120 S. GOTO. 



and contains a single small nucleolus. It appears in some cases to be 

 more weakly stained than the protoplasm; but this is, I believe, owing 

 to the entire absence of any granules from the nuclear fluid, which, 

 in consequence, appears very clear and transparent. However, 

 the nucleus itself stains equally or often a little more deeply 

 than the protoplasm. As these yolk-cells grow and become larger, the 

 protoplasm becomes moi-e and more coarsely granular and stains less, 

 until at last the whole protoplasm is replaced by the yolk -granules 

 already described. In some cases I have observed cells one half of 

 which contained only yolk-granules, while the other half had still 

 a coarsely gi'anular protoplasm (PI. IV, fig. 7). Lorenz'' had 

 perhaps the young yolk-cells above described before him when he men- 

 tions the occurrence, directly inside the muscular layer of the body, of 

 small cells (0.006 mm.) having a strong affinity for carmin, and the 

 central parts of which remain clear and contain each :i dot-like body. 

 Considering that the yolk-cells are drained off in considei-able numbers 

 during the period of reproductive activity, one would very naturally 

 expect to meet with the phenomena of division among the young 

 yolk-cells, and it appears to me somewhat remarkable that I have never 

 been able to observe in them any division either direct or mitotic. 



Yolk-ducts — In most genera the primary yolk-ducts that pro- 

 ceed from the lobes could not be observed; but in Tristoimun I could 

 often observe them in sections (PI. XXI, fig. 7). They have a very 

 small calibre, and are provided with a thin but distinct, membranous 

 wall. In other species I have often observed yolk-cells passing out 

 from the lobes, but the ducts seemed to close together as soon as the 

 cells had passed along, so that their presence could not be recognised 

 any longer. In Tristomum again, the secondary and tertiary ducts 

 wei-e usually filled with yolk-cells, and could be easily recognised in 



1). Lorons',— /. c. p. 5. 



