STRAY NOTES ON THE GAME-FISH 63 



so much must be inferred, that life-facts become propor- 

 tionately less capable of proof, more dependent on 

 deduction — or even conjecture- — than is the case with 

 birds or other terrestrial forms of life. The value of 

 work herein depends on correct apportionment of weight 

 to the few facts that come within the scope of vision. 

 • For that reason, while writing" of the observed habits 

 of kelts, I have purposely minimised the evidence in regard 

 to their taking" insect-food ; since an undue importance 

 may possibly be attached thereto. For it is natural 

 to conclude when kelts are observed to gulp down surface- 

 flies, that such is their regular means of subsistence. 

 But if that were the case (instead of being, as I hope 

 to show, merely the exception), the phenomenon would 

 be of unmistakable daily demonstration. For in early 

 spring, the rivers swarm with kelts ; if these large fish 

 had to satisfy voracious appetites in such manner, the 

 waters would boil with their constant rises and plunges — 

 as it actually does, on a minor scale, with those of trout. 

 And that demonstration would be the more conspicuous 

 owing to the greater bulk of kelts as compared with trout ; 

 and secondly, because of the essentially different construc- 

 tion of the two. For whereas river-trout are enabled 

 by their possession of air-glands, inflatable at will, 

 to float buoyantly in mid-water, so that, during a "hatch" 

 of fly, they can lie poised within an inch or two of the 

 surface, ready to snap up the floating food with a minimum 

 of effort and of splashing ; yet the case of the migratory 

 Salmonidcs is entirely different. They; not being so 

 provided with air-glands, are incapable of floating thus, 

 poised at ease in mid-water, or near its surface. The 

 salmon, while in fresh water, must perforce rest on the 

 bed of the stream, his weight supported on some convenient 



