Originally this diary was going to cover almost twice as much territory but when I saw how long it had become already I decided to split off the development of population genetics between 1918 and the early 30s into a separate diary. The first third of the past century is particularly fascinating in the history of evolutionary biology for two reasons. First it is the period when the basis of the modern theory of evolution was laid down (next diary). The second is that the development of genetics, which allowed for a much more complete theory of evolution, caused a lot of conflict delaying progress in the field for well over a decade (this diary).
I'd like to acknowledge at the outset that most of the material in this diary was obtained from the book: The Origins of Theoretical Population Genetics by William Provine, published by The University of Chicago Press in 1971. This work was Provine's doctoral dissertation (he got his PhD from Chicago the year before). Provine is probably the foremost historian of evolutionary genetics in the world (it is a specialized field) and has been on the faculty at Cornell for many decades. So I am cheating a bit here and taking advantage of Dr. Provine's expertise.
Also kossack Bernard Pliers published some comments with excellent links on the early history of genetics in part 3 of this series. I encourage those of you with an interest to check them out.
Recap - The Problem of Inheritance
In earlier installments of this series I've mentioned that one of the biggest troubles faced by Darwin and those that followed after him in the last third of the 19th century was heredity. Everyone knew that characteristics were inherited by offspring from their parents but no one knew how. The common view of blending inheritance seemed clearly incompatible with natural selection because blending removed variation that is necessary for natural selection to cause evolutionary change. Darwin eventually came to a more of less Lamarckian view (inheritance of acquired characteristics) as that was the only way to generate enough variation to allow natural selection to operate. Other biologists in the late 19th century also held Lamarckian views. The idea that the environment could directly cause changes in characteristics and that these changes could be passed on to offspring was an obvious mechanism for maintaining variation on which natural selection could act. However no one could actually demonstrate this kind of inheritance.
As discussed in my digression (in between parts 2 and 3) Gregor Mendel (1822-1884) had published his work on genetics in the 1860s and it was not recognized by the scientific community of the time. Mendel proposed particulate inheritance which would not destroy variation in the way that blending inheritance would. August Weismann (1834-1914) a prominent German evolutionary biologist proposed the germ plasm theory in the 1880s. The germ plasm(what today we we call testes and ovaries) is the tissue that produces reproductive cells and is thus the tissue responsible for inheritance. Weismann firmly rejected Lamarkian inheritance.
Rediscovery of Mendelism
Mendelian inheritance was rediscovered independently by three different researchers in 1900 and Mendel was rapidly but posthumously given credit. Please see the digression for more details on on how Mendelian inheritance works. Although Mendelism put the inheritance of acquired characteristics out of contention, a conflict was already underway among evolutionists and the rediscovery of genetics only exacerbated the fight.
Discontinuous and Continuous Variation
Basically the conflict was not over evolution but over the importance of natural selection to evolution and the nature of inherited changes in phenotype (physical characteristics of organisms). As mentioned last time in part 3, some of Darwin's strongest supporters such as Thomas Huxley and Galton believed that the primary mechanism for evolutionary change was sudden jumps in form. They didn't reject natural selection but didn't believe it was capable of causing organisms to evolve into different types of life. This is a theme we shall see repeated several more times in this series. Some evolutionary biologists, typically those studying evolution on large scales (e.g. comparing different species, studying the fossil record), have tended towards the idea that major evolutionary changes require something more than 'just' natural selection acting on existing variation within populations. They see gaps in form between different kinds of organisms and hypothesize that something in addition to natural selection is needed to bridge those gaps. Other evolutionary biologists have tended to the idea that natural selection is capable of producing the diversity of life on earth. In later diaries, when we have covered more evolutionary theory and possible genetic mechanisms we will return to these ideas.
In contrast to the views of Galton and Huxley were two of Galton's close colleagues Karl Pearson (1857 - 1936) and Raphael Weldon (1860-1906) who were firmly in the natural selection camp. Galton, Pearson, and Weldon were all members of what came to be known as the biometricians. They applied statistical techniques such as regression and correlation to the study of inheritance. In doing so they developed many of the standard techniques of statistical analysis. Pearson and Weldon were convinced that natural selection acted on variation that existed, even in very small degrees, in almost anything that could be measured. In other words they emphasized continuous variation such as is seen in many traits such as height. Evolution occurred by gradually changing the average value of the trait in a population rather than a sudden abrupt change.
In the other camp was William Bateson (1861-1926) a firm believer in evolution through discontinuous variation, in other words more abrupt changes between organisms that placed them into clearly defined categories rather than along a continuous scale. Bateson performed a series experiments examining inheritance of traits such as floral symmetry among others, demonstrating inheritance of discrete differences in traits over and over again.
Although Bateson was not among those who rediscovered Mendelian inheritance in 1900 he was performing experiments that would have caused him to do so within a year or two. Bateson rapidly became one of the foremost advocates of Mendelian inheritance and took what was already a heated controversy to a whole new level. The biometricians (other than Galton) did not regard Mendelian inheritance as important to evolution while Bateson regarded the major differences between different genotypes in Mendelian crosses as vital to evolution and the continuous variation studied by the biometricians as trivial. In fact it was claimed that the inheritance of continuous variation detected by Pearson's statistical techniques was not of long duration and could not cause any real change in populations.
Mendelian Inheritance vs Natural Selection and Continuous Variation
Why did Mendelian inheritance get set up as an opposing idea to evolution by natural selection? The reason is that the easiest way to observe Mendelian inheritance is to study characteristics that are affected primarily by a single gene with a major effect on the trait. Early geneticists focused on genes that had discontinuous variation.
To use a cat example that I've used before. The agouti locus in cats produces cats with (at least some) agouti (banded) fur if the cat is genotype AA or Aa but solid colored cats if the genotype is aa. Because the fur color was inherited in discrete categories (cats had one fur type or the other) it was easy to see how a change in genotype in population could suddenly produce a new form.
Cat ear length is a characteristic that varies continuously. It probably is under the control of many genes. The Mendelian pattern on inheritance is not going to be obvious if cats with different ear lengths are crossed because of the many genes involved. Instead you will probably see something that resembles blending inheritance because of averaging the effects of the genes of the two parents. Also cats with slightly different ear lengths are not as obviously different as cats with different coat colors because other cats in the population will have other intermediate ear lengths.
Because Mendelian inheritance was understood to some extent (the pattern was understood if not the underlying mechanism) and variation in continuously varying traits was not as well understood, it was possible or the Mendelians to argue that continuous variation was fairly meaningless and that it was Mendelian inheritance itself that was driving evolution in a pattern of sudden abrupt changes. Another factor in the conflict was that understanding the biometricians' arguments required some mathematical sophistication. Not a lot by today's standards but more than most biologists at the time, including Bateson, were capable of.
Bateson had been feuding with Pearson and Weldon for several years prior to the rediscovery of Mendelism and the discovery only intensified their feud. It was an ugly and personal conflict. Because of the prominence of these individuals in the British biological establishment the effects of the conflict were fairly widespread and even published in newspapers. Students got caught in the crossfire. Weldon died very young in 1906. Pearson was angry that so much of Weldon's time and talent was spent in arguing with Bateson.
Galton managed to largely stay outside the fray. He was linked to both sides because he was the founder of the Biometricians and thus a senior colleague of Pearson and Weldon. However Galton's views more closely coincided with those of Bateson. Both sides tried to claim him as one of their own. Galton was, like the other surviving figures of the early Darwinian period (Wallace and Hooker) an old man at this point, a senior figure of authority in British biology.
Resolution of the Conflict
The resolution of this conflict is that there was no real conflict. The variation underlying the continuous traits of the biometricians is also Mendelian. It is simply caused by many genes with individually small effects. This was demonstrated by a number of different geneticists doing experiments in different biological systems in the period immediately preceding WW1. William Castle (1867-1962 - not the same guy as the horror film director) was an American geneticist. Castle was originally an advocate of discontinuous evolution and a follower of Bateson. Gradually, after many breeding experiments done by himself and his students on both rats and guinea pigs, he came to realize that variation in traits with Mendelian inheritance could be altered by selection. As a side note, one of Castle's students in this period was Sewall Wright about whom we will hear much more in the next diary.
Herman Nilsson-Ehle (1873-1949) was a Swedish agricultural biologist. Edward East (1879-1938) was an American agricultural biologist working in Illinois and later at Harvard. Working on crop plants both Nilsson-Whle and East demonstrated that continuous variation was caused by Mendelian inheritance of many genes at the same time. Nilsson-Ehle also clarified the role of sexual reproduction is creating new genetic combinations in offspring and thus generating variation.
Thomas Hunt Morgan (1866-1945) was probably the most important figure in early genetics. He began research using 'fruit flies', Drosophila melanogaster, establishing them as the most important system for studying genetics. The ease and speed of breeding Drosophila in large numbers meant that Morgan's lab could do experiments on scales that researchers working on mammals or plants could only dream. Morgan and his colleagues observed the formation of new variation through visible mutations (such as flies with different colored eyes). Morgan developed the idea that natural selection caused evolution by causing favorable mutations to increase in frequency.
The experimental work had now established a model in which traits were inherited through units called genes. Different versions of these genes are known as alleles. A trait could be affected by one gene or by many genes. Genetic variation within populations was at least partly maintained because sexual reproduction produced offspring with different combinations of alleles.
And, crucially, we have the idea that evolution proceeds through the changing of the frequency of alleles in populations. Particular versions of genes become more or less common.
The Hardy Weinberg Equilibrium
We'll end this diary with an important theoretical result that forms the baseline for the work discussed in the next diary. If you have taken introductory biology as an undergraduate or had a good high school class in biology it may be familiar to you. It is the Hardy-Weinberg equilibrium.
This body of theory disproves the early idea of Udny Yule (1871-1951) that dominant alleles will tend to win out and become more common than recessive alleles which he published shortly after the rediscovery of Mendelian inheritance. This is a widespread misconception among students learning genetics today. The idea was convincingly disproved by the British mathematician G.H. Hardy (1877-1947) and the German physician Wilhelm Weinberg (1862-1937) independently in 1908.
The HW equilibrium is a mathematical proof that Mendelian inheritance alone will not change allele frequencies. If a series of assumptions are met that prevent any evolutionary force from acting on the population then the allele frequencies will stay the same in a population forever. The assumptions are unrealistic (no population is infinitely large, very few populations are completely isolated from other populations, etc) but that isn't really the point. The point is that the simple act of passing in genes will not, in and of itself, cause evolution.
The math of HW is fairly simple. I will first remind you that genes are locations on chromosomes known as loci (singular locus) and there are different versions of a gene called alleles. In most organisms an individual would have two copies of a gene. These could be two copies of the same allele (homozygous) or one copy each of two different alleles (heterozygous). This is known at the genotype (again see the digression for more details.
If you have a gene (locus) with two alleles A and a then each of them will have a frequency. The frequency of A is p and the frequency of a is q. If p = 0.6 that means 60% of the alleles for that gene in that population are A. Necessarily the frequency of a is 0.4. So p=0.6, q=0.4 and p+q=1.
HW assumes random mating among all individuals in a population. This means that when you look at a zygote (fertilized egg) drawn at random from the population the probability that it gets allele A from the mother is 0.6, the probability it gets allele a from the mother is 0.4, the probability it gets allele A from the father is 0.6 and the probability it gets allele a from the father.
From this you get that the probability that the fertilized egg has the genotype AA is 0.6*0.6 = 0.36 = p2
The probability that the fertilized egg has the genotype aa is 0.4*0.4= 0.16 = q2.
The probability that the fertilized egg got the A allele from its mother and the a allele from its father is 0.6*0.4 = p*q = 0.24
The probability that the fertilized egg got the a allele from its mother and the A allele from its father is 0.4*0.6 = q*p = 0.24
The last two sentences are two different ways of getting the same result (a fertilized egg with the genotype Aa). So the frequency (probability) for Aa fertilized eggs is 2*p*q which in this case is 0.48.
Thus if no evolutionary force is acting on the gene then a locus with allele frequencies p and q will result in offspring genotype frequencies of p2, q2, and 2pq.
And, as this diary is already long enough, I won't go through the next step but if you look at the allele frequency in this generation it is exactly the same as it was in the previous generation. In other words p and q will never change.
I'm going to add a personal comment here in that the teaching of HW so widely is a bit unfortunate as its relevance to evolutionary biology is somewhat subtle and often misunderstood. I was taught it badly the first time round and although I could do the math quite easily I completely failed to understand why it was relevant. I think this is a common occurrence.
One important consequence of HW is seen in genetic diseases such as cystic fibrosis which are caused by a single recessive allele. In this cases individuals with genotypes AA and Aa are healthy while genotype aa are individuals with the disease. In many cases these diseases result in aa individuals with little or no likelihood of ever having children, especially in the era before modern medicine. Why does natural selection not remove the a allele from the population?
The simple explanation is that as q gets smaller (the a allele becomes more rare) the frequency of aa individuals (q2) gets smaller even faster. For example
If q is 0.1 (10% of the alleles are a) then q2 is 0.01 (1% of the individuals have genotype aa). In this case 2pq = 0.18 (90% of the a alleles are in heterozygotes and not subject to selection)
but
If q is 0.01 (1 in a hundred alleles is a) then q2 is 0.0001 (1 in ten thousand individuals is aa) In this case 2pq is 0.0198 (99% of the a alleles are in heterozygotes and not subject to selection).
So the more rare the disease causing allele is, the greater the proportion of those alleles will be heterozygotes that are not sick. Eventually the a allele reaches a frequency at which the very small number removed by selection will be balanced by new mutations producing new a alleles.
The next phase in evolutionary biology is the development of models showing how different evolutionary forces will act to change the genetic composition of populations. That is the topic of the next diary.