The behaviour of the clouds will depend upon the relative shares in their production borne by interconvection pure and simple and by the wave oscillations. If the stratum is one in which cloud would actually be formed independently of the up-and-down movements, all this will be able to do will be to arrange the cloudlets at their birth, and these will then continue to exist, drifting with the general horizontal movement of the air like any other cloud of the same order. On the other hand, if the production of cloud is dependent upon the vertical oscillations, the cloudlets or lines of cloud will move with the air waves, and their rate of motion and direction of motion will be determined by the rate and direction of the waves, which may be quite different from that of the air at that stratum as a whole. The ascending waves will be marked by lines of cloud generally rounder and better defined on their advancing sides, while the descending troughs will be marked by clear intervals. Wave movements of the necessary kind are frequently very complicated, and it is not by any means a rare occurrence to see the wave lines in one part of the sky at all sorts of angles with similar lines in other parts, or even to see two or more sets of waves at different altitudes crossing one another. Either phenomenon is always accompanied by rapid changes in the cloud, and the rippled structure is short-lived. This was the case with the clouds shown in Plate 54. Plate 53, on the contrary, shows great uniformity in the wave lines, and although the vertical oscillation is probably the main cause of condensation, the form was unusually persistent. Irregular patches of wave disturbance, affecting a plane occupied by cirro-stratus vittatus, are shown in Plate 57. In this case the wave systems only touch the cloud plane here and there, and the places of contact varied rapidly. It is pretty clear from this photograph that the idea of the waves being formed at a surface of contact between two diverse currents will not suffice. The bands of the cirro-stratus are for the most part unbroken and unaffected; it is only here and there that the wave region touches them. [Illustration: Plate 57. WAVED CIRRO-STRATUS. (_Cirro-stratus Undatus._)] The conclusions at which we have arrived are simple, and there is little room for doubt as to their main correctness, but there are numerous minute features presented by these beautiful cloud patterns which await interpretation, and they reveal complicated oscillatory movements in the air which are difficult to account for, whether we seek their originating causes or the mechanics of their motions. CHAPTER IX CLOUD ALTITUDES