The science of life : fully illustrated in tone and line and including many diagrams

BOOK 3

uranium-lead in having an atomic weight of 208.

Now, the bearing of all this on the computing of geological dates is very simple. This disintegration does not go on haphazard and wildly. It is timed. Each of these transformations takes place at its own definite rate, as regularly and inevitably as the swing of a pendulum. ‘The physicist assures us that if we took a definite quantity of radium, say a milligram, then after 1,700 years there would be only half of it left—exactly halfthe rest would have turned into helium, lead, and traces of the intermediate steps. Uranium, on the other hand, is a slow disintegrator ; you would only lose half your uranium after 4,500 million years. Thorium, too, has its own particular speed of metamorphosis. But since the helium and the lead which are thus generated do not undergo any further changes, we can calculate precisely how long it would take for any given proportion of lead to be accumulated in a mineral containing uranium or thorium. And so, if we can find a rock of, say, Lower Carboniferous Age (III E 1) in which, when it was first formed, radio-active minerals containing uranium or thorium were crystallized out, we can tell its age in years by measuring the proportion of helium and of lead which has been produced in these minerals since they were first locked up in their rocky prison.

Lord Rayleigh measured the amount of helium produced by uranium-bearing minerals, and found that a gram of uranium would take nine million years to produce a cubic centimetre of helium, which agrees with calculations based on the rate of stepby-step transformation. As regards lead, it can be calculated that a ton of uranium gives rise to 1/7,400th of a gram of lead every year. ‘Thus, ifon analysis we find 1 per cent. of lead in a uranium-containing mineral, this proportion must have taken 1/ggth of 7,400 million years to accumulate ; while if there had been to per cent. of lead, the time needed would have been 1/gth of 7,400 million years. The rate for thorium is slower, but the principle is the same.

The one assumption made is that the radium clocks never change their rate of going ; and physicists, after trying in vain to alter that rate by every conceivable means, are agreed that they never do. There are various small mathematical corrections to be made, and various precautions to be taken in the practical procedure; but all such difficulties can be readily overcome. How they are overcome is clearly set forth in

256

THE SCIENCE OF LIFE

CHAPTER 5

Professor Holmes’ little book, The Age of the Earth.

There, too, will be found an account of other methods of estimating geological time—as by the rate of erosion of the land, by the rate of deposition of new sediments, by the rate at which the sea, which is always receiving salts and always evaporating, grows salter, by timing geological processes against periods of known length, such as the precession of the equinoxes with its 21,000 years cycle, and so forth. ‘These all supply checks upon our radio-active estimates.

In addition there are one or two methods by which, at certain pages of the record of the past, we can get even an estimate in single years. For example, trees form annual rings, and some of the Big Trees of California take us back several thousand years. When ice-sheets cover part of a country, the flood produced by each summer’s melting lays down a layer of clayey deposit over the neighbouring regions; and Baron de Geer has shown, by counting these layers in Southern Sweden, that the retreat of the last ice-sheet in Scandinavia took over 10,000 years. But these more detailed methods commonly apply only to the immediate past, and to periods of time that are but one or two ticks of geology’s clock.

There is also a further method applicable to any period where marine shells were fossilized. It is based on the fact that many bivalve shells show annual growthrings like trees, and that the annual growthrings are divided up into minor rings, each new addition to the shell being made after a feeding period. These minor rings vary much in breadth, according to outer conditions. Thus each annual growth-ring bears its own individual stamp in the number and size of its separate feedingrings ; and the same stamp will be impressed on all the shells in the same neighbourhood. By comparing a whole series of shells in the successive layers from one locality we can identify the separate years by the pattern they have left on the growth-rings, and so count their succession. This method is being worked out by Prof. W. M. Winton of Texas; it should give valuable results whenever thick layers of one type have been laid down in one place, and contain the shells of bivalves.

Ultimately, by such means as these, we may carry a surprisingly detailed calendar back for hundreds of thousands and even millions of years into the past from the