But no sketching or drawing can make records of cirrus or alto clouds with the speed and accuracy necessary for careful study. Photography is really the only way in which the amazing wealth of detail can be truthfully portrayed. Yet even the camera has its limitations. It does not record colour, and completely fails to delineate the forms of alto-stratus, stratus, or nimbus, if they are present in the most typical condition, that is to say, when they cover the whole sky with a uniform tint. It is only when these forms are more or less broken up that a photograph, or anything other than a carefully coloured picture, will represent them at all. Cloud photography, even of the most delicate and brilliant varieties, is easy enough when the right methods are followed; but these are not the same as those which are right for portraiture or landscape work of the usual kind. The background of blue sky produces almost the same effect on the plate as the image of the cloud itself, and the whole art consists in an adequate exaggeration of the minute difference so as to reveal the details of form and structure. A slow plate--the accompanying illustrations have all been taken on Mawson and Swan’s photo-mechanical plates--extremely cautious development, and sometimes intensification of the image, are all that is necessary; but the process becomes easier if, instead of pointing the camera to the cloud, it is directed to the image formed in a properly constructed black mirror. Many of the following studies have been taken by this method, and details of the camera and processes employed will be found in a later chapter, for the convenience of any one who may be inspired to take up a fascinating branch of photography. It has been said that reference will be made to the average altitudes of the different types of cloud, and to the actual altitude of some of the varieties shown. The question will, no doubt, have occurred to some as to how those altitudes have been measured. The methods are all more or less complicated, involving rather laborious calculation. They generally depend upon simultaneous observations made from two stations at opposite ends of a measured base line. Sometimes the observations are made directly by pointing an instrument at each station to some agreed point of the cloud. It is obvious that the two directions must converge to this point. If the convergence is measured, the exact distance from either station can be calculated, and if the angle between the cloud-point and the horizon beneath it is noted, it is a simple matter to deduce the actual altitude of the cloud. At other places the observers have relied upon the comparison of photographs simultaneously taken from the two stations. In this method it is necessary to know the exact direction in which the camera is pointed, and the position of the image upon the plate then gives the direction of the cloud as seen from that particular station, and the subsequent calculations are the same.