How Do We Know that Fossils are old?
Entry by: Richard Dawkins
When a living creature dies, it usually decays and is lost to history. Very occasionally, however, rocks take up some kind of permanent imprint of the body and preserve it for us to see even after millions of years. This is a fossil. Often a fossil retains only the external shape of the body. Better quality fossils are formed when minerals dissolved in water seep in and replace the molecules of the corpse. When this happens, the crystal structure of the rock retains details of the internal structure of the animal. That's how fossils are formed, but how do we know how old they are?
The best way to date fossils is by measuring radioactive isotopes in surrounding rocks. Elements like potassium, uranium, lead and so on come in more than one version, depending upon how many neutrons they have in their atomic nucleus. These different versions are called isotopes. Some isotopes are radioactive, meaning that they decay, at a fixed and known rate, into a completely different element. Physicists understand why this happens, and they know the rate at which each isotope decays. So radioactive isotopes act as miniature clocks, thousands of tiny clocks embedded through the substance of a rock.
The rate of decay - the rate at which the clock ticks - is measured as the half-life, which is the time it takes for the quantity of isotope to be halved. Potassium 40, for example, decays to argon 40 with a half-life of 1.3 billion years. The half-life of strontium-90 is 28 years. You might wonder why we speak of "half-life" not just "life". If you have 100 grams of strontium-90, after 28 years it will be halved to 50 grams. So, won't it all have decayed after 56 years? No, no, absolutely no! It takes 28 years for the first half of the strontium-90 to decay. Then it takes another 28 years for half of what was left to decay. So, after 56 years, there will still be a quarter of the original strontium-90 left: 25 grams. And after another 28 years there will be half as much again, and so on. So you can see that there never comes a definite time when all the strontium-90 is gone. It just sort of dies away forever, although there does come a time when, for practical purposes, you can say that it is all gone. Strontium-90 decays much too fast to be useful for dating fossils. But the half-life of potassium-40 is 1.3 billion years, and that we can use. Plus lots of other isotopes with a great range of half-lives.
We can measure how much potassium-40, say, there is in a piece of rock. But that is not enough to date the rock, because we don't know how much potassium-40 there was when the rock first formed. We solve this problem by also measuring the content of argon-40, the decay product. The ratio of the two tells us how much time has elapsed since the process of radioactive decay started. But that still is not enough information. It tells us the age of a rock only if all the clocks in the rock were 'zeroed' at the same time. This is true of volcanic rocks like granite. What does it mean to 'zero' a radioactive clock?
When molten lava solidifies to make granite, little pockets of radioactive isotopes, for example potassium-40 are trapped, with none of their decay products (no argon-40 in our example). The same is true of all the pockets in any particular chunk of granite, and it is true of all the different radioactive elements. All the clocks are zeroed at the moment when the rock was formed. So by measuring the ratio between the amount of radioactive isotope and the amount of its decay product (for example the ratio of potassium-40 to argon-40) and knowing the half-life (from laboratory measurements) you can calculate how many millions of years have elapsed since the rock was formed.
Unfortunately, fossils are not found in volcanic rock - molten lava is not conducive to preserving an animal's shape! Fossils are normally found only in sedimentary rocks - hardened mud, silt or sand, in which the corpse can lie peacefully for long enough to permit the fossilisation process to get going. There are radioactive clocks in sedimentary rocks. The particles of mud and sand that settle to make the sediments were earlier ground down from volcanic rocks, and they were zeroed at some point in history. But the grains and particles that compacted together to make a particular sediment of sandstone or limestone are all of different ages, 'zeroed' at different times - perhaps they were ground down from volcanic rocks of different ages. So we can't use the radioactive clocks in the fossils themselves, nor in the sedimentary rocks in which they lie. We have to use volcanic rocks that happen to lie in the vicinity. For example, if a fossil is sandwiched between a piece of granite that is 100 million years old and another piece of granite that is 95 million years old, we know its age is between those two.
It is through these methods that we know how old the Earth is (about 4.6 billion years) we know when famous fossils such as the ape woman "Lucy" lived (about 3.2 million years ago). We know when the dinosaurs suffered their catastrophic mass extinction (65 million years ago). The transitional fish Tiktaalik, representative of our fish ancestors that first came out of water onto land, lived about 375 million years ago.
Since the different isotopes have different half-lives, they are good for different ranges of ages. Carbon dating is good for the archeological timescale, potassium- 40 dating is good for a broad range of ancient fossils, other isotopes are most sensitive for other age ranges, but there is plenty of overlap between their ranges, and it is very satisfying that they agree - within expected margins of error - with each other when they are used to date the same specimens.