82 EDWARD DRANE CRABB. 



2, 3, 4, 5) and are accompanied by spermatozoa of the same kind 

 (Fig. 9). Since, from smear and sectioned preparations, it ap- 

 pears that the conoid spermatozoa are found more often than the 

 round-headed kind in the upper part of the hermaphrodite duct 

 of different individuals and since the latter have not been found 

 in the acini, it appears that the two types are not different kinds 

 of sperms, but merely represent different ages. Thus it is as- 

 sumed that the round sperm heads found in ova from the her- 

 maphrodite duct were conoid when they entered the ova, probably 

 while the latter were high up in the acini. The heads then began 

 to swell, lost their tails and by continued swelling assumed sphe- 

 rical forms (Figs. 6, 19, 21, 25) and finally disintegrated; the 

 nuclus of one of them normally becoming reorganized to form a 

 vesiculated and later a definite pronucleus (Figs. 26-41). Among 

 sperms which have never entered an ovum I have traced stages 

 of metamorphosis paralleling those which occur within the egg, 

 except that spermatozoa within the sperm receptacle form a pro- 

 nucleus which 'does not possess karyomeres (Fig. 44). The 

 dumb-bell-shaped sperm heads which occur among ova in the her- 

 maphrodite duct are probably fused spermatozoa such as Retzius 

 (Wilson, '25, p. 305) describes for the gastropod Turritclla. 



That these round-head sperms are not the result of faulty fixa- 

 tion is shown by the fact that they are regularly obtained with 

 Zenker, Perenyi and modified Flemming solutions. In view of 

 the fixations used the number and definiteness of these structures 

 in the ova and among the spermatozoa practically precludes the 

 possibility of their being cell inclusions. Neither should these 

 be classed with the " apyrene, oligopyrene, hyper-pyrene " and 

 other similarly abnormal sperms such as Gatenby ('16) de- 

 scribes for L. stagnalis, Ankel ('24) for Bytliinia and others for 

 other forms, for these abnormalities are not sufficiently numerous 

 to play any part in the relative numbers of conoid and round- 

 head individuals. 



B. Ovarian Insemination. 



Although I have not been able to show conclusively that ovarian 

 ova are sometimes inseminated in pond snails, I have seen in- 

 stances, especially in L. colnuiclla, in which ripe spermatozoa ap- 

 pear to have penetrated ovarian ova. Some additional evidence 



