d. Glaciation near end of Proterozoic in Australia, Norway,

and China.

3. Paleozoic. (1/4 of geological time)

a. Late Ordovician(?). Local in Arctic Norway.

b. Silurian. Local in Alaska.

c. Early Devonian. Local in South Africa.

d. Early Permian. World-wide and very severe.

4. Mesozoic and Cenozoic. (1/4 of geological time)

a-b. None definitely determined during Mesozoic, although there appears to have been periods of cooling (a) in the late Triassic, and (b) in the late Cretacic, with at least local glaciation in early Eocene.

c. Severe glacial period during Pleistocene.

This table suggests an interesting inquiry. During the last few decades there has been great interest in ancient glaciation and geologists have carefully examined rocks of all ages for signs of glacial deposits. In spite of the large parts of the earth which are covered with deposits belonging to the Mesozoic and Cenozoic, which form the last quarter of geological time, the only signs of actual glaciation are those of the great Pleistocene period and a few local occurrences at the end of the Mesozoic or beginning of the Cenozoic. Late in the Triassic and early in the Jurassic, the climate appears to have been rigorous, although no tillites have been found to demonstrate glaciation. In the preceding quarter, that is, the Paleozoic, the Permian glaciation was more severe than that of the Pleistocene, and the Devonian than that of the Eocene, while the Ordovician evidences of low temperature are stronger than those at the end of the Triassic. In view of the fact that rocks of Paleozoic age cover much smaller areas than do those of later age, the three Paleozoic glaciations seem to indicate a relative frequency of glaciation. Going back to the Proterozoic, it is astonishing to find that evidence of two highly developed glacial periods, and possibly four, has been discovered. Since the Indian and the African glaciations of Proterozoic times are as yet undated, we cannot be sure that they are not of the same date as the others. Nevertheless, even two is a surprising number, for not only are most Proterozoic rocks so metamorphosed that possible evidences of glacial origin are destroyed, but rocks of that age occupy far smaller areas than either those of Paleozoic or, still more, Mesozoic and Cenozoic age. Thus the record of the last three-quarters of geological time suggests that if rocks of all ages were as abundant and as easily studied as those of the later periods, the frequency of glacial periods would be found to increase as one goes backward toward the beginnings of the earth's history. This is interesting, for Jeans holds that the chances that the stars would approach one another were probably greater in the past than at present. This conclusion is based on the assumption that our universe is like the spiral nebulæ in which the orbits of the various members are nearly circular during the younger stages. Jeans considers it certain that in such cases the orbits will gradually become larger and more elliptical because of the attraction of one body for another. Thus as time goes on the stars will be more widely distributed and the chances of approach will diminish. If this is correct, the agreement between astronomical theory and geological conclusions suggests that the two are at least not in opposition.

The first quarter of geological time as well as the last three must be considered in this connection. During the Archeozoic, no evidence of glaciation has yet been discovered. This suggests that the geological facts disprove the astronomical theory. But our knowledge of early geological times is extremely limited, so limited that lack of evidence of glaciation in the Archeozoic may have no significance. Archeozoic rocks have been studied minutely over a very small percentage of the earth's land surface. Moreover, they are highly metamorphosed so that, even if glacial tills existed, it would be hard to recognize them. Third, according to both the nebular and the planetesimal hypotheses, it seems possible that during the earliest stages of geological history the earth's interior was somewhat warmer than now, and the surface may have been warmed more than at present by conduction, by lava flows, and by the fall of meteorites. If the earth during the Archeozoic period emitted enough heat to raise its surface temperature a few degrees, the heat would not prevent the development of low forms of life but might effectively prevent all glaciation. This does not mean that it would prevent changes of climate, but merely changes so extreme that their record would be preserved by means of ice. It will be most interesting to see whether future investigations in geology and astronomy indicate either a semi-uniform distribution of glacial periods throughout the past, or a more or less regular decrease in frequency from early times down to the present.

2. The Pleistocene glacial period was divided into at least four epochs, while in the Permian at least one inter-glacial epoch seems certain, and in some places the alternation between glacial and non-glacial beds suggests no less than nine. In the other glaciations the evidence is not yet clear. The question of periodicity is so important that it overthrows most glacial hypotheses. Indeed, had their authors known the facts as established in recent years, most of the hypotheses would never have been advanced. The carbon dioxide hypothesis is the only one which was framed with geologically rapid climatic alternations in mind. It certainly explains the facts of periodicity better than does any of its predecessors, but even so it does not account for the intimate way in which variations of all degrees from those of the weather up to glacial epochs seem to grade into one another.

According to our stellar hypothesis, occasional groups of glacial epochs would be expected to occur close together and to form long glacial periods. This is because many of the stars belong to groups or clusters in which the stars move in parallel paths. A good example is the cluster in the Hyades, where Boss has studied thirty-nine stars with special care.[120] The stars are grouped about a center about 130 light years from the sun. The stars themselves are scattered over an area about thirty light years in diameter. They average about the same distance apart as do those near the sun, but toward the center of the group they are somewhat closer together. The whole thirty-nine sweep forward in essentially parallel paths. Boss estimates that 800,000 years ago the cluster was only half as far from the sun as at present, but probably that was as near as it has been during recent geological times. All of the thirty-nine stars of this cluster, as Moulton[121] puts it, "are much greater in light-giving power than the sun. The luminosities of even the five smallest are from five to ten times that of the sun, while the largest are one hundred times greater in light-giving power than our own luminary. Their masses are probably much greater than that of the sun." If the sun were to pass through such a cluster, first one star and then another might come so near as to cause a profound disturbance in the sun's atmosphere.

3. Another important point upon which a glacial hypothesis may come to grief is the length of the periods or rather of the epochs which compose the periods. During the last or Pleistocene glacial period the evidence in America and Europe indicates that the inter-glacial epochs varied in length and that the later ones were shorter than the earlier. Chamberlin and Salisbury, from a comparison of various authorities, estimate that the intervals from one glacial epoch to another form a declining series, which may be roughly expressed as follows: 16-8-4-2-1, where unity is the interval from the climax of the late Wisconsin, or last glacial epoch, to the present. Most authorities estimate the culmination of the late Wisconsin glaciation as twenty or thirty thousand years ago. Penck estimates the length of the last inter-glacial period as 60,000 years and the preceding one as 240,000.[122] R. T. Chamberlin, as already stated, finds that the consensus of opinion is that inter-glacial epochs have averaged five times as long as glacial epochs. The actual duration of the various glaciations probably did not vary in so great a ratio as did the intervals from one glaciation to another. The main point, however, is the irregularity of the various periods.

The relation of the stellar electrical hypothesis to the length of glacial epochs may be estimated from column C, in Table 5. There we see that the distances at which a star might possibly disturb the sun enough to cause glaciation range all the way from 120 billion miles in the case of a small star like the sun, to 3200 billion in the case of Betelgeuse, while for double stars the figure may rise a hundred times higher. From this we can calculate how long it would take a star to pass from a point where its influence would first amount to a quarter of the assumed maximum to a similar point on the other side of the sun. In making these calculations we will assume that the relative rate at which the star and the sun approach each other is about twenty-two miles per second, or 700 million miles per year, which is the average rate of motion of all the known stars. According to the distances in Table 5 this gives a range from about 500 years up to about 10,000, which might rise to a million in the case of double stars. Of course the time might be relatively short if the sun and a rapidly moving star were approaching one another almost directly, or extremely long if the sun and the star were moving in almost the same direction and at somewhat similar rates,--a condition more common than the other. Here, as in so many other cases, the essential point is that the figures which we thus obtain seem to be of the right order of magnitude.

4. Post-glacial climatic stages are so well known that in Europe they have definite names. Their sequence has already been discussed in