We must look for another explanation. It occurred to Helmholtz, one of the most eminent investigators in the domain of the mechanical theory of heat, that, instead of the meteorites, parts of the sun itself might fall towards its centre, or, in other words, that the sun was shrinking. Owing to the high gravitation of the sun (27.4 times greater than on the surface of the earth), the shrinkage would liberate a great amount of heat. Helmholtz calculated that, in order to cover the heat expenditure of the sun, a shrinkage of its diameter by 60 m. annually would be required. If the sun’s diameter should only be diminished by one-hundredth of one per cent.—a change which we should not be able to establish—the heat loss would be covered for more than 2000 years. That seems at first satisfactory. But if we proceed with our estimate, we find that if the sun went on losing as much heat as at present for seventeen million years it would have to contract within this period to a quarter of its present volume, and would therefore acquire a density like that of the earth. Long before that, however, the radiation from the sun would have been decreased so powerfully that the temperature on the earth’s surface would no longer rise above freezing-point. Helmholtz, on this argument, limited the further existence of the earth to about six million years. That is less satisfactory. But we know nothing of the future and must be content with possibilities. Not so, however, if we calculate back with the aid of Helmholtz’s theory. According to this theory, and according to Helmholtz’s own data, a state like the present cannot have existed for more than ten million years. Since, now, geologists have come to the conclusion that the petrefactions which we find in the fossil-bearing strata of the earth have needed at least a hundred million years for their formation, and more probably a thousand million years, and since, moreover, the still more ancient formations—the so-called precambrian strata—have been deposited in equally long or still longer periods, we see that the theory of Helmholtz is unsatisfactory.