THE SYSTEM OF ARTIFICIAL FLIES. 17 



make it quite impossible for Trout to discriminate 

 minutely between the various unnatural imitations 

 of natural flies, whether in form or tint ; (5) and 

 render it doubly important that the imitation in- 

 sect should be as characteristic and " fly-like" as 

 possible in shape, lest the fish should fail to per- 

 ceive the resemblance altogether. 



6. General shape, general colour, and size, are all 

 that can be distinguished by the fish. These are 

 the points, therefore, to be kept in view in the con 

 struction of artificial Trout-flies. 



The next step is to reduce these propositions to 

 practice ; and the argument naturally takes some- 

 what the following form : If when presented to 

 them in the only manner in which we can present 

 them, nice varieties of imitations, and shades of 

 shape and colour cannot be distinguished by Trout, 

 the great mass of flies now used are obviously 

 unnecessary, and where either the colour or outline 

 is confused, are mischievous. It would be better 

 on every ground to select two or three of the 

 most favourite and distinctive families of flies, and 

 imitate them only ; not in their varieties, or even 

 species, but, as it were, in their types and using 

 those colours only which represent the prevailing 



B 2 



