528 KNOWER. [Vol. XVI. 



thick ectoderm and, in this case, form an especially large mass 

 under the posterior end of the band. 



Turning to PI. XXXI, Fig. 37, we find several important 

 changes. It is a section of the stage in PI, XXXI, Fig. 27, 

 before the appearance of cephalic lobes. The embryo now 

 forms an elongated band bent over the posterior pole of the ^g^. 

 The mesoderm has followed the growing posterior end and has 

 become arranged in a thinner layer. Its anterior cells appear 

 to have retained their primary position, as in the preceding 

 stage, but the greater mass of mesoderm has been carried back 

 with the elongating ectoderm, leaving only a single layer 

 beneath the middle of the embryo. This growth of the meso- 

 derm is, I believe, accomplished independently of the ectoderm, 

 by a multiplication and rearrangement of its own cells. There 

 is still a sharp division between the two layers. The growth 

 seems to be more active at the posterior end, while the middle 

 region appears to be pulled out, as it were, the anterior end 

 remaining stationary. The size of the yolk-cells still precludes 

 a later origin of entoderm from these. There is no trace of 

 entoderm up to the time of the segmentation of the germ-band. 

 The ectoderm just in front of the anterior limit of the meso- 

 derm has grown further forward than in the preceding section 

 (PI. XXXI, Fig. 36). This anterior extension of the ectoderm 

 will continue in later stages, and give rise to the cephalic lobes. 



The effect of the backward elongation of the germ-band on 

 the amnion, whose cells are now apparently multiplying but 

 seldom, is well shown in the section before us. Posteriorly it 

 still retains to a slight degree the character of the ectoderm, 

 though much thinned out. Anteriorly the amnion has been 

 stretched out by the pull from behind into a very thin mem- 

 brane of flattened cells. I have found but few dividing nuclei 

 in the later stages of the amnion, the membrane appearing to 

 be stretched rather than to actively grow. This is beautifully 

 seen in surface preparations, where the amniotic cells, now 

 much larger than those of the more rapidly multiplying ecto- 

 derm, stand out in bold relief, lying closer together posteriorly. 



The oldest stage of the germ-band just before segmentation 

 is dissected-off from an Qgg like that in PI. XXXI, Fig. 29, 



