98 



BULLETIN OF THE BUREAU OF FISHERIES. 



fastened, containing an apei'ture large e.iougb to admit the head of a fish. The body 

 of the animal was loosely bandaged, the cloth being tacked to the vertical piece of 

 wood in order to restrain movement (fig. 2). A piece of sheet rubber was then per- 

 forated and drawn over the head in such a way that it fitted snugly around the l)ody, 

 just behind the gill covers. The mai-gin of the rubber being tacked to the wooden 

 frame, a nearly water-tight partition resulted, which served to bound ofi' the head 

 from the trunk region. The frame, bearing the fish, was now immersed in a rectan- 

 gular jar of water, the head being uppermost, except in two specified cases. Fresh 

 (or salt) water was supplied to the jar through a rubber tube, while another tube 

 carried salt (or fresh) water into the mouth, allowing it to pass out through the gills. 

 A second bandage held the head in position, while the pectoral fins were securely 

 bound to staples. 



periment lo determii 



, playefl by the gills 



The following six experiments were made with the carp. The fishes were 

 weighed before and after the treatment, the same precautions being taken as have 

 already been described for earlier experiments. 

 Experiment 113. 



New York, May, 1905." One carp (fed 2-i hours previously), weight Sti.l 

 grams. Salt water (1.014) through gills, fresh on body. After 5 hours alive, 

 though feeble; weight 80. i) grams. The fish thus lost 6 per cent. 

 E.V periment IH.. 



One carp (fed 24 hours previously), weight 63.6 grams. Fresh water 

 through gills, salt on body. After 5i hours fish well; weight 63.8 grams. 

 Weight thus practically unchanged. 



« Date and place the same lor all of this series. 



