The formation of ice on the bed of the stream was preventing this catastrophe; and so, while seemingly unable to enforce her own law, Nature was really fulfilling her design. The law that water when frozen is expanded, and so floats, was inapplicable to the phenomena under review. All the contents of the watercourse, those which were fluid and those which were stable, were chilled below the freezing-point ; and one of the contents might almost as well as any other have been at the bottom, or in the middle, or on the top.

All nature is but art unknown to thee, All chance direction which thou canst not see, until you look with care. The apparent breach of the law was explained by the consideration that the water, all of it so cold that no part was heavier than another, remained in motion only because the necessity of falling a thousand feet in five miles did not allow it to appear in its true character, which was that of ice. The secret was revealed whenever it had something to rest behind, or to cling to : as when it became frankly ice in the lee of the stones on its bed, or clustered in beads on my cast of flies.

Recollection of the effects of temperatures upon the water will help in a study of the influence of temperatures upon the trout. Whilst approaching this subject with a sense that it is complex, I am not without hope of being able to present considerations which will divest it of much mystery.

Often you hear an angler explaining away an empty basket by saying that the weather on the water was too " muggy " or too " close " ; but you never hear him saying that it was too warm. In his estimation heat in itself is no hindrance to his efforts: it is only the conditions which sometimes accompany heat that are a trouble. On the other hand, he will often tell you, without hesitation, that the weather has been too cold. Cold, he will say, puts down the trout.

The proposition, which is usually absolute, made without reference to times or seasons, is not in accord with experience. This must speedily be realised by all whose wanderings in pursuit of trout extend from the South of England to the Highlands. The climates of these places are not the same. In Cornwall, or in Devonshire, or in Hampshire, a shower of snow in March is so unusual as to be noticeable; in the Highlands, until the end of April, it is as common as a shower of rain, and is not a freak even so late as Whitsuntide. Besides, fishing in the North begins much earlier than in the South. From the Thames to the Test it is not considered sportsmanlike to seek trout until April; but in the North they are fair game a month before that. It is in the Highlands that this problem of temperature is to be looked into most scientifically : it is there the data are most comprehensive.

What, then, do we find in the North ? Do one's experiences early in the season afford sanction for the common belief that the trout are kept down by cold ? They do not. " Snow brew," admittedly, is unfavourable. Anglers do not expect good baskets from a flood which is the result of snows quickly melting in a thaw, and undoubtedly the sport is poor. The explanation, I think, lies mainly in the action of cold upon the earthworms. A warm flood, a flood which comes with spring rain when the country is free from snow, entices the worms to the surface of the soil, and hurries many of them down the hillsides to the streams, to feed the trout; but melting snow chills the earth more than the snow itself, and " snow brew " on the hillsides and on the fields causes the worms to keep to their winter quarters, which are farther down than a spade goes at a stroke. A flood of that kind bringing no food into the streams, the fish are not on the outlook ; and, unless it happens to run up against the very mouth of one of them, the angler's worm is unregarded, as a Mayfly would be in August. When the melted snow has been drained off to the sea things wear a different aspect for the angler. The temperature may be even lower than it was when the " snow-brew " floods were out; but that does not matter. The trout will come at the flies. Even if the temperature is such that your fingers and feet are numb, during the first few weeks of the season, when the weather seems to be free from those thundery and other obscure conditions which are a misfortune to the sportsman later in the year, the fish rise well any day and all day. While the water itself is of normal temperature, the temperature of the air is unimportant. The readiness of the trout to rise is not stopped even by a shower of snow.

Very soon, however, there is cause for reconsideration. On a running water the sport of one day is pretty much like that of the day before, with the differences that it is sometimes arrested by conditions which, for our present purpose, we will assume to have little direct relation with the temperature, and that its quality increases as the fish gain in strength and agility ; but what has come over the lake? Only last week, let us say, this drift by the north shore yielded many trout; but now a rise is rare. What has happened ?

In order to understand the phenomena of sport in lakes, it is desirable that we should first realise that still water differs from running water in an important respect. A stream is of the same temperature all through. It is just as cold, or as warm, on the surface as at the bottom; just as cold, or as warm, at the sides as in the middle. A lake lacks this equality of temperature. Its waters are much less quickly transfused. It is obvious, for example, that if in April there is a sudden freshet from the high lands where snow still lies in drifts and corries, all round the points at which the hill streams enter there will be places where the lake is colder than it is in the middle. There is a still more powerful though less observable cause for inequality of temperatures throughout a lake. We have seen that before ice begins to form on still water the body of the water has to be reduced to freezing-point. Waters that are shallow, therefore, are covered with ice sooner than those which are deep. That is why there is skating in St. James's Park earlier than on the Serpentine. Also it explains why, whilst the Scottish Championship is run for on Lochleven, Lochlomond is almost constantly free from ice. Lochleven is so shallow that it is covered with ice after frost of a few days' duration; Lochlomond is so deep that long before the process of transfusion has been sufficient the "cold snap" has given way. In a lake inequalities of temperature are produced also by the direct action of the sun. When the sun beats down upon the water the deeps are less quickly warmed than the shallows; and the shallows on the south side, having no backing of land to retain the heat, axe less quickly warmed than those of the north, where, besides striking aslant upon the water itself, the sunbeams beat directly on the banks, by which part of their warmth is caught and thrown upon the lake.