Mr. F. S. Conant on the Ghcetognaihs. 203 



Oviducts als eine Art Saraentasche functioniren, wie der oft 

 in ihr vorgefundene Inhalt lebender Spermatozoen bevveist." 



Grassi finds that the wall of the ovisperm duct is inter- 

 rupted at irregular intervals, and that through these inter- 

 spaces spermatozoa maj pass into the ovary before the ova 

 are mature, sometimes penetrating the unripe ova and causing 

 degenerative changes. These interruptions in the wall exist 

 along the whole length of the duct, and Grrassi thinks that 

 probably the ova make their way through them into the duct, 

 thus passing into the duct anywhere throughout its length. 



I may say, in brief, that while my preparations show 

 nothing of these interruptions in the wall of the duct, the 

 actual course of events points to the correctness of Grassi's 

 supposition that the ova pass through some such interspaces. 

 They do not, however, appear to pass into the " ovisperm 

 duct," but into another entirely new channel. The " ovi- 

 sperm duct " seems to function, at least throughout all but 

 the posterior part of its course, as a receptaculum seminis, 

 and Hertwig's supposition is found true in this I'egard. 



Sagitta hispida, which is found in abundance at Beaufort, 

 offers a favourable opportunity for the study of the passage of 

 the ova into the oviduct in the living object. With this 

 species the egg-laying occurs at about sunrise, as is stated 

 also by Grassi for the Mediterranean form Sagitta hexaptera. 

 Boveri quotes Fol as finding the eggs of Sagitta bipunctata 

 laid at sunset ; and he found the same himself in March, but 

 the process was often much delayed, so that at times he waited 

 till 11 or 12 o'clock in vain. The exact time with S. hispida 

 is found to depend somewhat upon the temperature. While 

 the eggs were generally laid about 5 o'clock in July and 

 August, on a cooler morning than usual the process might be 

 found to occur an hour later ; and by artificial cooling in an 

 ice-chest overnight the laying might be delayed as much as 

 three hours. Such artificially retarded eggs developed regu- 

 larly on the whole, but it was found safer to make the obser- 

 vations at the earlier normal hour. 



To follow the process, then, the eggs are found in the ovary 

 the first thing in the morning — at 5 o'clock in the case that 

 we will select — with the large germinal vesicle still showing, 

 but lying eccentrically toward the mid-line of the animal. By 

 5.16 the germinal vesicles have entirely disappeared. At 5.22 

 each egg has flattened somewhat from before backward, and 

 a small cell is seen clearly at the outer end opposite the point 

 where the germinal vesicle was. The ova of the Chteto- 

 gnaths arise from the germinal epithelium on stalks (see 

 Grassi, Taf. xi. figs. 9 and 10), and the small cell standing 



14* 



