Now, if we have a quantity of air, say at sea-level, damp but not saturated, and it is caused to ascend, either because it is warmer and therefore lighter than the surrounding air, or for some other reason, as it moves upwards the pressure upon it will decrease, it will expand, and in the act it will be steadily cooled. This cooling may after a time bring it down to the same temperature as the rest of the air at its particular level. If so, it will no longer be lighter, and the ascent will come to an end. But before this state of affairs is attained it may have reached its saturation point, and cloud production will begin. It is true that the rarefaction of the air tends to enable it to retain more vapour than it could if it were cooled without change of density. The temperature of the air being fixed, its holding power increases with decrease of pressure. But this increase is much less than the diminution due to cooling, and the result in nature must be similar to what we can see happen under the receiver of the air-pump. The condensation of water introduces another factor of great importance. It has just been said that the ascending air may be cooled so rapidly as to be reduced to the same temperature as the rest of the air at that level, and if so the ascent will end. Clearly the cessation or persistence of the upward motion depends upon whether the diminution of temperature per 100 metres of ascent is most rapid in the rising column or in the air outside it. As long as the ascending air is warmer than that outside, but at its own level, so long will ascent continue. Now, as long as no condensation was taking place, the rate of cooling would follow a simple law which produces a cooling of 1 degree for about 100 metres of ascent; but as soon as water vapour begins to pass into the liquid form, a large quantity of heat is set free, and the rate of cooling is consequently greatly lessened. Cloud production tends, therefore, to accelerate ascent, and the greater the amount of condensation, the more important will this consideration become; though, on the other hand, when once the cloud is formed, it tends to stop the rising current by shading the air and ground beneath it. On an ordinary day the rate of decrease of temperature as we ascend is rather less than the value given above, and uprising currents are soon checked. If they do extend far enough to reach cloud production, the clouds will be small, forming the smallest variety of cumulus. This is shown in Plate 43. Small irregular uprising currents have just been able to reach far enough up to have their summits tipped with cloud. [Illustration: Plate 43. SMALL CUMULUS. (_Cumulus Minor._)]