The part of Jeans' work which specially concerns us is his study of the probability that some other star will approach the sun closely enough to have an appreciable gravitative or electrical effect, and thus cause disturbances in the solar atmosphere. Of course both the star and the sun are moving, but to avoid circumlocution we shall speak of such mutual approaches simply as approaches of the sun. For our present purpose the most fundamental fact may be summed up in a quotation from Jeans in which he says that most stars "show evidence of having experienced considerable disturbance by other systems; there is no reason why our solar system should be expected to have escaped the common fate." Jeans gives a careful calculation from which it is possible to derive some idea of the probability of any given degree of approach of the sun and some other star. Of course all such calculations must be based on certain assumptions. The assumptions made by Jeans are such as to make the probability of close approaches as great as possible. For example, he allows only 560 million years for the entire evolution of the sun, whereas some astronomers and geologists would put the figure ten or more times as high. Nevertheless, Jeans' assumptions at least show the order of magnitude which we may expect on the basis of reasonable astronomical conclusions. According to the planetary hypothesis of sunspots, the difference in the effect of Jupiter when it is nearest and farthest from the sun is the main factor in starting the sunspot cycle and hence the corresponding terrestrial cycle. The climatic difference between sunspot maxima and minima, as measured by temperature, apparently amounts to at least a twentieth and perhaps a tenth of the difference between the climate of the last glacial epoch and the present. We may suppose, then, that a body which introduced a gravitative or electrical factor twenty times as great as the difference in Jupiter's effect at its maximum and minimum distances from the sun would cause a glacial epoch if the effect lasted long enough. Of course the other planets combine their effects with that of Jupiter, but for the sake of simplicity we will leave the others out of account. The difference between Jupiter's maximum and minimum tidal effect on the sun amounts to 29 per cent of the planet's average effect. The corresponding difference, according to the electrical hypothesis, is about 19 per cent, for electrostatic action varies as the square of the distance instead of as the cube. Let us assume that a body exerting four times Jupiter's present tidal effect and placed at the average distance of Jupiter from the sun would disturb the sun's atmosphere twenty times as much as the present difference between sunspot maxima and minima, and thus, perhaps, cause a glacial period on the earth.