THE HABITS OF THE SALMON. 43 



(perhaps mineral), from which the main supply of 

 the water in the loch is derived. It is not easy 

 to account for this in any other way. 



If the respective temperatures of the Bush, 

 Bann, Erne, and Bundrowse rivers were taken on 

 a cold day in spring, and we found that the water 

 in the Bush and Bundrowse rivers was warmer than 

 that in the Bann and the Erne, then we could 

 account for the two last-named being later rivers — 

 the Bann being the latest river in Ireland. 



Even without that information, I think my state- 

 ments strongly support the assumption that the pre- 

 sence of early salmon and their movements in fresh 

 water is generally dependent upon the state of the 

 temperature. It is as much beyond the power of 

 man to alter temperature as to alter the character 

 of rivers, and therefore any attempt to convert a 

 late river into an early one, or to induce salmon to 

 run into fresh water earlier than is their usual habit 

 would, in my opinion, be an utter failure. 



In his book on Salmon Problems, Mr. Willis 



