Without laying too much stress on these visionary concepts of life a million million years hence, we may nevertheless think of this as the period in round numbers after which the inevitable wastage of the sun’s weight is likely to drive life off the earth. Venus, with a mean temperature some 60 degrees higher than the earth, is probably rather too hot for life at present. But after a million million years, the temperature of Venus will have fallen by 40 degrees, and what the earth is now, Venus may perhaps be somewhere between one and two million million years hence. Whether life will then inhabit Venus we cannot know, and it would be futile to guess, but there is at least a chance that as the earth fails, Venus may step into its place. Possibly Venus may be followed by Mercury in due course, but the present evidence is that Mercury is devoid of atmosphere, in which case it is hard to imagine it as a home for life at all resembling that which now inhabits the earth. So far we have considered only the normal course of events; a variety of accidents may bring the human race to an end long before a million million years have elapsed. To mention only possible astronomical occurrences, the sun may run into another star, any asteroid may hit any other asteroid and, as a result, be so deflected from its path as to strike the earth, any of the stars in space may wander into the solar system and, in so doing, upset all the planetary orbits to such an extent that the earth becomes impossible as an abode of life. It is difficult to estimate the likelihood of any of these events happening but they all seem very improbable, and the first and last highly so. Let us disregard them all. A danger remains which cannot be so lightly dismissed. Let us first state it in technical language. The sun is a main-sequence star, and is moreover very near to the left-hand edge of the main-sequence in the Russell diagram (p. 278). Beyond this edge is a region of the diagram which is completely untenanted by stars. We have supposed this region to be untenanted by stars because the stellar configurations it represents would be unstable. Stars pass through it rapidly until they find a stable configuration, and so end up in a region which can be permanently tenanted by stars. Now the next stable configurations beyond this region are those of the white dwarfs, and as these are less massive as a class than the main-sequence stars, the general trend of stellar evolution appears to be from main-sequence star to white dwarf. On this view the white dwarfs must have previously been main-sequence stars which wandered across the left-hand edge of the band of stable configurations and then fell through the unstable region until they resumed stability as white dwarfs.