MIGEATOET FISHES OE ASIA. 203 



direct, wliicli is said by some naturalists to be ejected ]per anum 

 after it has been used. Native fishermen assert that the midday- 

 is the best time to net tanks, as the fishes are nearer the surface. 



In order to prove the correctness or the reverse of Indian fresh- 

 water fishes having two normal modes of respiration, I instituted 

 a number of experiments, which I do not propose detailing in full 

 (see ' Proc. Zool. Soc' I. c), but some of which I must advert to 

 for the purpose of illustrating my views. 



Before, however, I commence this, I think the following propo- 

 sition will be admitted — that if fish aerate their respiratory organs 

 from air contained in the water in which they reside, there cannot 

 normally he any necessity for their rising to the surface to take in 

 atmospheric air direct. 



Of course, under certain abnormal conditions, all species rise to 

 the surface, as I have already pointed out ; but the following ex- 

 periment will show that the freshwater fishes of India are divisible 

 into two distinct classes as regards respiration. 



I took a glass globe which I nearly filled with fresh water and 

 into which I put a Barbel (Barbus) and a Walking fish (Ophioce- 

 phalus): the first used its gills freely, but did not come to the sur- 

 face. The second moved its gills sluggishly, but every now and 

 then ascended to the surface ; and each time it did so, it appeared 

 to blow out a bubble of air. These were normal conditions of 

 respiration as invariably observed in these examples of the two 

 classes under ordinary circumstances. 



I then took a piece of net, which, by means of a ring of split 

 rattan, could be pushed into the globe of water, thus dividing the 

 interior as by a diaphragm. This J inserted about 2 inches below 

 the level of the water, so that it became impossible for the 

 contained fishes to rise to the surface and obtain atmospheric 

 air. 



The result in all instances was similar. The fishes which moved 

 their gills rapidly and did not normally rise to the surface remained 

 unaffected ; those which normally rose to the surface became 

 more and more sluggish, and died in a longer or shorter period, 

 apparently in a ratio corresponding with whether they had been 

 quiescent or excited, the quiescent ones living the longest. 



Conclusive as the above results appeared to be, I tried another 

 series of experiments. I stitched a bandage round the gill-openings 

 oS both these classes of fishes. Those which normally used their 

 gills rapidly and did not rise to the surface soon died j those 



18* 



