theory of stability is really so clear that it is a wonder Laplace did not see it straight away. But he was so intensely mathematical that he often did not see the simplicity of the results which he attained by complicated mathematical analysis. Suppose you shift a quantity of ice from one place to another on the globe; and suppose, instead of

sea-water of its actual density, we had ideal water of a twentieth part of the density of sea-water, then the attraction of this solid mass of ice upon the water would be calculated simply by the attraction of the ice upon the ideal water. The figure would be such that the surface of the ideal water would be everywhere perpendicular to the surface of the

globe. But when we deal with real sea-water we have a piece of very delicate and nice mathematical book-keeping which reminds one of certain rules in Compound Interest. The ice attracts the fluid it displaces; but the displaced fluid itself attracts the remaining fluid and so contributes to the resultant attractive force. First calculate the amount of the attraction on the water due to the ice-cap alone; then calculate the increase of the attractive force due to the displaced water; then calculate this increase, to the second degree of approximation, and so on; and thus you get at the result. There have been different calculations founded on largely varied assumptions for data. Mr. H. D. Heath and the Rev. O. Fisher made calculations which differed somewhat widely from mine as well as from Croll's, but I believe them all to be consistent. My result, 380 feet, seemed to be immensely smaller than the others. In point of fact, Croll omitted to notice that mine referred to an ice-cap gathering on an ideal set of islands, and that I supposed the whole land to be distributed uniformly, and the ice to be

placed upon the top of these islands. I took the actual proportion of the area of land to water, roughly estimated, as being one-third land and two-thirds water. To bring my result into comparison with the others quoted, you must therefore treble it, because I took only one-third area as covered with ice. Three times 380 1140, which is my number on a certain supposition; while Heath's is 650 on a supposition not quite the same as mine. Pratt's estimates is still greater-something like 2,000 feet.

=

Now, if I could say anything to throw light upon the real question of extensions of ice in the southern or northern polar regions, and the effect of such extensions upon the sea level and upon the climate in past times, I should feel my attempt was certainly not insignificant. But I cannot even look upon such an attempt; I can merely point out certain fallacies and set certain limits to former suppositions. We cannot have an ice-cap on the Antarctic continent 12 miles thick, as Mr. Croll has calculated. I can bring substantial evidence against this. But Mr. Croll's argument does not

at all stand upon that number; he is satisfied with a small fraction of it, 3,000 or 5,000, or 12,000 feet, instead of 65,000 feet. He is satisfied with an icecap of a comparatively moderate thickness as a sufficient cause for some most important fluctuations of sea level which geological history proves to have taken place. It seems to me that Croll is here meeting my case, and we may find the most probable explanation of some of our familiar changes of sea level—familiar even to people who are not geologists-in Croll's supposition of shiftings of ice either on the Antarctic or the Arctic hemisphere.

In the first place I shall ask you to imagine that the Antarctic continent for some unknown cause had at one time a distribution of ice over it thicker by 1,000 feet than at another time. To be more accurate I would say 1,200 so as to correspond with the area equivalent to 1,000 feet of water. Mr. Murray has made a very careful estimate of the area of the Antarctic continent, which shows that the area is about one-fortieth of the area of the whole earth. Croll makes it more; but Mr. Murray has given us the more recent and more

probable estimate. Now I shall merely ask you to think of a great ice-cap melted off the Antarctic continent, an ice-cap 1,200 feet thick equivalent to 1,000 feet depth of water. Imagine this mass of ice melting and flowing into the ocean; it would just raise the level of the ocean by one-fortieth of a thousand feet, a quarter of a hundred, or 25 feet. Our latest change of sea level here on the Firth of Clyde was only 10 feet. We do not know exactly the date, but it is quite certain that it was not very many thousand years ago. The water-level in the Firth of Clyde was then 10 feet higher than it is now, and that change of level would, on the theory I have stated, involve the melting of only about 400 feet of ice from the southern continent, which would raise the water 10 feet all over the world. When the theory of gravitation is taken into account in the manner I have indicated, the water thus brought to one of the poles and converted into an ice-cap, or the water that flows away from the melting ice-cap leaving a deficiency of solid water, does by its own gravitation always exaggerate the effect. Before we go into any