Oral Gestation in the Gaff-Topsail Catfish, Felichthys Felis. 43 



for 2 years and 6 months, weighed 3.5 grams. A larva with body- 

 walls which had just completely closed over the yolk, the seam of 

 closure only being visible, after being in 4 per cent formalin for 4 

 years and 3 months, measured 93 mm. and weighed 9 grams. This 

 fishlet was in all respects normal and was just in the stage when the 

 young begin to lead a free life. To make this point clear, the reader 

 is asked to contrast the young fish in figures 11 and 12 of plate iv with 

 the eggs with comparatively early embryos shown in figures 8, plate in, 

 and 9, plate iv. 



This increase in length of over 45 mm. (a doubling) and in weight 

 of 5.5 grams (almost a trebling) can not be accounted for by yolk 

 alone the young feed in the mouth of the parent. At some future 

 time it will be shown that the larvss reared in running sea-water 

 grew faster and when fed by hand were less voracious than those 

 kept in a diluted sea-water filtered many times, the latter feeding 

 ravenously when given bits of oyster. The conclusion drawn in 1907, 

 which has remained unchanged, is that the young feed while in the 

 mouth of the father by filtering out of the respired sea-water, by means 

 of their closely set gill-rakers, minute Crustacea to satisfy their hunger. 

 In this way only can their great increase in size and weight be 

 accounted for. 



Von Ihering (1888) is the only investigator who has given any data 

 whatever on the growth of the young while in process of brooding. 

 Of Arius commersonii he records that the eggs in early development 

 stages average 2.5 grams, while at a time when the larvae are 60 mm. 

 long they weigh 4.3 grams. The increase in weight, he argues, is due 

 to the intake of nourishment by the embryo. 



LENGTH OF THE PERIOD OF GESTATION. 



It is impossible to state this either from observation or direct experi- 

 ment. It being impracticable, at the time this research was carried on 

 at the Beaufort Laboratory, to isolate the fish in pairs, direct observa- 

 tion of the period of incubation was not feasible. All efforts to effect 

 artificial fertilization proved abortive, nor was I ever able to get eggs 

 in early segmentation stages. Thus it is impossible to fix the time of 

 fertilization, which would give the time of the beginning of incubation. 

 Further, it has been impossible to carry any one set of early eggs 

 through to time of hatching (i. e., bursting the egg-shell) and of the 

 complete closing of the body-walls over the diminished yolk-sac as 

 shown in figure 12, plate iv, at or about which time the young are set 

 free from the paternal mouth. 



Confronted by these impossibilities, the best that can be done is 

 to make the closest approximation possible from the data at hand. 

 Now it has been shown in the section on the breeding season that the 

 eggs are "laid" under ordinary conditions of weather (i. e., temper- 



