The near neighbourhood of the ground is not essential. As long as the temperature of the air at any level is rising, so long interconvection must occur. The process will be independent of the presence or absence of wind. All that wind can do is to mix up the air at different levels, breaking the system of currents and reducing it to, so to say, a finer texture, or producing eddies, if strong enough, which direct the currents and gather them into definite channels. The final result in any case is that, with rising temperature, water vapour is steadily borne upwards from the ground. As it ascends the air becomes cooler, and yet retains its water vapour. When the rising currents are large they mix little with the descending dry air, and on reaching a certain level condensation takes place, and we have the beginning of a cumulus. If they are of a more moderate size they will ascend less rapidly, the admixture with descending air will bear a larger proportion to the whole, and the plane at which condensation will begin will be higher, and then each small column will be tipped with a ball of alto-cumulus. Make the interconvection currents smaller still, and the cloud plane will be lifted yet higher, and we shall have cirro-cumulus or cirro-macula. Now, the more even the distribution of temperature on the ground the less the probability of coarse interconvection, and the same is true of any higher stratum of air, provided it is free from disturbing influences from outside. If, therefore, we have large currents near the ground, ending, as they must, in cumulus, it has already been explained that these clouds stop the action, and the general system of large currents will be restricted to the region in which they occur. At some distance above the lower clouds the only difference will be that water vapour has been brought up to their level in great abundance. Smaller systems of interconvection can then exist, and so we may have the spectacle of several layers of cloud--cumulus capping the great currents of lower regions, alto-cumulus forming the summits of the smaller currents of intermediate regions, and cirro-cumulus floating far above both.