Saturday, July 25, 2009

What Darwin Said (Part 3): Heredity   posted by DavidB @ 7/25/2009 04:20:00 AM

This series of posts attempts to identify the key propositions of 'Darwinism', in the sense of what Charles Darwin himself believed, and to assess their current standing. Part 1 dealt with 'The Pattern of Evolution'. Part 2 considered the 'Mechanisms of Evolution'. Part 3 considers Darwin's views and speculations on the subject of Heredity. I will cover the subject under the headings:

What Darwin got right
What Darwin got wrong
What Darwin didn't know

Note on sources

As previously, I will cite the Origin from Charles Darwin: The Origin of Species: a Variorum Text, edited by Morse Peckham, 1959, reprinted 2006. Darwin regarded the Origin as an 'abstract' of his theory of evolution, and intended to present it more fully in several longer works. The 2-volume Variation of Animals and Plants under Domestication (1868) carries out this intention for the chapters in the Origin on 'Variation under Domestication', 'Laws of Variation', and part of 'Hybridism'. The rest of the grand plan was never completed.



Darwin's most fundamental proposition on heredity is that 'like produces like' [Origin, 85]; in other words, offspring resemble their parents, or in modern technical terms, there is positive heritability for most traits. Darwin believed this to be the case both for relatively large variations ('sports') and for the smaller variations that he called 'individual differences'. This is fundamental to his theory because 'any variation which is not inherited is unimportant for us'. [85] Darwin's evidence for this basic proposition is taken mainly from domesticated animals and plants, and is set out at length in the first volume of Variation, with 100 pages on pigeons alone. He believed that the evidence was overwhelming, and dismissed objections as merely 'theoretical'. Modern genetic theory and experiment generally supports Darwin.


Before Darwin it was often argued that changes in animals and plants under domestication were only temporary, and that in the wild they would always revert to the wild form. Darwin accepted that if conditions changed, the characters of the animals or plants would also change as a result of natural selection, but he did not accept that they would always revert exactly to the original wild form. He emphasised that the new forms could continue indefinitely if conditions were constant and the animals or plants bred only among themselves: 'the antiquity of various breeds clearly proves that they remain nearly constant as long as their conditions of life remain the same' [Variation, vol.2, 416]. Darwin's belief in the potential permanence of change was challenged by Francis Galton, who believed that small quantitative changes would wash out in a few generations by 'regression towards the mean', and only the discontinuous variations known as 'sports' would resist this regression. Karl Pearson later showed that Galton's arguments rested on a misunderstanding of his own principle of regression.


Another difficulty for Darwin was the widespread belief that variations under domestication could never go beyond the normal limits of differences between species. One of the recurring points made in Variation is that the differences in external morphological characters between domesticated varieties were often greater than those between recognised wild species, or even genera: see for example Variation vol. 1, 36 (dogs), 49 (horses), 70 (pigs), 115 (rabbits), 133, 157 (pigeons). From modern knowledge of genetics there is no reason to believe in any inherent limit to cumulative variation, though there may of course be physiological limits, e.g. to the speed of race-horses.



As discussed in Part 2, like most biologists of his time Darwin believed in the inheritance of acquired characteristics (IAC). (See especially chapter 5 of the Origin, and chapter 24 of Variation.) It is now generally accepted that, with some special exceptions, IAC does not occur. (In the last few years some biologists have argued that 'transgenerational epigenetic inheritance' is widespread, but even if this turns out to be true, it has little to do with IAC in the traditional sense.) In Darwin's defence it may be said that in his day there were many phenomena which could most easily be explained by IAC. There also appeared to be strong experimental evidence to support it. Darwin accepted the claims of the French physiologist Brown-Sequard to have produced IAC in guinea pigs. In the first edition of Variation Darwin described this as 'the most remarkable and trustworthy fact' in support of IAC (vol.2, 24), while in the second edition, following further reports by Brown-Sequard, and apparent independent replication, he upgraded this to 'conclusive evidence' (2nd. edn, 1875, vol.1, 488). It was only after objections were raised by Weismann to the very possibility of IAC that biologists generally began to question the kind of evidence previously used. An example is the fact that animals living in caves are often congenitally blind. Sometimes their eyes have degenerated or disappeared, sometimes they are present but non-functional. Darwin thought that IAC resulting from disuse was the best explanation, as 'it is difficult to imagine that eyes, though useless, could be in any way injurious to animals living in darkness', and their deterioration could therefore not be explained by natural selection. Nor could it be explained by 'economy of growth' in cases where the eyes were still full-sized. It would now generally be agreed that when selection for any complex adaptation is relaxed, the genes underlying that adaptation will accumulate mutations, most of which will tend to destroy the adaptation. But this explanation was hardly available to Darwin, and it is a departure from a purist interpretation of natural selection, since it assumes that phenotypic change in the absence of selection is biased in one direction (towards deterioration). Another case that is difficult to explain by simple ideas of natural selection is the inheritance of callosities of the skin, as in camels, etc. Such callosities may be produced by friction in the lifetime of an animal, but in some species they are already developed before birth, and it seems most unlikely that genetic mutations would purely by chance produce congenital callosities just where they are needed. In such cases modern biologists would probably appeal to C. H. Waddington's concept of 'genetic assimilation', which is strictly consistent with natural selection but does give a causal role to the experience of the animal's parents, and therefore has the appearance of IAC (see John Maynard Smith, The Theory of Evolution, 3rd. edn., under index reference to Waddington.)


Darwin believed that all hereditary variation has some cause, usually change in the environment. This would not now be accepted. Strong environmental influences, such as radiation or powerful chemicals, can induce mutations, but in general mutation is due to chance failures in the mechanism of replication.


One belief about heredity prevalent in Darwin's time was that of telegony: the idea that a male who mates with a female may also influence her offspring from subsequent matings with other males. This was widely believed by animal breeders, and supported by a few apparently well-documented cases, the best-known of which was that of Lord Morton's mare. Lord Morton had mated one of his mares with a quagga, producing hybrid offspring, but one of the subsequent matings of the same mare with an ordinary horse had produced offspring with stripes, resembling a quagga. This and other cases were accepted by Darwin as proof that telegony could occur. They would now be dismissed as coincidence or bad observation (e.g. unobserved 'sneak' matings).

In writing this I assumed that the idea of telegony was long dead, but I have found the following at a dog-breeding website:

This subject, "Does the first impregnation of the female have any influence upon the progeny of subsequent breeding to other sires?" has been for years, and still remains, a disputed question. Scientists are arrayed on both sides of the question. Among dog breeders, the popular opinion is that it does, and many of the breeders who look with suspicion upon a bitch which has suffered a misalliance have had personal experiences with which to support their position

Some ideas die hard! So far as I know there is no support whatever for telegony among modern scientists.


When a plant is grafted onto a stock of a different species, the vegetation growing from the graft sometimes combines characteristics of both species, and can be propagated permanently by cuttings. Darwin attributed this to a true hybridisation of the two species, involving a permanent mixture of their hereditary material. The phenomenon observed is apparently a genuine one (if rare) but would now be attributed to the formation of chimaeras (mixtures of genetically different cells) , rather than a true hybridisation.


Darwin accepted claims that pollen occasionally affected the tissues of the mother plant, and not just the offspring. This would now be rejected as coincidence or faulty observation.


What Darwin didn't know about heredity was the whole of modern genetics, from Mendelian ratios to DNA and the genetic code. He (and others of his generation) lacked even the most basic facts about the mechanism of heredity, such as the role of chromosomes and the process of recombination. In Darwin's time it was not even certain that fertilisation of an egg involves only one sperm (or pollen grain), and Darwin accepted evidence that sometimes more than one was required.

But I think the most serious gap in Darwin's knowledge was the concept of segregation. The idea that an offspring receives a quasi-random selection of genetic material from each parent, which together makes up the unique genotype of the offspring, is entirely absent from Darwin's thinking. It was not impossible for such a concept to be formed in Darwin's time, since Mendel achieved it, and Francis Galton came close. Darwin himself was aware of many facts which are easily explained by segregation. It is often supposed that Darwin believed in a doctrine of 'blending inheritance', in which an offspring is always intermediate between its parents, but this is an oversimplification of his position. He was aware of many traits that do not 'blend', and of some that appear 'prepotent', while others may be 'latent' and reappear in later generations. These are close to the Mendelian ideas of dominant and recessive. But Darwin never worked out a quantitative theory of such phenomena, and he lacked the grasp of basic combinatorial mathematics which guided Mendel and Galton in that direction.


Darwin attempted to account for the phenomena of heredity with his 'Provisional Hypothesis of Pangenesis', as set out in chapter 27 of Variation. This is quite an elaborate theory, but it can be summarised in four main points:

a. each unit of the body (and probably each cell) produces particles, known as gemmules, which are capable in the right circumstances of reproducing units of the same type

b. gemmules of different kinds are diffused through the fluids of the body

c. in sexual reproduction, gemmules from all over the body are packaged into sperms and ova

d. in the process of individual growth and development, gemmules give rise to cells of the kind from which they were themselves derived, in the proper order and place determined by their 'mutual affinities'.

By these principles Darwin attempts to explain not only the normal course of reproduction and development, but such supposed phenomena as IAC and telegony. IAC, for example, can be explained by assuming that the greater or lesser use of an organ influences the number of gemmules produced and entering the reproductive system.

In the light of modern knowledge the theory of pangenesis is almost entirely false. Its only useful component is the principle that the units of heredity are physical particles, rather than, say, fluids, immaterial influences, or vibrations of some kind, as supposed in some rival theories. The most obvious weakness in the theory is the idea that gemmules are diffused through the body and collected into the sperms and ova (sometimes known as the 'transport hypothesis'). No-one had ever observed such mobile gemmules, and Darwin's only defence was to point out that they must be extremely small, like many disease organisms (e.g. those responsible for smallpox) which had not been observed but must exist. By the end of the 19th century it was widely accepted that the material of heredity was contained in the chromosomes, which left no place for loose gemmules circulating round the body.

The main reason for Darwin's adoption of the transport hypothesis was the need to explain IAC. If IAC in the traditional sense occurs, then there must be some feedback from the affected parts of the body to the reproductive organs. The transport hypothesis meets this need. Darwin deserves some credit for recognising the need for a mechanism to explain IAC, which had previously been lacking. Historically, the theory of pangenesis was important in highlighting the implausibility of such a mechanism, and therefore of IAC itself. Francis Galton was the first to seriously question the existence of IAC, and his scepticism about it arose directly from his own attempts to test pangenesis through experiments in blood transfusion. De Vries, Weismann, and other pioneers of modern genetics, were also influenced by the need to develop alternatives to Darwin's theory. For example, De Vries (one of the rediscoverers of Mendel's laws) called his own model of heredity 'intracellular pangenesis', as it was based on Darwin's theory with the 'transport hypothesis' left out.

Summing up, Darwin had little (correct) knowledge about the mechanisms of heredity, and much of what he thought he knew was in fact wrong. Fortunately, this had little effect on his theory of natural selection, as this depended only on the existence of an adequate supply of heritable variation, from whatever source. Darwin concluded from his study of domesticated animals and plants that an adequate supply of heritable variation did exist, and this has been amply vindicated by modern genetics.