266 DR. J. B. GATENBY ON THE GERM-CELLS AND 



6. The Structure of the Amphiblastula Larva. 



The amphiblastula larvae of Grantia compressa develop under the 

 flagellated collar-epithelium, and are partially or wholly surrounded by a 

 capsule of nutrient squamous cells. In PI. 23. figs. 28 and 32 the capsule 

 is marked LGrC ; at first I believed that the capsule was really part of the 

 larva — that is embryonic tissue — but Professor Dendy kindly pointed out its 

 true nature ; this partly investing membrane is formed of maternal cells, 

 and when the larva breaks forth from the sponge, the capsule is left behind. 

 The subject of (his nutrient capsule will be dealt with below. 



When the larva, after having become free, is examined alive, it will be 

 noted that the histocytes or flagellated cells, while being clear at their outer 

 extremities, contain in their inner ends abutting against the blastocoel, a 

 large number of brownish-yellow yolk-spheres : in other species of sponge- 

 krvse, the colour of the yolk may be much brighter. In the fresh, the 

 archeocytes or posterior granular cells are found to contain fewer yolk- 

 spheres, or at least they are not grouped in such a wajr as to form a 

 conspicuous coloured band as in the flagellated cells. 



Now if the sponge is fixed in Champy's chrome-osmium mixture, and 

 treated as elsewhere described (4) by the Champy-Kull method, in the 

 embryos true yolk is greenish-brown, mitochondria are red, and nuclear 

 substance is bluish-green. Moreover, by the Flemming-without-acetic 

 acid and Iron Hsematoxylin method, the yolk is greenish-brown, the 

 mitochondria are black. By Altmann's method, yolk is greenish-brown and 

 mitochondria red. 



As is well known, if one fixes the sponge in highly coagulative fluids such 

 as alcohol, corrosive acetic or Bouin's fluid, the posterior granular cells 

 become more chromophile and granular on staining than the flagellated 

 hystocytes. 



Nevertheless, I cannot say that this more marked granulation in the 

 posterior cells is traceable to a greater content in the latter of either 

 mitochondrial or yolk spheres. On the contrary, most of the granular cells 

 contain fewer grauules of yolk or mitochondria than the histocytes. This 

 can be seen on examinino- PI. 23. fies. 28 and 32. 



Some amphiblastula? embryos prepared in alcohol and stained in carmine 

 showed the posterior or lower cells much denser and chromophile than the 

 flagellated cells, while yolk and mitochondria were absent or, at the best, 

 hardly observable. I trace the difference in qualitjr of the granular and 

 flagellated cells not to their contained granules, but to certain differences in 

 the ground-cytoplasm, which in the granular cells is coarser and more 

 chromophile on fixation in alcohol-acetic-corrosive liquids. 



The cytoplasm of the granular cells is extremely oxyphile, nuclei are 

 amphophile and contain a true plasmosome : in iodine the granular cells go 



