In observing cloud-motion attention should be given only to the sky overhead; at any lower angle the parallax due to viewing the clouds obliquely deprives the observation of value. It is also necessary to distinguish between the movement of the upper and of the lower clouds, as these are floating in very different parts of the atmosphere. It is comparatively rarely that the motion, say of nimbus and cirrus, is in the same direction. On a lofty mountain, strata of cloud which from below were seen to be cumulus may be passed through as layers of mist, and on emerging from them their upper surface may be seen below one. In many mountains the cloud-belt is as sharply defined as the snow-line, and its variations should be carefully observed. Clouds should occasionally be photographed as a record. This should be done especially when a type of cloud comes to be recognised as a usual one, for while exceptional forms may prove interesting, a record of the usual forms is certain to be valuable. In this connection a protest may be made against the horrible custom of some amateur and of many professional photographers of printing in clouds from some stock negative in their pictures of scenery. The cloud is an essential part of a picture, and it is better to leave an over-exposed sky of natural cloud than to insert a beautiful representation of a cloud-form which may be one never visible in the particular place or at the particular season. _Mist, Fog and Haze._—Mist or fog at low levels will of course be recorded whenever observed, and its density and duration noted. A good way to define the density of thick fog is to measure the number of yards at which an object becomes indistinguishable, and the most convenient object for the purpose is a person. Light mists lie over water or marshes at certain hours in particular seasons, and their behaviour should always be observed. It often happens that the distant view from a height is obscured by a haze not due to moisture, and this appearance should be noticed with a view to discovering its cause. The smoke from forest or prairie fires in Canada sometimes produces so thick a haze as to put a stop to surveying operations for weeks at a time. Haze is often due to dust blown from deserts, or ejected from volcanoes, and sometimes to swarms of insects.