314 mr. g. e. dobson on the [May 19, 



ceased to respire, and lay at the bottom of the vessel. To show that 

 their deaths were not due to a deleterious condition of the water, it 

 was carefully drawn off (a small portion sufficient to cover the fish 

 being allowed to remain) and fresh substituted, without, however, 

 permitting the fish to reach the surface. The fresh water seemed 

 to produce no change in their condition. At 5 p.m. the largest 

 specimen lay on its side ; at 7 p.m. all were dead. Both the speci- 

 mens in water with free access to the surface were alive at this hour ; 

 but one died at 10 p.m., probably from wounds it had received ; the 

 other was found dead next morning*. 



Exp. 3. — A specimen of SaccobrancJius singio was placed in the 

 same jar, having the netted diaphragm similarly fixed 1 inch beneath 

 the surface of the water. It made but few attempts to reach the 

 surface, and died in four hours. Another specimen placed in the 

 jar afterwards swam about in a very lively manner, and made several 

 attempts to reach the surface of the water ; it died in two hours. 



The difference in the time required to cause suffocation was pro- 

 bably due to the greater exertions made by the second specimen, 

 which exhausted the supply of air faster. 



Other specimens kept for observation in my aquarium with free 

 access to the air were observed, when approaching the surface of the 

 water, to discharge a considerable amount of air from the gill-openings, 

 immediately afterwards placing the muzzle above the water. 



Exp. 4. — A specimen of Amphipnous cuchia \ was placed in the 

 glass jar, which was more than half filled with fresh water. The 

 netted diaphragm was omitted, in order to observe its mode of respi- 

 ration. The habits of the animal while in the water were noted 

 during eight days, with the following results : — It was found that the 

 fish either lay in a semitorpid state at the bottom or remained sus- 

 pended in the water, keeping its head resting against the side of the 

 jar about 1 inch or less below the surface. The used-up air was 

 expelled from the gill-apertures, at first slowly, one bubble succeeding 

 another till four or five were discharged ; the remaining air was then 

 expelled by a single effort, the fish immediately sinking to the bottom 

 of the jar, most probably as the direct result of suddenly increased 

 specific gravity. Usually in about two minutes afterwards the animal 

 raised itself by muscular effort, placed the muzzle at the surface of 

 the water, separated the lips slightly (precisely as a snake does when 



* I do not consider the above-described experiment as conclusively proving 

 this species to be a mixed air- and water-breather ; for although very many 

 specimens were procured by me for the purpose of examining their respiration, 

 the greater number died before I could find an opportunity for making the 

 necessary experiments ; and even those examined were most probably in a 

 sickly condition, as shown by the death of the other two specimens, not confined 

 beneath the surface of the water, a few hours later. When placed in the 

 aquarium these fish invariably kept swimming round and round, having the 

 extremity of the muzzle at the surface of the water, or remained vertically sus- 

 pended with the mouth at the surface. This seems to me to indicate that 

 unmixed air was also required for the purpose of respiration. 



t For a description of the respiratory organs in this fish, see Owen's ' Ana- 

 tomy of Vertebrates,' vol. i. pp. 481 & 487. 



