Some of the flies are of Scotch patterns; some of Irish; some of English. The flies are types of texture and of shape. They are not standards of size. Size ranges between that of the large and that of the small fly shown on the last in the series of pictures. Why salmon flies differ so much in size I will explain immediately. Meanwhile it is to be noted that only seventy-six flies are depicted on the plates. Where are the other hundreds ? They are scattered all over the British Isles. It may be thought that our pictures of salmon flies are too few. Instances of successful fishing with many a fly that our plates exclude will be recalled. In making our selection Mr. Malloch and I have not been unmindful of that consideration. Quite well we know that many a fisherman will find that his favourite fly has been rejected. The fact is that a fly with which a salmon might possibly be caught could be made of any conceivable mixture of fur and feathers. In essaying a presentment of trout flies it is possible to be explicit and exhaustive. The flies are imitations, approximately exact in colour, size, and shape, of insects on which the trout are known to feed. Such are the flies the images of which appear in Trout Fishing. Exactitude of similar kind is not possible in a work on Salmon Fishing. Whilst the science of trout flies is natural, that of salmon flies is empiric. Nature, outside the mind of the trout, tells us exactly what trout flies should be like, and exactly how many they should be; but Nature, outside the mind of the salmon, is on that subject dumb. In fishing for trout we see insects and the trout rising at them, and thus, by a simple process of thought, know what artificial trout flies should be made of; but in relation to the salmon we have no such guidance. We do not see any fly at which the salmon habitually rises. " He likes something to snap at,'" we say ; "it is probable that the more luscious the offering seems the more it will be approved ; let us, therefore, blend colours appetisingly."" That argument, or something like it, is the genesis of every fly that has been cast upon a salmon river. The argument indicates why salmon flies are much more various than trout flies. As Nature shows us nothing to copy, the scope for speculation and invention is practically unlimited. Fishermen have been for ages, and still are, devising new combinations of tinsel and feathers and fur. The whole of this volume would not give space enough for pictures of all the flies with which salmon have been tempted.

How, then, it may be asked, is the selection now presented to be justified ?

It is to be justified by the fact that really, after all, we do know a good deal about the salmon's taste as regards flies. How the knowledge has been gleaned may best be illustrated by an incident. Two men, one of whom was Mr. Watson Lyall, who told me the happenings, were fishing on a river from the same boat. By lunch-time one rod had caught seven salmon, and Mr. Lyall had caught none. All the fish had been taken on a Dusty Miller. As it was the only Dusty Miller on board, "I'll give it to you for the afternoon," said the successful fisherman, " and try some other fly myself." With that Dusty Miller in the afternoon Mr. Lyall caught six salmon, and his friend, using other flies, had none. The hook was broken at the bend by a seventh fish. It was nearly nightfall then; but had the Dusty Miller held out, other salmon would have come in. They kept rising at the barbless lure, which for a little while longer was cast to them in wonderment at their determined preference.

This incident affords very clear proof that salmon do not rise at random. One fly is not at all times as good as another. The fish do undoubtedly have preferences. It was on the basis of vigilant and comprehensive observation of these preferences that the task of selecting flies for presentation in this volume proceeded. The selection has been made with extreme care; considering the guidance of which I have had the privilege, I may say, also, that it was made from the richest and most minute knowledge. Although there are other flies which are successful now and then, those which are figured in this book, it is confidently believed, are types of the best flies known within the United Kingdom. For each of them there has been, and will be again, many an occasion as propitious as that which befell the Dusty Miller.

Even so, we are still far from an approximately complete philosophy of experience as regards salmon flies. Few though they be in relation to the whole of the flies in use, absolutely they are many. " Seventy-six of them ! " I can imagine some one exclaiming. " It would be a week's work to give each a half-hour's trial! How am I to know which four or five to trust when I have only a day to spare ? "

That question cannot be answered precisely. All the knowledge we possess is general. Large flies are appropriate in spring and autumn; in summer smaller flies are best. Flies with silver bodies begin to be attractive as the colder weather comes. The lower the stream, the more sombre should be the lure. When the river is tinged by flood water the gaudiest flies are the most likely to be successful.

These are general rules, dependent, of course, upon normal conditions. Usually the rivers are fairly full early in the year; usually they are small in summer; usually they begin to rise as autumn is approaching. Sometimes, however, the normal order is broken, and then we find that the rules are not without exceptions. Here again an incident may be helpful.

Just as Jock, in the estimate of Serjeant Mulvaney, was " a deceivin' fighter," the Tay is a deceivin' river. Almost every part looks as if it would be excellent for trout; but that is judging from experience on ordinary streams. The Tay is not ordinary. It is an enormous burn rather than being a river. Contrast it with the Thames. That is a temperate stream, sedate, not often in a terrifying flood, and hardly ever extremely low; but sometimes the Tay is so small that in certain broad parts you could wade across, and at other times it rises twelve feet in as many hours. The Thames is within stable banks; those of the Tay are frequently broken, and every flood makes changes in some of its channels. Thus a stranger in Tayside must often find himself at a loss. He may say to himself, "This side-stream, much less rough than the main flow, will hold some good trout," and then find, after delicate casting of his flies, that evidently it holds no trout at all. It would never occur to him that, far from having been for ages as he finds it, the side-stream was not there until a few weeks before, when, being in violent mood, the Tay cleft a few new bypaths for itself. On the other hand, some of the side-streams, long-established ones, yield excellent sport. Not far from Aberfeldy, for example,The idiosyncrasies of the Tay, however, are so many that I must do no more than suggest their nature. Were I to dwell upon them in detail, I should have to defer a pleasant task, which is to say how we fared on the afternoon of March 1, 1905, when the trout fishing season opened in Scotland.