theory of Secular changes of climate. This terminates the part of the inquiry relating to oceanic circulation.

We now come to the crucial test of the theories respecting the cause of the glacial epoch, viz., Warm Inter-glacial Periods. In Chapters XV. and XVI. I have given a statement of the geological facts which go to prove that that long epoch known as the Glacial was not one of continuous cold, but consisted of a succession of cold and warm periods. This condition of things is utterly inexplicable on every theory of the cause of the glacial epoch which has hitherto been advanced; but, according to the physical theory of secular changes of climate under consideration, it follows as a necessary consequence. In fact, the amount of geological evidence which has already been accumulated in reference to inter-glacial periods may now be regarded as perfectly sufficient to establish the truth of that theory.

If the glacial epoch resulted from some accidental distribution of sea and land, then there may or may not have been more than one glacial epoch, but if it resulted from the cause which we have assigned, then there must have been during the geological history of the globe a succession of glacial epochs corresponding to the secular variations in the eccentricity of the earth's orbit. A belief in the existence of recurring glacial epochs has been steadily gaining ground for many years past. I have, in Chapter XVIII., given at some length the facts on which this belief rests. It is true that the geological evidence of glacial epochs in prior ages is meagre in comparison with that of the glacial epoch of Post-tertiary times; but there is a reason for this in the nature of geological evidence itself. Chapter XVII. deals with the geological records of former glacial epochs, showing that they are not only imperfect, but that there is good reason why they should be so, and that the imperfection of the records in reference to them cannot be advanced as an argument against their existence.

If the glacial epoch resulted from a high condition of eccentricity, we have not only a means of determining the positive date of that epoch, but we have also a means of determining

geological time in absolute measure. For if the glacial epochs of prior ages correspond to periods of high eccentricity, then the intervals between those periods of high eccentricity become the measure of the intervals between the glacial epochs. The researches of Lagrange and Leverrier into the secular variations of the elements of the orbits of the planets enable us to determine with tolerable accuracy the values of the eccentricity of the earth's orbit for, at least, four millions of years past and future. With the view of determining those values, I several years ago computed from Leverrier's formula the eccentricity of the earth's orbit and longitude of the perihelion, at intervals of ten thousand and fifty thousand years during a period of three millions of years in the past, and one million of years in the future. The tables containing these values will be found in Chapter XIX. These tables not only give us the date of the glacial epoch, but they afford, as will be seen from Chapter XXI., evidence as to the probable date of the Eocene and Miocene periods.

Ten years ago, when the theory was first advanced, it was beset by a very formidable difficulty, arising from the opinions which then prevailed in reference to geological time. One or two glacial epochs in the course of a million of years was a conclusion which at that time scarcely any geologist would admit, and most would have felt inclined to have placed the last glacial epoch at least one million of years back. But then if we assume that the glacial epoch was due to a high state of eccentricity, we should be compelled to admit of at least two glacial epochs during that lapse of time. It was the modern doctrine that the great changes undergone by the earth's crust were produced, not by convulsions of nature, but by the slow and almost imperceptible action of rain, rivers, snow, frost, ice, &c., which impressed so strongly on the mind of the geologist the vast duration of geological periods. When it was considered that the rocky face of our globe had been carved into hills and dales, and ultimately worn down to the sea-level by means of those apparently trifling agents, not only once or twice, but many times, during past ages, it was not surprising

that the views entertained by geologists regarding the immense antiquity of our globe should not have harmonised with the deductions of physical science on the subject. It had been shown by Sir William Thomson and others, from physical considerations relating to the age of the sun's heat and the secular cooling of our globe, that the geological history of our earth's crust must be limited to a period of something like one hundred millions of years. But these speculations had but little weight when pitted against the stern and undeniable facts of sub-aërial denudation. How, then, were the two to be reconciled? Was it the physicist who had under-estimated geological time, or the geologist who had over-estimated it? Few familiar with modern physics, and who have given special attention to the subject, would admit that the sun could have been dissipating his heat at the present enormous rate for a period much beyond one hundred millions of years. The probability was that the amount of work performed on the earth's crust by the denuding agents in a period so immense as a million of years was, for reasons stated in Chapter XX., very much under-estimated. But the difficulty was how to prove this. How was it possible to measure the rate of operation of agents so numerous and diversified acting with such extreme slowness and irregularity over so immense areas? In other words, how was it possible to measure the rate of sub-aërial denudation? Pondering over this problem about ten years ago, an extremely simple and obvious method of solving it suggested itself to my mind. This method-the details of which will be found in Chapter XX.-showed that the rate of subaërial denudation is enormously greater than had been supposed. The method is now pretty generally accepted, and the result has already been to bring about a complete reconciliation between physics and geology in reference to time.

Chapter XXI. contains an account of the gravitation theories of the origin of the sun's heat. The energy possessed by the sun is generally supposed to have been derived from gravita tion, combustion being totally inadequate as a source. But something more than gravitation is required before we can

account for even one hundred millions of years' heat. Gravitation could not supply even one-half that amount. There must be some other and greater source than that of gravitation. There is, however, as is indicated, an obvious source from which far more energy may have been derived than could have been obtained from gravitation.

The method of determining the rate of sub-aërial denudation enables us also to arrive at a rough estimate of the actual mean thickness of the stratified rocks of the globe. It will be seen from Chapter XXII. that the mean thickness is far less than is generally supposed.

The physical cause of the submergence of the land during the glacial epoch, and the influence of change in the obliquity of the ecliptic on climate, are next considered. In Chapter XXVI. I have given the reasons which induce me to believe that coal is an inter-glacial formation.

The next two chapters-the one on the path of the ice in north-western Europe, the other on the north of England ice-sheet-are reprints of papers which appeared a few years ago in the Geological Magazine. Recent observations have confirmed the truth of the views advanced in these two chapters, and they are rapidly gaining acceptance among geologists.

I have given, at the conclusion, a statement of the molecular theory of glacier motion-a theory which I have been led to modify considerably on one particular point.

There is one point to which I wish particularly to direct attention—viz., that I have studiously avoided introducing into the theories propounded anything of a hypothetical nature. There is not, so far as I am aware, from beginning to end of this volume, a single hypothetical element: nowhere have I attempted to give a hypothetical explanation. The conclusions are in every case derived either from facts or from what I believe to be admitted principles. In short, I have aimed to prove that the theory of secular changes of climate follows, as a necessary consequence, from the admitted principles of physical science.

CHAPTER II.

OCEAN-CURRENTS IN RELATION TO THE DISTRIBUTION
OF HEAT OVER THE GLOBE.

The absolute Heating-power of Ocean-currents.-Volume of the Gulf-stream.— Absolute Amount of Heat conveyed by it.-Greater Portion of Moisture in inter-tropical Regions falls as Rain in those Regions.-Land along the Equator tends to lower the Temperature of the Globe.-Influence of Gulf-stream on Climate of Europe.-Temperature of Space.-Radiation of a Particle.-Professor Dove on Normal Temperature.-Temperature of Equator and Poles in the Absence of Ocean-currents. - Temperature of London, how much due to Ocean-currents.

The absolute Heating-power of Ocean-currents.-There is perhaps no physical agent concerned in the distribution of heat over the surface of the globe the influence of which has been so much underrated as that of ocean-currents. This is, no doubt, owing to the fact that although their surface-temperature, direction, and general influence have obtained considerable attention, yet little or nothing has been done towards determining the absolute amount of heat or of cold conveyed by them or the resulting absolute increase or decrease of temperature.

The modern method of determining the amount of heateffects in absolute measure is, doubtless, destined to cast new light on all questions connected with climate, as it has done, and is still doing, in every department of physics where energy, under the form of heat, is being studied. But this method has hardly as yet been attempted in questions of meteorology; and owing to the complicated nature of the phenomena with which the meteorologist has generally to deal, its application will very often prove practically impossible. Nevertheless, it is · particularly suitable to all questions relating to the direct