But notwithstanding the exceedingly high temperature of the exterior, enforced upon it for the brief time of transit through the atmosphere, it is probable that all large meteorites, if they could be reached at once on striking the earth, would be found to be cold, because the smooth, black, varnishlike crust which always incases them as a result of intense heat is never thick. On one occasion a meteor which was seen to fall in India was dug out of the ground as quickly as possible, and found to be, not hot as was expected, but coated thickly over with ice frozen on it from the moisture in the surrounding soil. As to the composition of shooting stars, and their probable mass, and its effect upon the earth, our data are quite insufficient. The lines of sodium and magnesium have been hurriedly caught in the spectroscope, and, estimating on the basis of the light emitted by them, the largest meteors must weigh ounces rather than pounds. Nevertheless, it is interesting to inquire what addition the continual fall of many millions daily upon the earth makes to its weight: somewhere between thirty and fifty thousand tons annually is perhaps a conservative estimate, but even this would not accumulate a layer one inch in thickness over the entire surface of the earth in less than a thousand million years. Many hundreds of the meteors actually seen to fall, together with those picked up accidentally, are recovered and prized as specimens of great value in our collections, the richest of which are now in New York, Paris, and London. The detailed investigation of them is rather the province of the chemist, the crystallographer and the mineralogist than of the astronomer whose interest is more keen in their life history before they reach the earth. To distinguish a stony meteorite from terrestrial rock substances is not always easy, but there is usually little difficulty in pronouncing upon an iron meteorite. These are most frequently found in deserts, because the dryness of the climate renders their oxidation and gradual disappearance very slow. The surface of a suspected iron meteorite is polished to a high luster and nitric acid is poured upon it. If it quickly becomes etched with a characteristic series of lines, or a sort of cross-hatching, it is almost certain to be a meteorite. Occasionally carbon has been found in meteorites, and the existence of diamond has been suspected. The minerals composing meteorites are not unlike terrestrial materials of volcanic origin, though many of them are peculiar to meteorites only. More than one-third of all the known chemical elements have been found by analysis in meteorites, but not any new ones.