hitherto been seen. This apparent increase of size can be observed almost anywhere in the mountainous regions of the Moon, but the diminution due to melting is rare, and requires first a surface without deep crevices, where the snow can collect, and, second, a very thin layer of snow, such as can be vaporised by the heat of the Sun in the course of the lunar day. Such surface conditions occur on portions of the maria, and sometimes on crater rims. Thus in Plate 9B the rim of Pallas [1.4, 7.9] is seen to be white with snow. In Plate 9C a large part of this whiteness has disappeared, and in 9D and 9E the rest of it has vanished.
When the snow disappears toward the end of the lunar day it is impossible to determine whether the disappearance is due to evaporation or whether it is due to a low Sun which cannot shine into the crevices of the surface, but when the snow disappears in the middle of the day, as in this case, the only possible explanation, as far as we can see, is that it has evaporated.
On the lower slopes of most of the craters, wherever the frost or snow occurs, its apparent area increases with the rising Sun, indicating the existence of deep crevices in which it collects. The extent of the snow shown on a photograph of these regions depends largely on the exposure and development of the plate, so that no satisfactory evidence can be reached by this method.
At this point, however, visual observations come to our aid. The eye is so sensitive to minute contrasts of light that we can sometimes draw the outline of the brightest portion of the snowy region, and as this area diminishes in size under the effect of the Sun's rays we can obtain a series of drawings showing the changes in shape and gradual diminution in size of this brightest region. The eight smaller figures on Plate C give some sketches made in Jamaica showing the changes in the outline of some of these small white regions after they have been illuminated for different periods by the Sun. It will be noticed that in the case of Linn6 the spot was smallest one terrestrial day after lunar midday, and then again increased in size. In the case of the crater Abulfeda e, on the other hand (see also Plate 8E [1.7, 2.7]), the spot continued to grow smaller until it disappeared, and was not seen again until the next "lunation," or limar day. The date of disappearance was about one day before sunset. Censorinus (see also Plate 6D [0.8, 0.4]) is another craterlet showing similar changes, except that the spot does not disappear until sunset.
Although the outlines of these white spots are sharply defined in the drawings for
