By means of the process of terrestrial radiation ice is artificially formed in Bengal, "where the substance is never formed naturally. Shallow pits are dug, which are partially filled with straw, and on the straw flat pans containing water which had been boiled is exposed to the clear firmament. The water is a very powerful radiant, and sends off its heat into space. The heat thus lost cannot be supplied from the earth--this source being cut off by the non-conducting straw. Before sunrise a cake of ice is formed in each vessel.... To produce the ice in abundance, the atmosphere must not only be clear, but it must be comparatively free from aqueous vapour." Considering, therefore, the important consequences attending both terrestrial and solar radiation, it appears to us that observations from radiation thermometers are of much more utility in judging of climate than is usually supposed. These observations are very scanty; and what few are upon record are not very reliable, principally from bad exposure of the instruments, while the want of uniformity in construction may be another cause. Herschell's actinometer and Pouillet's pyrheliometer, instruments for ascertaining the absolute heating effect of the sun's rays, should, however, be more generally employed by meteorologists. In comparing observations on radiation it should be kept in mind, that "the difference between a thermometer which, properly confined [or shaded], gives the true temperature of the night air, and one which is permitted to radiate freely towards space, must be greater at high elevations than at low ones;"[6] because the higher the place, the less the thickness of the vapour-screen to intercept the radiation. =83. Solar Radiation Thermometer.=--"As the interchange of heat between two bodies by radiation depends upon the relative temperature which they respectively possess, the earth, by the rays transmitted from the sun during the day, must be continually gaining an accession of heat, which would be far from being counterbalanced by the opposite effect of its own radiation into space. Hence, from sunrise till two or three hours after mid-day, the earth goes on gradually increasing in temperature, the augmentation being greatest where the surface consists of materials calculated, from their colour and texture, to absorb heat, and where it is deficient in moisture, which, by its evaporation, would have a tendency to diminish it."[7] It is, therefore, important to have instruments for measuring the efficacy of solar radiation, apart from those for exhibiting the temperature of the place in the shade. [Illustration: Fig. 63.]