by the coalescence of 

 its ventral lips. 



A comparison of this tubular thick-walled amnion with that 

 of the embryo in the Aphides, as described by Will [108J, 

 shows that it closely resembles it. The pre-formation of a thick 

 neural amnion in certain Rodents before the appearance of the 

 primitive streak and neural groove is a precisely parallel 

 phenomenon in Vertebrates. 



It is verv desirable to obtain more observations on the 



