GIGANTOCYPRIS MOLLERI 237 



The sphincter between end sac and duct is so similar in sections to that of Doloria 

 that there can be no doubt that the constitution of the structure in the two forms is the 

 same. In this case my tentative description of the valve in Doloria is certainly incorrect. 

 In that form I suggested (1931, p. 478) that the number of cells in the sphincter valve 

 was three and by comparing it with the antennal gland of Chirocephalus I implied that 

 these three cells, and hence their contained myofibrils, were connected directly to the 

 ectoderm. I have various series of thin sections of Gigantocypris, in one of which the 

 antenna was so orientated that the section is accurately transverse to the valve. From 

 these series there is no doubt whatever that the valve consists of four cells and that these 

 cells are not in any way connected to the ectoderm. Their myofibril apparatus is con- 

 tained entirely within themselves. Further, in the sections of embryos the sphincter 

 shows as a group of four cells (Fig. ija) and these form the junction between the 

 developing sac and duct but are not otherwise connected to any cells. 



This constitution of the sphincter is remarkable for in all previous descriptions of 

 similar structures, when the number of cells constituting the sphincter has been stated, 

 it has always been three (Vejdovsky, 1901). In one case, that of a Cyprid, I followed the 

 embryology and showed that three ectodermal cells become intercalated between end 

 sac and duct to form the sphincter (Cannon, 1925). Now in Gigantocypris we have 

 a case of an apparently homologous structure in which the number of cells is definitely 

 four, and there is no evidence to suggest that they have been connected to the ectoderm. 



The real difficulty, however, is not to justify the homology of a sphincter of three 

 cells with one of four cells, for, in the evolution of the larger Gigantocypris there is no 

 reason why an additional cell should not have been added to the original sphincter. The 

 difficulty is to explain how, if they are strictly homologous, the method of functioning 

 of the one can have evolved into that of the other. 



In Chirocephalus or Cypris the fibrils form a triangle whose three corners represent 

 the attachment of the fibrils to the exoskeleton. The sides of the triangle are curved 

 inwards. Obviously contraction of these fibrils will straighten out the sides of the 

 triangle and this will open the sphincter. 



In Gigantocypris the primordium of the sphincter consists of four pear-shaped cells 

 with their narrow ends surrounding a tubular space (Fig. 17a). In the adult the same 

 arrangement of cells is to be seen but the fibrils can now be distinguished. They are 

 arranged in two distinct sets — one set against the wall of the end sac whose fibrils run 

 on the whole at right angles to those of the other set which is against the commencement 

 of the duct (Fig. ijd). On the inner surface of the cells, that is close against the lumen 

 of the duct, fibrils from one set run down to meet the other (Fig. 17c). In this way the 

 fibrils which together surround the duct and form the so-called sphincter are arranged 

 along four sides of a tetrahedron. Contraction of these fibrils alone would presumably 

 close the sphincter but the whole system taken together is too complicated to allow of 

 any useful suggestion as to its method of functioning. At the same time it is clear that 

 it cannot work in the simple manner of the triangular sphincter of other forms, and 

 further it is difficult to see how an intrinsic tetrahedral system of fibrils can have evolved 



