30 Papers from the Department of Marine Biology. 



tail bent upward at a sharp angle, but without anywhere referring 

 to it, though in the adjacent text he has been discussing oral gestation 

 in Silurus militaris, which he figures without barbels. However, 

 some seven years after Bloch's death, Schneider, in editing the Sys- 

 tema Ichthyologice (1801) of his predecessor, says in a footnote to 

 Silurus hagre (p. 378): "I have observed the eggs of this fish about 

 to be hatched in the mouth." The larval form as figured by Bloch 

 on plate 365 has 6 barbels and lacks the filament to the dorsal fin. 

 Whether these are errors chargeable to the artist can not, of course, be 

 determined. Possibly this Httle fish is the young of some other silu- 

 roid described by Bloch in the context. At any rate it may be noted 

 here that Bloch is the first to describe oral gestation in siluroid fishes. 

 The whole matter is very obscure, but the data are given here that the 

 reader may draw his own conclusions. It is interesting, however, 

 to note that, immediately after bursting the tough eggshell, the httle 

 gaff-topsail has the same high- waving tail as that figured by Bloch. 

 (See fig. 10, plate iv.) 



THE BREEDING SEASON. 



The limits for egg-laying by the gaff-topsail at Beaufort can not 

 be set very narrowly, but an attempt will be made to indicate when 

 males with eggs may be taken. This research was initiated the first 

 week in August 1906, but only females long spent were taken. The 

 real search for the fish was begun in 1907, when the first female 

 was captured on June 4. Her ovary had in it no eggs over 10 mm. in 

 diameter, but did show a lot of recently evacuated follicles. On June 

 6, other females were brought in having ova as large as 20 mm. 

 Plainly, the breeding season was not over. 



As explained before, it was found impracticable to search for the 

 breeding males in 1907 until June 22, and the eggs taken on this day 

 had on them young in various stages, but the majoritj'' had the whole 

 yolk covered with the blood vascular system. 



The search for early stages of the eggs began in 1908 on June 3. 

 On this day a female with empty follicles was taken, as were others 

 two days later. Notwithstanding daily seinings, no eggs were taken 

 until June 11, and constant inquiries brought no news of the finding 

 of a single egg by any of the fishermen before this date. Our failures, 

 as explained elsewhere, were due to the freshening of Newport River 

 by heavy rains. The youngest of these eggs had the yolk ch-culation 

 covering quite half of the egg and were at least 10 days, probably 

 2 weeks old. 



In 1909, the first eggs were taken May 27. These had on them 

 forming embryos with a huge open blastopore behind. They were 

 probably from 5 to 7 days old. However, a number of females were 

 taken having enormous belUes due to the greatly swollen undischarged 



