Here I am again inflicting upon you another diary on the history of evolutionary biology. Previous diaries can be found here, here, here, and here. This diary covers the period from 1918 through the early 1930s. Although not noticed by very many people at the time, this era saw the theoretical work (and some experimental work as well) that established the fundamentals of modern evolutionary biology. As with part 4, I am greatly indebted to William Provine's book, 'The Origins of Theoretical Population Genetics' for much of the material in this diary.
Recap In the previous two diaries we saw that while evolution was widely accepted by biologists within a few decades after Darwin's publication of the 'Origin' the importance of natural selection was quite controversial. The rediscovery of Mendelian genetics in 1900 led briefly to two distinct schools of evolutionary biologists: the Mendelians and the Biometricians. The biometricians thought that Mendelian factors (i.e. genes) were not involved in the inheritance of most characteristics while the Mendelians believed that inheritance itself drove evolution.
By the end of first world war these conflicts had largely been resolved, at least at the genetic level. Evolution became defined as the process of changes in the genetic composition of populations of organisms. This was refined to define evolution as changes in the frequencies of alleles in populations over time. Hopefully you will recall that an allele is a particular version of a gene. Imagine a population of cacti on an island. The color of the flowers of these cacti is controlled by a gene with two alleles A1 and A2. Each individual will have two alleles. A1A1 individuals will have white flowers, A2A2 individuals will have red flowers, and A1A2 individuals will have pink flowers. I am also going to inform you that mating in these cacti is accomplished by pollinating animals moving pollen from one cactus to another. Each pollen grain is haploid and has either an A1 or an A2. It produces a sperm that unites with an egg in the cacti on which it arrives. The egg is also haploid.
The alleles A1 and A2 will have frequencies in the population. If all the cacti have white flowers then the frequency of A1 is 1.0 and the frequency of A2 is 0.0. If we have a population of 500 cacti and 240 have pink flowers, 80 have red flowers, and 180 have white flowers then the frequencies are as follows:
500 individuals = 1000 total alleles (2 per individual)
80 red individuals (A2A2) = 160 A2 alleles
180 white individuals (A1A1) = 360 A1 alleles
240 pink individuals = 240 A1 alleles and 240 A2 alleles
So the frequency of A1 = (240 + 360)/1000 = 0.6
The frequency of A2 = (240 + 160)/1000 = 0.4 as well by the same logic.
At the end of the last diary we discussed the Hardy Weinberg equilibrium which is a simple mathematical proof that without some evolutionary force acting on a population the allele frequencies will stay the same forever (evolution will not happen). So what are these evolutionary forces? In about 15 years following the end of WWI three men worked out the fundamental theory describing how evolution acts on genes in populations. Follow me below the fold to find out more.
Evolutionary Forces
There are about half a dozen forces that will change allele frequencies in populations. They are listed below in approximate order of importance.
1. Selection. Darwin's theory of evolution by natural selection can be rephrased in genetic terms. In the example that white flowers are visited by pollinators more often than pink or red flowers both of which get equal but smaller numbers of visits on a per flower basis. As a result the white allele (A1) ends up being in more seeds in the next generation.
2. Genetic Drift. Drift is very important but often harder for people to grasp. Drift refers to changes in allele frequency due to random differences in success at reproduction. In other words changes that have nothing to do with the trait in question. Imagine that there is a rock slide on the island that crushes most of the white flowered cacti. This doesn't have anything to do with the flower color - they just happened to be growing in the path of the rockslide. Less dramatically the seeds of the cacti may fall on bare rock, on soil, or in the ocean. Only one of these things will lead to the growth of new cacti. Random variation in the success of seeds with different genes will happen each generation.
3. Gene Flow (Migration). This refers to changes in allele frequency due to movement of individuals or gametes (eggs or sperm) between populations. Imagine that the mainland has a much larger population of cacti of the same species that are all red-flowered, perhaps because there is a pollinator on the mainland that only visits red flowers. Small numbers of seeds or pollen will get carried out to the island and will slightly change the allele frequency (increase the frequency of A2).
4. Mutation. Mutation is the accidental conversion of an individual allele into another allele due to some mistake in DNA replication (this is an over simplification and we can revisit this in a later diary). Mutations happen very rarely and most are detrimental but they are the ultimate source of the genetic variation necessary for evolution to act.
5. Meiotic Drive. This is a very specialized evolutionary force and it is unclear how widespread it is. It is the ultimate example of a 'selfish gene'. If this was occurring in our cactus it might take the following form. Individuals with pink flowers (A1A2) would only produce pollen and eggs with the A1 allele. During meiosis the cells that ended up with A1 would somehow kill the cells with A2. There are known examples of this type of system in both mice and fruit flies. It tends to have bad effects on the populations with those genes over time. We won't discuss this any further in this diary.
6. Nonrandom mating. If the cacti were pollinated by different insect species, some of which only visited white flowers and others of which only visited red flowers you would rapidly see a population that was no longer in Hardy Weinberg equilibrium and if the preferences were strong enough then the pink cacti might vanish and the red and white flowered cacti become separate species.
The great achievement in evolutionary biology in this era was the development of a body of theory that developed models of how these evolutionary forces would work, separately and in concert, on allele frequencies in a population. This work was done by four men who were able to bring strong mathematical backgrounds to bear to create the field of evolutionary population genetics. One of them, Sergei Chetverikov (1880-1959) unfortunately rarely gets credit for his pioneering work. His 1926 paper was written in Russian and not readily accessible to the academic world outside of the Soviet Union. Unfortunately, but not surprisingly, Chetverikov fell into the bad graces of Lysenko. Lysenko was the dominant figure in genetics in the Stalinist period of the USSR. He advocated a Lamarckian system of inheritance more in line with Marxist thought and purged those who stood up to him. Chetverikov spent the rest of his life in obscurity.
Fisher, Wright, and Haldane
The other three figures all eventually became quite well known to the world of evolutionary biology although in the period that they did most of their work it remained inaccessible to most of their colleagues because a lot of it was in a foreign language - MATH! These men were R.A. Fisher (1890-1962), JBS Haldane (1892-1964), and Sewall Wright (1889-1988). Given that they were all educated white males they were a disparate group. I'll introduce the three of them first, then summarize the major conclusions that can be drawn from their work and finally go through their differences which continue to influence the field to the present days.
Ronald Alymer Fisher is, in my opinion, one of the great scientists of the 20th century. He has been described (not by me but rather by Richard Dawkins) as the most important evolutionary biologist since Darwin. While I don't think this would be a universal sentiment, I think he would definitely be in the top ten on anyone's list (that is anyone who knew enough about evolution). But Fisher was also perhaps the most important statistician of the 20th century, developing statistical tests such as analysis of variance and also developed factorial design in experiments, a concept now so fundamental it is amazing to realize that it is less than a century old.
Fisher came to evolution from a mathematical background. He demonstrated in a paper in 1918 that the results of biometrical studies could be explained by Mendelian inheritance. This paper demonstrates how inheritance of many different genes results in particular patterns of genetic variation. Over the following decade Fisher built upon this model to show how the various evolutionary forces would alter allele frequencies in populations. The culmination of this work was his book, 'The Genetical Theory of Natural Selection' published in 1930.
Fisher is an interesting figure because he reveals the complexity of scientists as human beings. Clearly Fisher was genius, a towering figure in two different fields. Yet he was also clearly a figure of his time. Born into a prosperous but non-aristocratic British family Fisher was a political conservative and a devout Anglican. He was a committed believer in eugenics and the latter part of his book is devoted to an analysis of the danger to the British empire from the upper classes failing to reproduce. Later in life he was extremely critical of the early attempts to demonstrate a link between smoking and cancer.
Sewall Wright was born in Massachusetts but grew up and spent much of his life in the midwest. His father taught at Lombard College in Galesburg, Illinois where Wright grew up and was an undergraduate. Unlike Fisher, Wright was not a formally trained mathematician, was trained as a biologist and largely self taught in math. Wright's Ph.D. advisor was William Castle who we encountered in the previous diary. After grad school Wright took a job with the USDA. He began working on series of papers on the effects of inbreeding on genetic structure. His theoretical work along with experiments carried out by the USDA gave him a strong conviction in the importance of interactions between genes in determining fitness.
Wright then developed theoretical models of the action of evolutionary forces based on his earlier work on inbreeding, culminating in a long paper, entitled 'Evolution in Mendelian Populations' on the topic in 1931.
Wright's long life is notable for its lack of incident. In contrast to his two British colleagues he was very mild mannered, not that he was not afraid to stand up for his ideas. In his position at the USDA and then his later faculty position at the University of Chicago his theoretical work was largely a sideline to his day job as a practicing geneticist and teacher. He retired from Chicago in 1955 and moved to the University of Wisconsin. This is an interesting contrast to Fisher and Haldane who moved to institutions in Australia and India at the end of their careers. Instead of moving around the world, Wright moved to an adjacent state. He spend an amazing third of a century in 'retirement' in Wisconsin during which he refined and gathered support for his ideas. He published a series of four books summarizing his news on evolution. Volume four was published when he was almost 90. His last paper was published in December of 1987, right around the time of his 98th birthday and over 75 years after he began his research career. That was the end of my first quarter in my PhD program and I remember getting the journal in the mail. I was struck by two things in his paper: it had an obviously hand drawn figure and it was accepted without review the day it was received by the journal.
John Burdon Sanderson Haldane came from an aristocratic and intellectual Scottish family. Haldane had early interest in genetics, apparently dating to a talk he heard at age eight about the newly rediscovered field of Mendelism. Haldane was influenced by T.H Morgan and the Drosophila geneticists which started him on the path of seeking theoretical explanations for results generated by other scientists.
Haldane wrote a series of papers in the 1920s describing the action of various processes on genes in populations. In 1931 he summarized his work in a series of lectures which were published as a book 'The Causes of Evolution' in 1932.
Haldane, as a person, seems by far the most interesting of the three. He was a Marxist, wrote children's fiction and poetry, and was extremely quotable (I'll include some quotes in a comment). He also wrote popular works on evolution. Late in his life he emigrated to India, angry over British imperialism in the Suez. He took up Indian citizenship and habitually wore Indian clothing. At the end of his life he composed whimsical poetry to the carcinoma cells that were killing him.
Points of Agreement
The three men worked independently and used different mathematical techniques. They were all aware of the work of the others and Fisher and Wright apparently each corrected errors in the work of the other man. All three of them came to the same basic conclusions about the fundamental ways in which selection, drift, etc worked in populations. Their primary concern was to conclusively demonstrate how Darwinian evolution and Mendelian inheritance interacted. However they differed in their views on the relative strengths and efficacy of the various forces in nature. We will cover that in the next section.
Here are some fundamentals of evolutionary genetics.
A. Drift is more powerful in small populations relative to large populations. Individual chance events have proportionately greater effects in small populations than in large populations.If you think about coin tossing this makes intuitive sense (I hope). I you toss 10 fair coins you expect to get 5 heads and 5 tails but it is not at all unlikely that you would get 6 heads (or even 7 or 8). However if you flip 10,000 fair coins you expect to get 5,000 heads and 5,000 tails. Getting 5,003 heads is not all unlikely but this has a minute effect on the frequency of heads. Getting 6,000 heads to of 10,000 is extremely unlikely and that is what would be necessary to get the equivalent change in frequency to flipping 6 out of 10 heads.
B. As a result of A, selection tends to be a more powerful force in larger populations. In small populations drift will often counteract selection.
C. In the absence of other evolutionary forces drift will cause genetic variation to decrease within populations and to increase among populations. Imagine many small islands with cacti on them and no movement of pollen or seeds among islands. If the populations are small the frequencies of A1 and A2 will fluctuate a lot on each island. In each case eventually the frequency of either A1 or A2 will drop to zero just due to chance. Barring mutation then all the cacti on that island will only have that one allele into the future. Some of the islands will end up 'fixed' for A1and some for A2. So some islands will have all white-flowered cacti and others will have all red-flowered cacti. This is not the result of differences in environment among the islands (in this example) simply due to chance in evolutionary history. This process will happen in any population not subject to other evolutionary forces but on average takes longer in larger populations.
D. Selection can have several different effects on genetic variation in populations. Directional selection will cause some alleles to become more common and others to become more rare such that the population changes phenotype (e.g. red becomes more common and pink and white become more rare). Stabilizing selection favor the current allele frequencies and keep the population where it is (preventing drift from causing changes). Disruptive selection will tend to split the population into two (or more) groups (e.g. white and red flowered cacti are favored and pink flowers are selected against).
D. Only fairly small amounts of migration among populations are necessary to counteract the effects of drift. So if some pollen blows between islands then the allele frequencies on the islands will remain similar unless selection is acting differently on the different islands.
E. Rare alleles, such as those caused by mutations, will frequently be lost due to drift even if they are favored by selection.
These results could be derived from models for single genes or for cases involving many genes in which the effects of the genes are additive (in other words the effect of each allele at each gene is affected by the other genes present). This is by no means a complete list.
Points of Disagreement
All three of them agreed about the fundamental properties of the evolutionary process. Where they disagreed were the nature of populations, the strength of selection in nature, and the nature of gene interactions.
The simplest models of evolution are ones in which selection is strong, drift is weak, and each gene can be regarded as relatively independent of other genes in terms of its overall effect on fitness. What do we mean by this. Imagine their are two different genes A and B each with two alleles (A1, A2, B1, B2). Imagine further that red flowers (A2) are favored over other colors. This would be true no matter what the genotype was at the other gene (B1B1, B1B2, or B2B2). In other words if A2 is better than A1 it is always better than A1.
This is the Fisherian view of the universe. Fisher saw evolution are proceeding primarily in large populations and he thought each gene could be thought of as evolving more or less independently of other genes. So in Fisher's world selection is the paramount force. Fisher did not actually believe that natural selection on individual genes was particularly strong, in actuality he tended to believe that selection was fairly weak, but in large populations strong selection is not necessary for selection to 'win' over drift.
In the Fisherian universe favorable mutations would arise and increase in frequency fairly often. Most mutations wouldn't make it but because population sizes were large there would be enough mutations that selection would cause the increase in favorable ones fairly often. Weak selection would cause these favorable mutations to slowly become more common. This theoretical result matched the observation that most characteristics showed considerable genetic variation. This variation would tend to be removed by selection or drift and Fisher's ideas showed how some variation could be maintained at all times.
In contrast Wright was very concerned with the problem of epistasis. He thought that specific alleles from multiple genes lead to high fitness only in certain combinations. To return to the cactus example again imagine a second gene with alleles B1 and B2. Imagine that having the genotype A1A1B1B1 resulted in high fitness and the genotype A2A2B2B2 gave even higher fitness but that all other combinations (such as A1A1B2B2) resulted in very low fitness. If the population starts out consisting of all A1A1B1B1 individuals it is very hard to evolve to a population that is all A2A2B2B2 because along the way you are going to get other genotypes with very low fitness.
Wright used a very powerful metaphor to describe this view of evolution. He called it an adaptive landscape. The genotype A1A1B1B1 is an adaptive peak. Changing any one of the four alleles involved to A2 or B2 will reduce the average fitness. A2A2B2B2 is a high adaptive peak but it is necessary to cross an adaptive valley to get to it.
Fisher was not convinced that this was a serious problem, both because he did not think epistasis was that common and powerful and because he felt that the environment was variable enough that the adaptive landscape would not be stable over time. Instead it would be more like the surface of the ocean in which the location of peaks and valleys would change. Genotypes that were peaks at one point in time would become valleys later on.
Wright obviously did think this was a problem and he developed a model of evolution to overcome it. This model is known as the shifting balance model of evolution. I'm going to spend a bit of time on this, not because it has been determined to be the best model of evolution, but because it is more difficult to understand that the Fisherian view.
In Wright's model, there are many small populations with minimal movement of genes between them. All of these populations have a genotype at an adaptive peak. Because of the small population size drift is fairly powerful force. One or more populations can move through an adaptive valley and end up on a higher adaptive peak (i.e. evolve genotypes with lower fitness through drift and then end up with genotypes with even higher fitness than originally).
Once a population has got to the new adaptive peak it should grow rapidly due to its high average fitness. Individuals with the new and improved genotype will migrate to other populations and spread the genotype.
Because of the important role of drift in Wright's model many people have felt that Wright de-emphasized selection. In fact Wright thought selection was very strong but that it tended to act on combinations of genes rather than individual alleles.
Haldane's studies of natural systems led him to believe that natural selection was generally very strong. Originally of the opinion that gene interactions were not important he later came to regard them as somewhat important. Haldane doesn't seem to have established a school of thought to the same extent that Fisher and Wright did. More so than the other two, a lot of his theoretical work was oriented towards explaining specific natural phenomena and research results.
All three men did significant work on specific evolutionary problems as well (e.g. Fisher's theory of runaway sexual selection) but I'm going to only focus on the big picture in this diary.
The Aftermath
As I mentioned, this work was not widely appreciated at the time of its initial publication. In the 1930s it became more widely appreciated and was the partial instigation for what is now known as the Modern Synthesis (the subject of the next diary). We will move beyond genetics and discuss evolution in the context of a number of other biological disciplines once again.
Fisher's view on evolution were enormously influential. British evolutionary biology has tended to favor the view that selection is powerful, largely as a result of his influence. W.D.Hamilton, probably one of the two or three most influential evolutionary biologists of the second half of the 20th Century was strongly influenced by Fisher although I am not sure they ever met. Fisher's ideas have also been influential in other countries as well.
Wright's ideas have been most influential in the US and perhaps Japan. The theoretical machinery he built has been very useful in studying population structure. The Shifting Balance Theory has been strongly criticized by many population geneticists. The main criticism is that it can only work under a fairly restrictive set of situations which are not regarded as realistic.
Still his ideas have been influential in many areas of evolution. Those interested in studying group selection (or the possibility of group selection), one of the most controversial topics in evolution over the last 60 years, have been interested in his models using populations split into small groups. Perhaps most importantly, the enormously influential Neutral Theory of evolution builds on ideas of drift and population subdivision. The neutral theory was developed by Japanese evolutionary geneticist Motoo Kimura who worked with Wright in Wisconsin.
Lastly Wright has been an inspiration to some who argued against the Fisherian world view. Stephen Jay Gould wrote Wright's obituary in the journal Evolution. He spent more time railing against what he called the 'Hardening' of the evolutionary synthesis to emphasize the power of selection above all else. Gould regarded Wright's Shifting Balance as very much in line with his own thinking (although Gould was not a geneticist).
Women and Evolutionary Genetics
As with most academic disciplines, in the late 19th and early 20th century evolutionary biology was an overwhelmingly male endeavor. On top of that, evolutionary genetics has been one of the more male dominated fields within biology. Presumably women, more so than male biologists, lacked the mathematical training necessary to participate during that era. As late as the 1990s I heard prominent male evolutionary geneticists speculate that the heavy math component was responsible for the high female grad student drop out rate (I remember my wife (to be at that point) pointedly disagreeing with the person who said this - a really gutsy move for a grad student, even a senior one). I will discuss these issues more in later diaries as well as the unfortunate severe under-representation of people of color in evolutionary biology which continues to this day.
At this point I did want to note that women played key roles in getting at least two of our subjects for this week on their way in their careers. Haldane's sister Naomi was breeding guinea pigs as a hobby and Haldane 'collaborated' with her in his first breeding experiments using her animals (she was a preteen at the time). Sewall Wright was introduced to genetics, evolution, and scientific research by a woman. His undergraduate instruction in genetics and evolution was taught by Wilhelmine Entemann Key (1872-1955). She was a native of Wisconsin and received a Ph.D. for the University of Chicago in 1901, one of the first female Ph.D.s from that institution. She taught for several years at Lombard College where Wright was an undergraduate. She encouraged him to go to graduate school and they remained friends until her death.
A Personal Note This diary is the first in the series to discuss individuals who were alive in my own lifetime (although admittedly both Fisher and Haldane died when I was a small child). I have come into contact with people who knew them and they are the first people whose work can really be considered modern in that their original works would still be read today for more than historical interest.
Twice in my life I have been housed in buildings where Sewall Wright once worked. During the first half of my doctoral program in Chicago my office was in Whitman Hall which once housed Wright's lab. The 'Sewall Wright Room' was our classroom/conference room and the place where I did my prelims. Unfortunately our department was moved out of that building and it has since been torn down. Later when I worked at the University of Illinois I worked in the Natural History Building, one of the oldest buildings on campus. Wright had apparently worked in that building during the academic year 1911-1912 when he did an MS at the U of I before heading to Harvard to work with Castle.
7:26 PM PT: Dedication
I'd like to dedicate this diary to two people who recently passed away. Julie Waters, presumably known to almost all of you, was among those who were unexpectedly enthusiastic about my bird evolution diaries and thus indirectly encouraged me to pursue more technical diaries such as this series.
My uncle (my mother's brother in law) passed away a few days ago. He was not a biologist but was scientifically trained and had a great enthusiasm for evolution.
