JBS Haldane

A Dialectical Account of Evolution

Source: Science & Society, Volume I, Number 4, Summer, 1937;
Transcribed: for marxists.org in May, 2002.

THIS paper requires a preliminary apology. It is written in Spain, and as the writer is taking a very minor part in the defence of Madrid he is unable to consult many works of reference. Readers to whom practice is as important as theory will excuse any inadequacy in citations.

Evolution is pretty generally accepted as an historical fact. But some biologists and many popularizers of biology believe that Darwin's account of how and why it happened is incorrect. This is of course true in one sense. Darwin was not infallible. But because we have had to modify Dalton's ideas about atoms we do not say that he was wrong. We say that he was not completely right. And as no one (except His Holiness the Pope, speaking ex cathedra on a matter of faith or morals) is always completely right, this need not trouble us.

The difficulties which biologists encounter in explaining evolution are of two kinds. One arises from the time scale. If a paleontologist finds any noticeable difference between fossils in strata of which one was laid down 100,000 years after the other, he says that the animal in question was evolving rather quickly. It is evident therefore that the change produced in the laboratory in the course of ten years may be a very crude model of evolutionary change, and may be actually misleading. It is also probable, in fact fairly certain, that our account is incomplete because some processes are occurring too slowly to detect.

The other difficulty is more serious. The discoveries of different workers seem to contradict each other flatly. And here a dialectical approach is essential. "Development is struggle of opposites. Two historically observed conceptions of development are: (I) development as decrease and increase, as repetition; and (2) development as a unity of opposites and their reciprocal correlation. The first conception is dead, poor and dry; the second is vital. It is only this second conception which offers the key to understanding the self-movement of everything in existence. .. ." [1]

If we adopt this point of view, our approach will differ from Darwin's. Darwin was at great pains to stress the variability of plant and animal species. We shall look for populations showing as little variation as possible, because we want to see how variation arises. Such populations were first obtained, studied and described by Darwin's contemporary, de Vilmorin, in the course of agricultural practice. Indeed, de Vilmorin was so busy turning out seeds, from which he made a considerable fortune, that his theoretical account of his work was not very clear, and passed almost unnoticed.

Let us take a wheat plant whose ancestors have been self-fertilized for twenty generations back, or a pair of mice which are brother and sister, and whose ancestors have been brother and sister for fifty generations. We find that their progeny are very much alike, and that what differences there are are not inherited. The more fertile wheat plants do not yield more fertile progeny than the less fertile. The lighter-colored mice have no lighter children than their darker brothers and sisters.

This is at first sight very paradoxical; nevertheless it agrees quite well with the account of an organism given by those physiologists who reject vitalism and mechanism alike, for example the late J. S. Haldane. The living organism is a union of opposites, such as anabolism and catabolism. It reacts to changes in the environment so as to bring itself back to normal. It eats and drinks to restore its material condition after doing work; it develops towards a definite form even if its normal course of development is upset by external interference. It heals wounds, recovers from infections and so on. A sufficiently gross interference kills it, but if it is not killed it struggles back towards the normal structure and function.

It is clear that a group of organisms with these properties could never evolve. At most, in different environments, they could give rise to local races, which however would not differ innately, but only as the result of the environments in which they lived. And since the differences between members of a pure line are not inherited, selection will not work within it. This has been repeatedly proved for artificial selection, and would no doubt be true for natural selection also. So when Johannsen, following up de Vilmorin, gave a full account of pure lines, a few biologists thought that he had disproved Darwinism, while others (e.g., Lotsy) took the view that heritable variation could only arise from hybridization. The facts are otherwise. The pure line contains its own internal contradiction, though this develops so slowly that a farmer, who has to think only ten years ahead, can generally neglect it — though a paleontologist, who thinks in millions of years, cannot do so.

To understand why a pure line is not eternal, we must examine the intimate structure of the cell and see why a completely self-regulating, self-reproducing organism is a fiction, though quite a satisfactory one for many purposes. A cell contains a great many different kinds of organic molecules, for example sugars, oils and proteins. Most of these are put together on the spot; only a few of the simpler ones are found in the food or transported from other organs. We are only just beginning to find out how the large molecules are put together. But generally this seems to be a very complicated affair. Each molecule is fashioned by the action of a great many other molecules, of the kind called "enzymes," each of which controls a particular chemical reaction — just as a machine is made by a number of other machines, few if any being like itself. Thus A is built out of food materials by B, C, D, E, F, etc. And if A is an enzyme it may take part, along with D, G, H, I, etc., in the synthesis of B. Thus if A is slightly altered to form A', this will cause some alteration in the cell's life; it may even kill it. But it will not produce a lot of new molecules of the type A'. If in a machine tool factory which uses its own products you alter the setting of a certain lathe, it will produce faulty tools, but it will not reproduce lathes with this particular fault.

There are however a certain number of exceptional molecules which, given the right chemical environment, reproduce themselves. Stanley of the Rockefeller Institute in New York isolated the first of them in 1935. It is a protein, probably a nucleoprotein, and shows no signs of life under ordinary conditions. But if you inject it into a tobacco plant it reproduces itself in such quantity as to kill the plant. It behaves like a disease germ. The fact that biologists dispute whether or not this and other viruses are alive is the best proof that the gap between living and non-living things is not absolute. The fact that it is a protein is a startling confirmation of Engels' statement in 1878 that "Life is the mode of existence of albuminous substances" [2] (or as we should now write, of the proteins, though actually only of some of them).

Now in the normal cell there are several thousands of different molecules with the property of self-reproduction. They are called "genes." We know that they are self-reproducing for the following reason: If a cell is subjected to X-rays, various molecules in it are damaged, and a big enough dose is fatal. After some cell-generations, recovery may be complete. But if there has been a permanent change it is generally found that a gene has been altered. That is to say that the gene A has been turned into A', and is reproduced in future as A'. The genes are too small to be seen even with a microscope, except perhaps in a few cases, but they are arranged in visible structures called "chromosomes" in the nucleus of the cell. The chromosomes, being composed of genes, share the property of the genes of self-reproduction. By means of X-rays the chromosomes can be broken and rearranged, and the new arrangement, provided it is compatible with cell life and regular nuclear division, is perpetuated indefinitely. If anything of the kind happened outside the nucleus we should be able to produce a group of characters inherited through females only, since the male parent supplies half the nucleus of the cell and little else, while the great bulk of the rest comes from the mother. But few such characters are known, and none have been produced by X-rays.

Now under normal conditions genes do change, and chromosomes do rearrange themselves, though much more slowly than under X-ray bombardment. For example, in one of the human chromosomes there is a gene whose function is to make the blood clot rapidly. Once in about 50,000 generations this gene changes, or mutates, so that it can no longer perform its function, and a man carrying such a changed gene becomes a haemophiliac, whose blood will not clot.

This process of natural mutation is strictly accidental. It is apparently due to too great a concentration of energy at a particular point, which alters a gene or chromosome. An accident does not mean an event with no cause. It means an event outside our control, and in this case outside the control which a living organism can exercise over its constituents. If one of Franco's "Made in Germany" bombs disintegrates me rather than some other comrade before I have finished writing this, it will be an accident, since the Fascist aviators are in much too great a hurry to escape the Loyalist pursuit planes to aim with any great care. Similarly if one gene rather than another mutates, that is an accident. Marx was very clear as to the importance of accidents. "World history would, indeed, be very easy to make, if the struggle was taken up only on condition of infallibly favorable chances. It would, on the other hand, be of a very mystical nature if 'accidents' played no part. These accidents themselves fall naturally into the general course of development and are compensated again by other accidents." [3]

Evolution is not "of a very mystical nature." It depends on accidents. In numerous species these accidents happen often enough to give rise to statistical certainty. If the gene for haemophilia arises afresh on an average once in 50,000 generations, it is very nearly certain that it will arise between 9000 and 11,000 times in the next 500 million people born. On the other hand, with a rare species such as the Indian elephant, comprising perhaps only 20,000 individuals, chance assumes a great importance.

The accidental character of mutation is clear in many other ways. Almost, though not quite, all mutations lower the fitness of an organism in its natural state. This is equivalent to saying that organisms are pretty well fitted to their environment (fitness is defined later) and any change due to chance is likely to be for the worse. If mutation were an adaptive phenomenon like the growth of a muscle when exercised, as Lamarck believed, this would not be so. Most mutations would be useful. The same would (I suppose) be true if mutation were a manifestation of the Life Force (whatever he, she or it may be). Naturally enough, biologists to whom dialectical materialism means nothing, or means a weapon of the abominable Marx, cannot understand how harmful mutations can be a condition of evolutionary progress. They therefore deny them any importance.

We have now taken our first step. The self-repairing, self-reproducing organism is negated by accidents of a certain type. It can no longer reproduce itself unchanged. But since it does reproduce itself in the changed form (say as a white mouse in place of a brown, or a beardless wheat in place of a bearded) the negation is negated. This dialectical process is called "mutation," and leads to inheritable variations within a species. If we do not look at it dialectically, we are apt to label it either as pathological or progressive. In fact it constitutes a union of these opposites.

Taking mutation as a given process, what do we expect from it? Certainly not evolution. Different mutations will affect different organs in different directions. The average man can get his best idea of the effects of mutation by going to a show of fancy poultry breeds. As a result of mutation the feathers can be any color of a large range from black to white. They can be longer or shorter than in the wild form. The comb may be doubled, as in a Sicilian Buttercup, or reduced to a mere button, as in a Malay. The number of toes may be increased, the rooster may have feathers like a hen, and so on. These changes are due to new genes which have arisen by mutation. Other genes produce still fiercer effects, for example young chickens with mere knobs for limbs, which die long before hatching.

Mutation alone, then, would cause every species to break down into a collection of freaks, some of which could only be preserved alive by a miracle. We have every reason to ask whether it is really of evolutionary importance. The answer is decisively "yes." Where two species can be crossed, and the hybrids of at least one sex are fertile enough to breed from, we can analyze the interspecific differences genetically. In such a case, it is generally found that a good deal of the difference between the two species is due to one or two genes, and that much of the remaining difference behaves statistically, as if it were due to a number of genes each having a small effect by itself. In some cases, and particularly where the species are so far apart that they cannot be crossed, or yield a sterile hybrid, we find differences of a higher order than genes — that is to say, differences in the arrangement of the chromosomes. As genes of new kinds and chromosomal rearrangements only arise by mutation, we cannot avoid attributing to it a certain evolutionary importance. The only question is whether it accounts for all or only some of the differences between species.

The antithesis to mutation, which nearly negates its effects, is natural selection. We can best understand it by studying the effects of artificial selection. Suppose that we start with a group of wheat plants or flies, and in each generation breed two hundred individuals and select the ten best, in some respect, as parents of the next generation. For the first five or ten generations we shall make dramatic progress. We may increase the yield of our wheat plants, on the average, by 30 per cent. We may nearly double the number of bristles on our flies. Darwin knew this. He did not know that after ten or twenty generations the process comes to an end. We started with a population containing several different genotypes — that is to say, sets of genes. We end by eliminating all but the one which best satisfies our criterion. We have got to a pure line, and further selection is useless.

When this fact was discovered it was at once stated that Darwinism was dead. But that cat has at least nine lives. The critics forgot the transformation of quantity into quality. In a large enough population, selection never gives us a pure line, because new genes are always turning up by mutation before selection has eliminated all the original heterogeneity. Hence an experiment on a small population is misleading as a model for the natural process.

In artificial selection we can select for anything we please — for example, for innate capacity to sing, to produce many eggs, to grow long hair, or to develop cancer, though unless the suitable genes are available selection is fruitless. But natural selection selects for one character only, which Darwin called fitness, but never defined rigorously. It can be defined only in mathematical terms, and R. A. Fisher and the writer were the first to do so. The full definition is complicated, but a simple example will serve. Consider a population of self-fertilized plants, say wild peas, which die down every winter and reproduce themselves annually from seed. Suppose several different genotypes, say purple, red and white, with and without tendrils. Now take one hundred plants of each sort in a particular environment, counted at a particular point in the life cycle, say the opening of the first flower, and see how many progeny they leave, on the average, on the same day or at the same point in their life cycle of next year. This average number is the fitness. In the long run the average fitness of any species is quite close to unity, except during rare periods of rapid spread or extinction.

A change in any quality may affect the fitness. Extra fertility will do so, other things being equal. But other things rarely are. A wild hen which produced two hundred eggs per year would certainly not be able to bring up two hundred chickens. She would be worse off than the mother of a mere ten, because she would be turning an immense amount of food into eggs which would never be hatched. Darwin laid a considerable stress on survival, rather than fertility, but a change in either will affect the fitness.

It is important to note that fitness must be averaged over a great many generations. The animals of a species may have a migratory habit which drives millions of them to destruction every year. But if as a result of this habit a single pregnant female reaches a land which would otherwise have been inaccessible, her progeny may colonize a whole continent. If the habit is hereditary, they will retain it for many generations, although it is disadvantageous to the vast majority of individuals. Before we condemn a character as disadvantageous we must consider its value, not merely on normal occasions, but during a catastrophe which wipes out most members of a species, such as a forest fire, an epidemic or a very abnormal summer or winter. And we must ask whether it may not be useful in some dangerous enterprise, such as crossing an ocean.

Unfortunately the word fitness has been used in many senses. For example, we can speak of fitness for Rugby football, the Civil Service or the gallows. It is sometimes assumed that fitness in some such respect implies fitness in the Darwinian sense. If it is true that in our existing English society the very stupid have the largest families, they are the fittest, since infantile mortalities do not differ enough between different groups to compensate for differences of birthrate. If our society is such that people with congenital mental defects are fitter than the average, that is a fault in our society. It is ridiculous to call these people unfit. They might be so in a different environment.

Above all, the truly ludicrous assumption is made that certain innate qualities enable their possessors to become richer under capitalism, that these qualities are strongly inherited, and that they have a relation with biological fitness. Insofar as rise in the economic scale depends on inherited factors, then these factors make for lessened fitness under capitalism, whatever they may do under some other system. For in most if not all capitalist countries the rich have fewer surviving children than the poor, and are therefore less fit. Economic success and biological success are entirely opposed to one another. This is one of the many contradictions of capitalism.

However, the duration of an economic system is measured in centuries, and covers a negligible period in the life of a species. Let us consider natural selection as we can actually observe it in a species which is not, like man, engaged in changing its own environment; or let us observe it in man, confining our attention to characters whose effect on fitness is invariable. In every species so far observed freaks turn up as the result of mutation, and the abnormal genes responsible are eliminated, in the long run, at exactly the same rate as they are reproduced. If abnormal genes are dominant, or if they are sex-linked, they are eliminated very quickly. But if they are recessive genes located on ordinary chromosomes, they are not eliminated until two of the same kind happen to come together in a single individual. Thus in man (or any other mammal) albinism is recessive. A normal person may carry one gene for albinism, but an albino must receive such a gene from both parents. In consequence there is an immense reserve of variability in a population, due to recessive genes which are harmful in the existing environment, but not necessarily so in a different one — a fact first proved experimentally by Tschetverikoff.

Just as mutation negates the fixity of a species, natural selection, to a first approximation, negates the negation, and we are apparently back where we were before, at a uniform population. To put the matter in different words, evolution, as Sewall Wright first realized, is a second order effect, due to the fact that two processes, which at first sight are in equilibrium, do not quite balance. This is an entirely normal situation in science. Consider such a process as mountain building. The weight of the rocks is very nearly balanced by the pressure of the earth's interior. The forces which build mountains may be small disturbances of this balance, or lateral thrusts. But they are too small to measure directly in many cases. Or consider the evolution of the solar system. To a first approximation the planets revolve in elliptical orbits, gravitation being negated by centrifugal "force." These orbits evolve, though it is far from certain how they evolve. If the negation were complete they would not evolve at all. Even the growth of a man or animal is due to the fact that anabolism and catabolism do not quite balance, though here the balance is far from exact. In man about 1 to 2 percent of the energy of the food eaten in the first fourteen years is used for growth, whereas in the evolutionary processes measured in millions of years the balance is very close indeed, and the negation almost completely negated.

Let us now consider the species in dynamic equilibrium between mutation and selection. In such a species a number of genes will be found to be fairly common which are slightly disadvantageous in the particular environment considered, but may be advantageous if the environment is changed. Thus Timofeeff-Ressovsky found that in Drosophila melanogaster white eye color (associated with lack of pigment in internal organs) diminished the life span except at very high temperatures. But at high temperatures it was actually advantageous. Dynamic equilibrium ensures that there will be a reserve of genes of this kind, so that a species may be expected to change fairly quickly in a new environment.

Darwin thought that evolution went on slowly and steadily, avoiding violent jumps. In this he was almost unquestionably influenced by the economic situation of his time. In an era of expanding markets it was natural to believe in steady progress, that

Freedom slowly broadens down
From precedent to precedent

and so on. This facile optimism is now replaced among intellectuals by the equally facile despair of a Spengler or a Huxley, on the one hand, or the grimmer optimism of Marxists[4] on the other.

It is possible that most mutations of evolutionary importance are very slight. The white polar bear may have arisen from brown or black ancestors as the result of a single mutation. Or it may have arisen by a series giving gray or yellow fur as an intermediate. Such small mutations would cause slighter upsets in metabolism than one large one, and these could be compensated for by the selection of modifying genes until the species was ready for the next step.

But changes in chromosome number must be abrupt and, so to say, revolutionary. An organism may have twenty-three or twenty-four pairs of chromosomes. It cannot have 23.317 pairs or any other intermediate number. In plants a doubling of the chromosome number is by no means uncommon, and has often led to the formation of a new species. This doubling commonly takes place as the result of hybridization. Thus the species of hemp-nettle Galeopsis pubescens and G. speciosa each have two sets of eight chromosomes. Their hybrid is very sterile, but from it Muntzing obtained a plant with two sets of chromosomes from each parent. As like chromosomes could pair, it was quite fertile, and closely resembled the species Galeopsis Tetrahit, with which it crossed readily, though sterile with the parent species. In fact Muntzing had made this species, and there is no doubt that it originated by the same process. Doubling may also occur without hybridization.

The new species produced by doubling the chromosome number may be better adapted than the old to certain conditions. Often it is better adapted to cold, and has a wider range northwards. Now consider a group of plants on an island where the climate is gradually getting colder. Every few years a plant with a doubled chromosome number is produced. It is less fit than its neighbors, and its progeny soon die out. But as the climate gets colder a revolutionary situation arises. The fitness of the two types becomes equal, and they struggle inconclusively. A further decrease of temperature ensures the victory of the new type.

Still more drastic changes in a plant are produced when only some of the chromosomes are doubled, so that instead of AABBCCDDEEFF we have AAAABBBBCCDDEEFF. In such cases also, evolution is not gradual.

Another type of revolutionary situation arises as a result of gene mutations. Suppose that two new genes arise frequently by mutation, each of which lowers fitness, while the two together raise it. Each alone will give worse sight. The two together will give better sight. (This is a hypothetical case; in the observed cases where two genes lower fitness alone and raise it together we do not understand how they interact.)

Now suppose each gene is recessive, and has a frequency of one in one hundred. The homozygous recessives will have a frequency of one in ten thousand, the double recessives will have a frequency of one in one hundred million. Even if it is very fit, it will be too rare to cause the genes to spread, in the face of the adverse selection exercised by the single recessives, which are ten thousand times as common. Nevertheless the population is potentially unstable.

The actual revolutionary situation may probably arise in two ways. A small population may be geographically isolated in which both genes are common enough to enable the double mutant to prevail. Or by a chromosomal rearrangement both the genes may be bound together so tightly that they act as a unit. In either case the new type will have a chance of ousting the old.

In fairness to Darwin we must admit that the fossil record is mainly one of slow change, anti that among those species which have been common enough to leave large numbers of fossils, these abrupt changes do not seem to have been very common. The probable reason for this will appear later.

Let us now consider the nature of fitness in greater detail. A given heritable change may have two effects (not of course mutually exclusive or even completely separable). It can make an organism fitter in relation to its environment of inorganic nature and other species. Or it can make it fitter in relation to its neighbors, the members of its own species. I have elsewhere called these changes increases in absolute and relative fitness. A few examples will make the matter clear.

Let us suppose that a plant develops a deeper root system. This will enable its descendants to live in drier soil, and if there is no loss in fitness in other respects, its descendants will cover a larger area. Now suppose that a plant in a cross-fertilized species which is normally fully fertilized produces more pollen, or pollen grains whose tubes grow quicker down the style. This plant will succeed in fertilizing a greater proportion of its neighbors than the average. Its characteristic, if heritable, will spread through the species. But as a result of this the species will not necessarily increase in numbers. It will be no fitter. It may be less fit, because an unnecessary amount of material is turned into pollen rather than seeds, roots, stems and leaves. A good deal of the evolution which has produced colored flowers has probably been due to selection for relative rather than absolute fitness. These structures are the supreme examples of publicity in pre-human organisms. By advertising their wares to the bees plants compete with members of their own species. But, like the advertisements which are so characteristic of capitalism, these probably represent a waste of material which could have been otherwise employed.

Or consider a species of fish. It may develop an adaptation which induces it to eat a hitherto untouched species of larval insect. Or it may develop the habit of eating the young of its own species. This may be a considerable source of fitness to the first few members practising it. In a population of a million fish the first few cannibals are very unlikely to eat their own young and thus extinguish the gene for cannibalism. The gene for cannibalism will always give its carriers a slight advantage in competition, and will tend to spread. But the final result will be to lower the fitness of the species. It is clear, however, that, in a small, isolated population a gene for cannibalism is much less likely to spread.

We see then that fitness may diminish as a result of the selection of the fittest. This result will however be found characteristically in successful species. So long as a species is rare, the struggle for life is a struggle with the environment. When it becomes common, it is largely a struggle with neighbors of the same species. This latter struggle will sometimes lead to the selection of valuable factors. But it will very often lower fitness. Further, in a crowded species the revolutionary situations arising from the isolation of small groups will be infrequent, and thus will tend to slow down the evolutionary process by preventing certain kinds of sudden change.

In fact a successful species tends to develop internal contradictions. It is now intelligible why the dominant groups of the past, such as the dinosaurs and the titanotheres, have left no descendants. They characteristically got bigger until they became extinct. Mere size is probably an advantage in the intraspecific struggle, especially the struggle for mates. It is not necessarily of any value in the struggle with the environment.

At this level the struggle between individuals becomes transformed into a struggle between species. And it is a struggle which places a premium on characters of a higher order, at least in animals. A species which does not indulge in unrestricted competition is less likely to lose in fitness than one which does so, should both of them become successful and crowded.

In man, if our knowledge were adequate, we should be able to distinguish genes with three different kinds of effect when substituted. For their "allelomorphs," as the alternative genes occupying the same place in a chromosome are called. First, genes making for economic success; second, genes making for increased fitness in the individual, and therefore biological success; and third, genes making for increased fitness of the species. In each case we must expect that genes which are advantageous in one environment will be disadvantageous in another. We must also remember that the human species transforms its environment, so that fitness is nothing absolute.

On the contrary, freedom is the recognition of necessity. At the moment it is quite likely that the most useful genes to the human species as a whole are those which cause spectacular abnormalities such as haemophilia, albinism or optic atrophy, and thus force us to investigate the laws of genetics, and so to achieve for our descendants a freedom from bondage to these laws.

At the present moment we have some slight evidence[5] for the inheritance of a make-up leading to individual economic success under capitalism, and to the loss of fitness which goes with that success. Capitalist eugenics assume without question that this make-up is desirable from the point of view of the species. It could be argued that the genes which are being eliminated in the bourgeoisie make for aggressiveness as well as intelligence, and that at the moment our species could do with fewer genes making for innate aggressiveness, if such exist. However, the whole question of the evolutionary effect of the class struggle has barely been raised as yet. And it is unlikely to receive an impartial answer until the class struggle has been liquidated.

I am fully aware of the inadequacy of this sketch. I have distinguished three Hegelian triads:

Thesis Antithesis Synthesis
Heredity Mutation Variation
Variation Selection Evolution
Selection of
the fittest
Consequent loss
of fitness
Survival of
noncompetitive species

I am perfectly aware that these represent a certain abstraction from reality. Thus selection probably affects the rate of mutation. An animal species alters its environment to some extent, and the human species does so to a great extent. Thus the evolutionary process itself affects the environment, which in turn determines its direction. Nevertheless, I hope I have shown that dialectical materialism furnishes us with a very powerful weapon for the interpretation of biological facts. It may well be that some theorists in the Soviet Union have attempted to apply dialectics to scientific problems for which simpler logical forms are better suited. Nevertheless, in other countries the attempt has not yet gone far enough.

Some writers, including biologists such as Hogben, while expressing a measure of sympathy with Marxism, have no use for those elements in it which are derived from Hegel. I have tried to show that, in biology at any rate, the intellectual technique of Marx, Engels and Lenin makes for clear thinking.


[1] V. I. Lenin. Materialism and Empirio-Criticism (New York, l927), p. 524.

[2] Anti-Duhring, (International Publishers, New York), p. 94.

[3] Letter to Kugelmann, April 17, 1871.

[4] It may be remarked that one of the first Englishmen killed while fighting for Spanish democracy was Cornford, a great-grandson of Charles Darwin, and a Communist.

[5] This evidence is far less convincing than is generally believed.

See responses to this article by A.P. Lerner and Haldane