147 



ginning of the spawning season , and kept confined in a latticed box 

 in the water off the First Commission's wharf at Wood's Holl during 

 the fall, winter and spring. Development slows up in full, but does not 

 come to a standstill even in winter. Sections of eggs taken December 

 1st. and January 1st. do not exhibit marked changes, but surface 

 views of the embryos, and the relative size of the black eye-spots (which 

 serve as a good measuring gauge) , prove that considerable progress is 

 made by the growing embryo, during this month. This is also seen in 

 the gradual absorption of food yolk. 



The period of incubation at Woods Holl is 10—11 months or 

 about 300 days 2 . 



In an egg with the segmentation of the superficial yolk comple- 

 ted, 34 cells or yolk pyramids are present. The constrictions of the 

 yolk are not simply superficial , but cleavage planes often reach down 

 half way to the centre of the egg. The nucleus with its rayed proto- 

 plasm lies near the centre of the convex face of each segment, but it 

 is still separated from the surface of the egg by a considerable layer of 

 yolk. The entire amount of protoplasm is thus distributed among the 

 yolk segments, none of it remaining in the undivided yolk below. Nu- 

 clear division is now radial, but at a later period after a large num- 

 ber of small segments have been formed, some of the nuclei begin to 

 divide tangentially, that is to delaminate (Fig. 5), and their pro- 

 ducts pass into the yolk below the superficial cells. Yolk cells are 



Fig. 5. 



Fig 5. Part of section of egg in a stage later than Fig. 3, in which the surface 

 consists of a pavement of very small, polygonal segments. This shows a single nucleus 

 which is delaminating. This egg contains several yolk cells. I = yolk sphere. 

 YP = yolk segment. Yolk somewhat diagramatic. 



formed at a similar stage in Alpheus by migration and possibly also 

 by delamination. 



2 An earlier statement to the effect that incubation lasted for a much shorter 

 period was based upon an erroneous assumption (v. Johns Hopkins University Cir- 

 culars. No. 80). 



