After the foregoing explanation, it is easy to see why at a given time the floating cloudlets should have a common base level. This is the height to which the air must attain before reaching its saturation temperature. Each cloudlet marks an uprising current, and the intervals show the position of the counterbalancing descending streams. A larger variety is shown in Plate 44. In this the level base and generally pyramidal shape is shown, and also the hard, rounded upper surface. The thickness of this cloud was about 500 metres. When clouds like these are visible, they may be the beginning of larger ones, and the only way to judge whether they are likely to develop into rain- or shower-clouds is to watch them. If they are seen to be growing larger, and particularly if detached fragments are developing into clouds, further growth is almost certain, and rain is probable. [Illustration: Plate 44. CUMULUS.] If great towering masses are making their appearance with little dark fragments between them, as shown in Plate 45, then smart showers may be confidently expected. The cloud figured was a shower-cloud, and the distance is seen through the veil of falling rain. The height and thickness of this particular cloud were measured just after its photograph had been taken. Its base was 1200 metres above the ground, and its summit was 1500 metres further. Its thickness from summit to base was, therefore, not much short of a mile, and the total contents of the cloud were probably between one and a half and two cubic miles. The upper contour is hard and rounded, as in the smaller cloud of Plate 44, but the whole cloud is much larger. [Illustration: Plate 45. LARGE CUMULUS. (_Cumulus Major._)] We have already explained that there seems to be a definite connection between the thickness of such clouds and the amount of precipitation from them. Small cumulus, less than 120 metres thick, rarely produces rain, and nothing like a heavy shower is likely unless the thickness exceeds 400 metres. In winter, especially in hard frost, snow crystals may fall from the smallest cloud, even from little fragments only a few metres thick, but the quantity of water so precipitated will, of course, be small. As long as the top of the cumulus is rounded and clearly defined, the conditions of aërial equilibrium are stable, and the growth of the cloud has been brought to an end by a stoppage of the ascending current. In Plate 45 the ascent has been hindered both by the mechanical action of the falling raindrops and by the cooling of the lower parts of the ascending column by the descent into it of the cool drops from its colder upper part. This is probably one of the chief reasons why a shower-cloud never maintains its activity as a rain producer for more than a very limited period. As the cloud drifts over the landscape, it seldom maintains its showery character for more than ten or twenty miles, often for much less.