MUSEUM OF COMPARATIVE ZOOLOGY. 223 



rations the duct in its backward growth is separated by a considerable 

 space from the ectoderm, and I have observed no instance in which it 

 was impossible to distinguish a perfectly sharp line between the funda- 

 ment of the duct and the overlying ectoderm. 



In describing the germ layers in Stage I., I referred to certain histo- 

 logical criteria which might be employed in determining to which germ 

 layer a given group of cells belonged. The most valuable of these is the 

 difference in the size and abundance of the yolk spherules, which even in 

 that early stage served to contrast sharply the mesoderm from the ecto- 

 derm. In later stages, this character is equally pronounced. "When 

 the duct appears, the cells which constitute it are not distinguishable in 

 histological features from those of the adjacent mesoderm, but are very 

 different from those of the neighboring ectoderm. It seems to me ex- 

 tremely improbable that the cells of the fundament of the duct, with 

 their numerous large yolk spherules, should have been recently derived 

 from those of the ectoderm, which are provided with only few spherules of 

 much smaller size. It would be entirely contrary to our conceptions 01 

 the physiological nature of yolk, if in the course of embryonic develop- 

 ment this material was increased instead of diminished in quantity. 



A similar argument seems to me to afford evidence that the duct 

 arises in situ. If the duct had grown freely backward from an anterior 

 proliferation, such growth would in all probability have been associated 

 with the consumption of yolk in the cells of the fundament, and the 

 spherules would be smaller or less numerous than those of the adjacent 

 mesodermal cells. This, however, is not the case. 



I conclude, therefore, that the segmental duct arises throughout its 

 entire length by a proliferation in situ of the somatopleure. Its posterior 

 end, however, grows across to the cloaca free from adjacent tissue. 



Returning to the pronephric pouch, I purpose describing the relations 

 of that organ to the somites. The section represented in Figure 29 

 (Plate IV.) shows graphically these relations. The plane of section in 

 this case was very nearly tangential to the somatopleure at the points 

 where the nephrostomes emerge. In this section it is evident that 

 the three nephrostomes lie precisely under the first, second, and third 

 somites, behind the ganglion nodosum. These correspond to the so- 

 mites which I have numbered II., III., and IV.; so that the proneph- 

 ric pouch remains in the same position as the pronephric thickening 

 of earlier stages. In Figure 21 (Plate III.) the last remnant of the 

 canal connecting the body cavity with the cavity of the protovertebrse 

 is faintly indicated (above the letters coel") in the same transverse 



