Barker (1651) makes the rod 'of a hasel of one piece, or of two pieces set together in the most convenient manner, light and gentle.' He recommends the use of a single hair next the fly,--'you shall have more rises,' which is true, 'and kill more fish,' which is not so likely. The most delicate striking is required with fine gut, and with a single hair there must be many breakages. For salmon, Barker uses a rod ten feet in the butt, 'that will carry a top of six foot pretty stiffe and strong.' The 'winder,' or reel, Barker illustrates with a totally unintelligible design. His salmon fly 'carries six wings'; perhaps he only means wings composed of six kinds of feathers, but here Franck is a better authority, his flies being sensible and sober in colour. Not many old salmon flies are in existence, nor have I seen more ancient specimens than a few, chiefly of peacocks' feathers, in the fly-leaf of a book at Abbotsford; they were used in Ireland by Sir Walter Scott's eldest son. The controversy as to whether fish can distinguish colours was unknown to our ancestors. I am inclined to believe that, for salmon, size, and perhaps shade, light or dark, with more or less of tinsel, are the only important points. Izaak stumbled on the idea of Mr. Stewart (author of _The Practical Angler_) saying, 'for the generality, three or four flies, neat, and rightly made, and not too big, serve for a trout in most rivers, all the summer.' Our ancestors, though they did not fish with the dry fly, were intent on imitating the insect on the water. As far as my own experience goes, if trout are feeding on duns, one dun will take them as well as another, if it be properly presented. But my friend Mr. Charles Longman tells me that, after failing with two trout, he examined the fly on the water, an olive dun, and found in his book a fly which exactly matched the natural insect in colour. With this he captured his brace.