Finally, the sun’s distance from the edge of the main-sequence cannot be estimated with anything like the degree of accuracy assumed in the foregoing calculations. The figure of 0·03 appeared as the difference of two much larger numbers, and although both of these can be estimated with fair accuracy, neither can be estimated with sufficient accuracy to justify us in treating their small difference of 0·03 as exact. The most we can say is that the sun is quite fairly near to the dangerous edge, but that any appreciable motion towards this edge is a matter of millions of millions of years. Another danger, of a more speculative kind, must also be mentioned. We have seen (p. 61) how every now and then a new star appears in the sky, shines with terrific brilliancy for a short time, and then either fades away entirely or continues to shine as an ordinary star. These apparitions are known as “novae”—new stars. In many cases the nova has been proved to be an ordinary star which was visible as a very faint star long before it appeared as a nova, flashed into brilliance for a brief span of life, and then lapsed back into commonplaceness, and it seems reasonable to suppose that all novae are of this kind, although the star may often escape detection until it assumes its brilliant nova state. These apparitions are by no means rare; something like six appear every year in the galactic system alone. Now if we suppose the galactic system to consist of 300,000 million stars, this means that, on the average, each star becomes a nova once in every 50,000 million years. What we would like to know is whether our sun is in danger of becoming a nova; for, if all kinds of stars run equal chances, it is likely to become a nova some twenty times in the next million million years. So far there is no agreement among astronomers either as to the physical causes which turn an ordinary star into a nova, or as to the physical conditions which prevail in novae. Various suggestions are in the field, but none of them wins general acceptance. It seems fairly certain that if our sun were suddenly to become a nova, its emission of light and heat would so increase as to scorch all life off the earth, but we are completely in the dark as to whether our sun runs any risk of entering the nova stage. If it does, this is probably the greatest of all the risks to which life on earth is exposed. Apart from accidents, we have seen that if the solar system is left to the natural course of evolution, the earth is likely to remain a possible abode of life for something of the order of a million million years to come.