Such a torrent consists of three often well-defined parts. First we have the numerous tiny rills which collect together to form what the French physiographers call a receiving basin (_bassin de réception_); then there is the stream proper forming a canal which drains the basin, while below, where the torrent debouches on the low ground, we find that it spreads out fanwise and throws down its load of débris to form a cone (_cône de déjection_). The torrent therefore already imitates a full-grown river, with its threefold division into mountain track, valley track, and plain track. It further illustrates the twofold work of the river, that of erosion and deposition. Observation on an unprotected surface after a heavy rainfall will illustrate another point which is of much interest in connection with the work of rivers. This is that the water has most excavating power, not, as might be supposed, in the collecting basin, but in the valley region, where the slope is still great, where the volume of the water is at its maximum, and where it has acquired a load of débris by means of which it carves out its bed. The excavation of the bed therefore proceeds from below upwards towards the collecting basin. The result is that the slope of the valley floor diminishes as we pass from the upper region to the lower, owing to the levelling effect of erosion. The process of levelling down cannot be carried beyond a certain point, the so-called base level of erosion, which in a lateral stream is determined by the point of junction with the main stream, and in a main stream by the point which marks sea or lake level, for obviously no point in the river valley can be worn down much below its mouth. When the work of a river is completed, the line which marks the profile of its bed should have a gentle and continuous slope downwards to base level. The existence of irregularities, of breaks in the smoothness of the slope, means that the work of excavation has not proceeded far, that the river is young. But it is not necessary to proceed to the laborious drawing of a profile in order to determine the extent to which the process of excavation has been carried. The existence of rapids, of waterfalls, the alternation of swift and slow-flowing reaches are all proofs that it has not been carried far. In short, if a river is navigable, the navigable reach at least is mature; if it is capable of furnishing power, that region at least is youthful. If, as sometimes happens, the middle course is navigable and slow-flowing, and the lower course broken by rapids and falls, then the probability is that earth movements have occurred, so that the two regions are of different age. This is a condition which occurs relatively often in the case of large rivers.