78 



ferencc to common air, and nearly in the same pro- 

 portion *. Dr Crawford observes, that if a portion 

 of atmospheric air be exposed in an inverted jar to 

 water which has had its air separated by boiling, the 

 purer part will be attracted by the water, and the 

 noxious portion will remain distinct in the vessel f. 

 The experiments of Mr Dalton, however, do not 

 allow us to suppose that water thus completely ana- 

 lyzes the air, for, according to him, if a quantity of 

 water, deprived of air, be agitated with atmospheric 

 air, the water will attract portions of each of its consti- 

 tuent parts the same as if they were presented to it 

 separately in their proper density J. From the opera- 

 tion of this distinct attractive power of water for the 

 two gases of which the atmosphere is composed, it 

 follows, that the oxygen gas, contained in water, will 

 bear a considerably larger proportion to the nitro- 

 gen gas than exists in the air of our atmosphere ; 

 and that as this oxygen gas is consumed by the re- 

 spiration of aquatic animals, more will be attracted 

 to supply its place, and to maintain the due respira- 

 bility of the air. 



64. The changes induced on the air by fishes are 

 effected by the branchia or gills, which in form and 

 structure vary greatly, according to the mode of 

 life of the fish. In eels, and those fishes which live 

 in impure water, the gills are supported by boney 

 arches, and are very large in proportion to the size 



* Thomson's Chemistry, vol. iv. p. 490. 



f On Animal Heat, p. 145. 



\ Manchester Memoirs, vol. i. new series. 



