THE STAEFISH SOLASTEE ENDECA. 45 



shaped masses not contained within cells, but penetrating everywhere between the 

 layers and compressing all the cavities and organs. 



Probably the hypenchyme-cells which Masterman (i8) describes as being budded 

 oflf from the hypoblast into the archenteron of Crihrella, and completely filling 

 up that cavity for a time, are loaded with yolk-granules. They may indeed be a 

 means of withdrawing a certain amount of yolk from tissues which would be 

 hampered in their activities by its presence in undue amount, the yolk thus segre- 

 gated being in a position where it can readily be digested and absorbed later on. In 

 Solaster there is a slight but apparently constant production of hypenchyme-cells 

 during the period when the archenteron is dividing into its three chambers. A few 

 such cells may usually be found both in the anterior and the posterior coeloms. 



There is, however, another method by which in Solaster a certain small amount 

 of surplus yolk may find its way into the interior of the archenteron. It seems to be 

 the rule in the re-arrangement of cells accompanying blastula formation that some 

 yolk-granules should escape to the free surface. At first I thought the condition was 

 abnormal, but the high proportion of cases in which it occurred convinced me that, 

 at any rate, it is not inconsistent with perfectly normal later development. At first 

 the granules are retained within the membrane of fertilisation which still surrounds 

 the egg, and one may watch them edd5'ing hither and thither under the action of the 

 cilia. As gastrulation proceeds they are swept into the widely open blastopore, at the 

 mouth of which they remain for a considerable time as a kind of plug which is 

 usually devoid of cellular structure, but sometimes contains a few nuclei evidently 

 belonging to cells that, like the granules, have been extruded from the surface. 



(5) The Yolk-Granules in Middle and Later Metamorphosis. — During this period 

 perhaps the most remarkable change in the proportionate size of permanent parts is 

 that undergone by the walls of the enteron. It will be remembered how narrow in 

 early development was the isthmus between the anterior and the posterior coeloms 

 which gave origin to the gut. In the middle and later free-swimming stages the gut 

 still remains small, and is lined by columnar epithelium of no great height. Afterwards, 

 it expands enormously, forming not only the sac-like stomach but also the nine pairs of 

 radial cseca, and much the greater part of this large cavity is lined by high columnar 

 epithelium. On staining deeply with osmic acid, the surprising fact emerges that the 

 whole of this epithelium now contains yolk-granules in profusion. Roughly speaking, 

 just prior to the formation of the mouth, one-half of the total yolk still present may be 

 put down as being contained within the cells of the enteron. A gradual transference 

 of granules from the mesenchyme to the hypoblast cells must have been taking place, 

 as it is quite impossible that the amount of yolk present in the small enteron of the 

 swimming larva could suffice for spreading out over the larger enteron at meta- 

 morphosis. I cannot say where the yolk-granules finally persist longest, not having 

 applied the necessary methods to old enough stages of Solaster. But in Asterias 



