What Mad Pursuit

by Francis Crick

  This compelling argument was shattered by Charles Darwin, who believed that the appearance of design is due to the process of natural selection. This idea was put forward both by Darwin and by Alfred Wallace, essentially independently. Their two papers were read before the Linnean Society on July 1, 1858, but did not immediately produce much reaction. In fact, the president of the society, in his annual review, remarked that the year that had passed had not been marked by any striking discoveries. Darwin wrote up a “short” version of his ideas (he had planned a much longer work) as The Origin of Species. When this was published in 1859, it immediately ran through several reprintings and did indeed produce a sensation. As well it might, because it is plain today that it outlined the essential feature of the “Secret of Life.” It needed only the discovery of genetics, originally made by Gregor Mendel in the 1860s, and, in this century, of the molecular basis of genetics, for the secret to stand before us in all its naked glory. It is all the more astonishing that today the majority of human beings are not aware of all this. Of those who are aware of it, many feel (with Ronald Reagan) that there must be a catch in it somewhere. A surprising number of highly educated people are indifferent to these discoveries, and in western society a rather vocal minority are actively hostile to evolutionary ideas.

  To return to natural selection. Perhaps the first point to grasp is that a complex creature, or even a complex part of a creature, such as the eye, did not arise in one evolutionary step. Rather it evolved through a series of small steps. Exactly what is meant by small is not necessarily obvious since the growth of an organism is controlled by an elaborate program, written in its genes. Sometimes a small change in a key part of the program can make a rather large difference. For example, an alteration in one particular gene in Drosophila can produce a fruitfly with legs in the place of its antennae.

  Each small step is caused by a random alteration in the genetic instructions. Many of these random alterations may do the organism no good (some may even kill it before it is born), but occasionally a particular chance alteration may give that particular organism a selective advantage. This means that in the last analysis the organism will, on average, leave more offspring than it would otherwise do. If this advantage persists in its descendents then this beneficial mutant will gradually, over many generations, spread through the population. In favorable cases every individual will come to possess the improved version of the gene. The older version will have been eliminated. Natural selection is thus a beautiful mechanism for turning rare events (strictly, favorable rare events) into common ones.

  We now know—it was first pointed out by R. A. Fisher—that for this mechanism to work inheritance must be “particulate,” as first shown by Mendel, and not “blending.” In blending inheritance the properties of an offspring are a simple blend of those of its parents. In particulate inheritance the genes, which are what is inherited, are particles and do not blend. It turns out that this makes a crucial difference.

  For example, in blending inheritance a black animal mated with a white animal would always produce offspring whose color was a blend of black and white, that is, some shade of gray. And their offspring, if they bred together, would always remain gray. In particulate inheritance various things can happen. For example, it could be that all the first-generation animals were indeed gray. If these were now mated together, we would obtain in the second generation, on average, one-quarter black animals, one-half gray animals, and one-quarter white. [This assumes that color is, in this case, a simple Mendelian character, without dominance.] The genes, being particulate, do not blend, even if their effects, in a single animal, blended, so that one white particle (gene) and one black particle, acting together in the same creature, produced a gray animal. This particulate inheritance preserves variation (we have mixed black, gray, and white animals after two generations, not just gray ones), whereas blending inheritance reduces variation. If inheritance were blending, the offspring of a black animal and a white animal mate, would produce gray animals indefinitely. This is obviously not the case. The fact can be seen clearly in humans: people do not become more and more alike as the generations go on. Variation is preserved.

  Darwin, who was a deeply honest man and always faced up to intellectual difficulties, did not know about particulate inheritance and was consequently very disturbed by the criticisms of a Scottish engineer, Fleeming Jenkin. Jenkin pointed out that inheritance (which, without realizing it, Darwin assumed to be blending) would not allow natural selection to work effectively. As particulate inheritance had not yet been thought of, this was a very damning criticism.

  What, then, are the basic requirements for natural selection to work? We obviously need something that can carry “information”—that is, the instructions. The most important requirement is that we should have a process for exact replication of this information. It is almost certain that, in any process, some mistakes will be made, but they should occur only rarely, especially if the entity to be replicated carries a lot of information. [In the case of DNA or RNA, the rate of making mistakes, per effective base pair, per generation must, in simple cases, be rather less than the reciprocal of the number of effective base pairs.]

  The second requirement is that replication should produce entities that can themselves be copied by the replication process or processes. Replication should not merely be like that of a printing press, when master plates make many copies of a newspaper but each newspaper cannot, by itself, produce further copies of either the press or the newspaper. [In technical terms, replication should be geometrical, not merely arithmetical.]

  The third requirement is that mistakes—mutations—should themselves be capable of being copied, so that useful variation can be preserved by natural selection.

  There is a final requirement that the instructions and their products should stay together [cross-feeding is to be avoided]. A useful trick is to use a bag—a cell, that is—to do this, but I will not dwell on this point.

  In addition, the information needs to do something useful, or to produce other things that will do useful jobs for it, to help it to survive and to produce fertile offspring with a good chance of survival.

  In addition to all this, the organism needs sources of raw material (since it has to produce copies of itself), the ability to get rid of waste products, and some sort of source of energy [Free Energy]. All these features are required, but the heart of the matter is obviously the process of exact replication.

  This is not the place to explain Mendelian genetics in all its technical details. However, I shall try to provide a glimpse of the astonishing results that a simple mechanism like natural selection can produce over long periods of time. A fuller and very readable account can be found in the early chapters of Richard Dawkins’s recent book, The Blind Watchmaker. One may wonder at the title of the book. Watchmaker obviously refers to the designer that Paley invoked to explain the imaginary watch found on the heath. But why “blind”? I cannot do better than quote Dawkins’s actual words:

  All appearances to the contrary, the only watchmaker in nature is the blind forces of physics, albeit deployed in a very special way. A true watchmaker has foresight: he designs his cogs and springs, and plans their interconnections, with a future purpose in his mind’s eye. Natural selection, the blind, unconscious, automatic process which Darwin discovered, and which we now know is the explanation for the existence and apparently purposeful form of all life, has no purpose in mind. It has no mind and no mind’s eye. It does not plan for the future. It has no vision, no foresight, no sight at all. If it can be said to play the role of watchmaker in nature, it is the blind watchmaker.

  Dawkins gives a very pretty example to refute the idea that natural selection could not produce the complexity we see all around us in nature. The example is a very simple one, but it drives the point home. He considers a short sentence (taken from Hamlet):

  METHINKS IT IS LIKE A WEASEL.

  He first calculates how exceedingly improbable
it is that anyone, typing at random (traditionally a monkey, but in his case his eleven-month-old daughter or a suitable computer program) would by chance hit on this exact sentence, with all the letters in their correct place. [The odds turn out to be about 1 in 1040.] He calls this process “single-step selection.”

  He next tries a different approach, which he calls “cumulative selection.” The computer chooses a random sequence of twenty-eight letters. It then makes several copies of this but with a certain chance of making random mistakes in the copying. It next proceeds to select the copy that most resembles the target sentence, however slightly. Using this slightly improved version, it then repeats this process of replication (with mutation) followed by selection. In the book Dawkins gives examples of some of the intermediate stages. In one case, after thirty steps, it had produced:

  METHINKS IT IS LIKE A WEASEL

  and after forty-three steps it had the sentence completely correct. How many steps it takes to do this is partly a matter of chance. In other trials it took sixty-four steps, forty-one steps, and so forth. The point is that by cumulative selection one can reach the target in a relatively small number of steps, whereas in single-step selection it would take forever.

  The example is obviously oversimple, so Dawkins tried a more complex one, in which the computer grew “trees” (organisms) according to certain recursive rules (genes). The results are too complex to reproduce here. Dawkins says: “Nothing in my biologist’s intuition, nothing in my 20 years’ experience of programming computers, and nothing in my wildest dreams, prepared me for what actually emerged on the screen” (p. 59).

  If you doubt the power of natural selection I urge you, to save your soul, to read Dawkins’s book. I think you will find it a revelation. Dawkins gives a nice argument to show how far the process of evolution can go in the time available to it. He points out that man, by selection, has produced an enormous variety of types of dog, such as Pekinese, bulldogs, and so on, in the space of only a few thousand years. Here “man” is the important factor in the environment, and it is his peculiar tastes that have produced (by selective breeding, not by “design”) the freaks of nature we see preserved all around us as domestic dogs. Yet the time required to do this, on the evolutionary scale of hundreds of millions of years, is extraordinarily short. So we should not be surprised at the ever greater variety of creatures that natural selection has produced on this much larger time scale.

  Incidentally, Dawkins’s book contains a fair but devastating critique (pages 37-41) of the book The Probability of God by Hugh Montefiore, the Bishop of Birmingham. I first knew Hugh when he was Dean of Caius College, Cambridge, and I agree with Dawkins that Hugh’s book “… is a sincere and honest attempt, by a reputable and educated writer, to bring natural theology up to date.” I also agree wholeheartedly with Dawkins’s criticism of it.

  At this point I must pause and ask why exactly it is that so many people find natural selection so hard to accept. Part of the difficulty is that the process is very slow, by our everyday standards, and so we rarely have any direct experience of it operating. Perhaps the type of computer game Richard Dawkins describes might help some people to see the power of the mechanism, but not everyone likes to play with computers. Another difficulty is the striking contrast between the highly organized and intricate results of the process—all the living organisms we see around us—and the randomness at the heart of it. But this contrast is misleading since the process itself is far from random, because of the selective pressure of the environment. I suspect that some people also dislike the idea that natural selection has no foresight. The process itself, in effect, does not know where to go. It is the “environment” that provides the direction, and over the long run its effects are largely unpredictable in detail. Yet organisms appear as if they had been designed to perform in an astonishingly efficient way, and the human mind therefore finds it hard to accept that there need be no Designer to achieve this. The statistical aspects of the process and the vast numbers of possible organisms, far too many for all but a tiny fraction of them to have existed at all, are hard to grasp. But the process clearly works. All the worries and criticisms just listed have no content when examined carefully, provided the process is understood properly. And we have examples, both from the laboratory and the field, of natural selection in action, from the molecular level to the level of organisms and populations.

  I think there are two fair criticisms of natural selection. The first is that we cannot as yet calculate, from first principles, the rate of natural selection, except in a very approximate way, though this may become a little easier when we understand in more detail how organisms develop. It is, after all, rather odd that we worry so much how organisms evolved (a process difficult to study, since it happened in the past and is inherently unpredictable), when we still don’t know exactly how they work today. Embryology is much easier to study than evolution. The more logical strategy would be to find out first, in considerable detail, how organisms develop and how they work, and only then to worry how they evolved. Yet evolution is so fascinating a subject that we cannot resist the temptation to try to explain it now, even though our knowledge of embryology is still very incomplete.

  The second criticism is that we may not yet know all the gadgetry that has been evolved to make natural selection work more efficiently. There may still be surprises for us in the tricks that are used to make for smoother and more rapid evolution. Sex is probably an example of such a mechanism, and there may, for all we know, be others as yet undiscovered. Selfish DNA—the large amounts of DNA in our chromosomes with no obvious function—may turn out to be part of another (see page 147). It is entirely possible that this selfish DNA may play an essential role in the rapid evolution of some of the complex genetic control mechanisms essential for higher organisms.

  But leaving these reservations aside, the process is powerful, versatile, and very important. It is astonishing that in our modern culture so few people really understand it.

  You could well accept all those arguments about evolution, natural selection, and genes, together with the idea that genes are units of instruction in an elaborate program that both forms the organism from the fertilized egg and helps control much of its later behavior. Yet you might still be puzzled. How, you might ask, can the genes be so clever? What could genes possibly do that would allow the construction of all the very elaborate and beautifully controlled parts of living things?

  To answer this we must first grasp what level of size we are talking about. How big is a gene? At the time I started in biology—the late 1940s—there was already some rather indirect evidence suggesting that a single gene was perhaps no bigger than a very large molecule—that is, a macromolecule. Curiously enough, a simple, suggestive argument based on common knowledge also points in this direction.

  Genetics tells us that, roughly speaking, we get half of all our genes from our mother, in the egg, and the other half from our father, in the sperm. Now, the head of a human sperm, which contains these genes, is quite small. A single sperm is far too tiny to be seen clearly by the naked eye, though it can be observed fairly easily using a high-powered microscope. Yet in this small space must be housed an almost complete set of instructions for building an entire human being (the egg providing a duplicate set). Working through the figures, the conclusion is inescapable that a gene must be, by everyday standards, very, very small, about the size of a very large chemical molecule. This alone does not tell us what a gene does, but it does hint that it might be sensible to look first at the chemistry of macromolecules.

  It was also known at that time that each chemical reaction in the cell was catalyzed by a special type of large molecule. Such molecules were called enzymes. Enzymes are the machine tools of the living cell. They were first discovered in 1897 by Eduard Buchner, who received a Nobel Prize ten years later for his discovery. In the course of his experiments, he crushed yeast cells in a hydraulic press and obtained a rich mixture of yeast juices. He w
ondered whether such fragments of a living cell could carry out any of its chemical reactions, since at that time most people thought that the cell must be intact for such reactions to occur. Because he wanted to preserve the juice, he adopted a stratagem used in the kitchen: he added a lot of sugar. To his astonishment, the juice fermented the sugar solution! Thus were enzymes discovered. (The word enzyme means “in yeast.”) It was soon found that enzymes could be obtained from many other types of cell, including our own, and that each cell contained very many distinct kinds of enzymes. Even a simple bacterial cell may contain more than a thousand different types of enzymes. There may be hundreds or thousands of molecules of any one type.

  In favorable circumstances an enzyme could be purified away from all the others and its action studied by itself in solution. Such studies showed that each enzyme was very specific, and catalyzed only one particular chemical reaction or, at most, a few related ones. Without that particular enzyme the chemical reaction, under the mild conditions of temperature and acidity usually found in living cells, would proceed only very, very slowly. Add the enzyme and the reaction goes at a good pace. If you make a well-dispersed solution of starch in water, very little will happen. Spit into it and the enzyme amylase in your saliva will start to digest the starch and release sugars.

  The next major discovery was that each of the enzymes studied was a macromolecule and that they all belonged to the same family of macromolecules called proteins. The key discovery was made in 1926 by a one-armed American chemist called James Sumner. It is not all that easy to do chemistry when you have only one arm (he had lost the other in a shooting accident when he was a boy) but Sumner, who was a very determined man, decided he would nevertheless demonstrate that enzymes were proteins. Though he showed that one particular enzyme, urease, was a protein and obtained crystals of it, his results were not immediately accepted. In fact, a group of German workers hotly contested the idea, which somewhat embittered Sumner, but it turned out that he was correct. In 1946 he was awarded part of the Nobel Prize in Chemistry for his discovery. Though very recently a few significant exceptions to this rule have turned up, it is still true that almost all enzymes are proteins.