Two other diagrams show the distribution of heat at the time of full-moon, one half of the curve showing the temperatures along the equator from the edge of the disc to the centre, the other along a meridian from this centre to the pole. This diagram (here reproduced) exhibits the quick rise of temperature of the oblique rim of the moon and the nearly uniform heat of the central half of its surface; the diminution of heat towards the pole, however, is slower for the first half and more rapid for the latter portion. It is an interesting fact that the temperature near the margin of the full-moon increases towards the centre more rapidly than it does when the same parts are observed during the early phases of the first quarter. Mr. Very explains this difference as being due to the fact that the full-moon to its very edges is fully illuminated, all the shadows of the ridges and mountains being thrown vertically or obliquely _behind them._ We thus measure the heat reflected from the _whole_ visible surface. But at new moon, and somewhat beyond the first quarter, the deep shadows thrown by the smallest cones and ridges, as well as by the loftiest mountains, cover a considerable portion of the visible surface, thus largely reducing the quantity of light and heat reflected or radiated in our direction. It is only at the full, therefore, that the maximum temperature of the whole lunar surface can be measured. It must be considered a proof of the delicacy of the heat-measuring instruments that this difference in the curves of temperature of the different parts of the moon's surface and under different conditions is so clearly shown. _The Application of the Preceding Results to the Case of Mars._ This somewhat lengthy account of the actual state of the moon's surface and temperature is of very great importance in our present enquiry, because it shows us the extraordinary difference in mean and extreme temperatures of two bodies situated at the same distance from the sun, and therefore receiving exactly the same amount of solar heat per unit of surface. We have learned also what are the main causes of this almost incredible difference, namely: (1) a remarkably rugged surface with porous and probably cavernous rock-texture, leading to extremely rapid radiation of heat in the one; as compared with a comparatively even and well-compacted surface largely clad with vegetation, leading to comparatively slow and gradual loss by radiation in the other: and (2), these results being greatly intensified by the total absence of a protecting atmosphere in the former, while a dense and cloudy atmosphere with an ever-present supply of water-vapour, accumulates and equalises the heat received by the latter.