Return to History Preamble

A Brief History of Evolutionary Genetics

Part 6 :  Dobzhansky

©Paul Handford, 1998

Dobzhansky's career exemplifies much that went on in evolutionary thinking from the 30s to the 70s

After the three 'giants' of the earlier era of mature theoretical evolutionary genetics, bringing us up to the time of the Second World War, many more important players come on the scene, too numerous to discuss in these brief notes.  But by anyone's standards, one of the most influential evolutionary geneticists from the times of Fisher, Wright and Haldane up through the Evolutionary Synthesis (see below) and on to modern times was Theodosius Dobzhansky.  There are several other very important contributors to our developing view of evolutionary populations, such as Ernst Mayr, G.G. Simpson and G.L. Stebbins, but we may single out Dobzhansky for two important reasons:  because one could easily substantiate a claim that he was truly the most influential of all the biologists of the Evolutionary Synthesis, through his multi-edition text (note 1);  and  because he provides a clear and fascinating illustration of a major trend in evolutionary genetical thinking from the thirties through to the recent past.  This concerns the importance accorded to random factors in evolutionary change.  In brief, in the space of 40 years - from the 30s to the early 70s - 'evolution by random drift' went from broad favour as part of evolutionary explanation to general, and pretty firm, disfavour.

This general shift in "science's attitude" towards the role of chance and the prevalence of non-adaptive characters may be seen in the writings of many of the major contributors to the Evolutionary Synthesis:  Wright himself, G.G. Simpson, Ernst Mayr, as well as in many rather more minor figures e.g. David Lack (note 2).  The whole trend has been dubbed "the hardening of the Modern Synthesis" (note 3)   This change appears to be only partly attributable to a changing empirical picture - it was to some good degree due to changes in "what seemed reasonable," and what yielded "desirable" consequences.

In our consideration of this sea-change in the general attitude of the evolutionist community, bear in mind three things:  first, that hardly anyone ever saw randomness - drift - as the exclusive, or even the major, agent of evolutionary change:  it was always regarded as at most a complement to selection, even by its enthusiasts.  Second, that the shift concerns attitudes to the balance of importance of drift and selection:  it does not concern absolutist claims for the exclusive importance of either, even on the part of all but the most rabid pan-selectionists - to this extent, evolutionists have always been, and remain, pluralist in outlook.  Rather, they have become, in Gould's words, 'polarized.'  Finally, we should clearly distinguish changing attitudes to processes going on within species (anagenesis) from those involved in the generation of new lineages (cladogenesis);  it was (and is) possible to see anagenetic change as being entirely driven by selection while at the same time affording a major role to non-adaptive processes in lineage splitting.  Much of what will be discussed in this section on Dobzhansky concerns the period of the Evolutionary Synthesis itself, but it will be useful to follow a particular theme with him before looking at the Synthesis in general.

Dobzhansky came from the naturalist tradition

Unlike any of the figures considered to this point, Dobzhansky had very limited mathematical skills (one of the reasons why he engaged in an extensive collaboration with Wright);  he was, by contrast, both by disposition and by training, a naturalist, interested in wild organisms for their own sake (even though he was also strongly motivated in his studies by the larger questions of the meaning of evolution in human society and ethics - of this more later.)  Like many such naturalists, whether amateur or professional, Dobzhansky became aware, early in his career, of the enormous variability among individual organisms of most taxa, and it is almost certainly no accident that most of these naturalist-biologists were also generally persuaded of the pervasiveness of adaptation - they commonly viewed much, if not most, of the variation - among individuals, populations and taxa - as representative of adaptive fit to circumstances.  Dobzhansky's view of nature was thus not the filtered, ordered and codified view of the fixed, defined, "types" considered by museum workers;  he, as so many others before and since, had the uncomfortable reality of exuberant diversity impressed indelibly upon him.  As we have seen earlier, this variability has struck the gamut of biologist as of greatly varying significance in the scheme of things:  to some the variability was inconvenient "mere fluctuation;"  to others, it was the stuff of evolutionary reality and possibility.

The great importance of population variability is explicitly acknowledged

One of the major contributions made during the entire Evolutionary Synthesis period (1940-1955) was the explicit, and central, recognition that species are composed of populations of variable individuals.  This keystone acknowledgment of the centrality of variability (which may, in the long view, be seen as the fruits of a 2000-year struggle against the dead weight of Platonic ideal forms - the return of scientific thought to the empirical attitudes of Democritus and the like after its long and distracted wanderings in the trackless wilderness of medieval scholastic sophistry) was to some large extent, the work of Dobzhansky, his fellow naturalists Ernst Mayr and G. Ledyard Stebbins, and the decidedly field-oriented paleontologist, G.G. Simpson.

Before he moved to the United States in 1927, to work in T.H. Morgan's Drosophila laboratory, Dobzhansky worked on ladybird beetles (ladybugs), and it is not an exaggeration to say that during his studies of their diversity across Eurasia, he conceived an enthusiasm, one might even say a love, for variation.  He was very well aware, like Darwin before him, that evolution depended on variation, and that continuing evolution depended on continuing variation.  Thus he was very concerned to understand how selection could happen, yet not stagnate through using up the variation upon which it worked.  Let us visit this problem in 1937, in his own words:

How adaptive variation could be maintained in populations under selection was a major problem

How should a species maintain variability of selectable traits (of adaptive significance) with which to face the unpredictable exigencies of future environments?  That was his compelling conundrum, and, given the general climate of thought in the 30s (as described in section 4 on Sewall Wright) it is perhaps no surprise that he seized on drift as his answer.  But why should he do this, given his general enthusiasm for adaptation?  -it was because the prevailing view of selection at the time was that it functioned to improve some globally-relevant adaptive fit through weeding out lower-fitness variants (as most variants were imagined to be), thereby generating genetic near-uniformity in natural populations of most organisms.  This view of selection, and the population constitution it generates goes by the name of the "classic model" and it will be discussed later and in the next section;  for now we can appreciate it as an alternative which Dobzhansky regarded as unwelcome - for he knew variation was abundant and ubiquitous, and this "classic" kind of selection seemed to remove it.

One may see this posture of Dobzhansky's as, in some senses at least, incoherent.  His naturalist's experience and intuition led him to suspect that geographic variation corresponded to significant variation in the environment i.e. that geographic variation was adaptive variation (a point of view he explicitly espoused early on in his ladybird days);  yet his variation-fostering drift notions suggest, rather, that much of such variation is non-adaptive.  It seems as though, despite his selectionist hunches,  he became persuaded by his initial failures in trying to demonstrate correlations between variants and environmental conditions in Drosophila, and cleaved to drift as the simplest explanation of this;  it also had the (to him) pleasant corollary of permitting variation to persist until such time as it might become of adaptive significance.

Incoherent or otherwise, it was his explicit posture in his 1937 text that substantial amounts of variation among individuals and among populations was non-adaptive, neutral, and due to drift.  But by the time of his 1951 edition, he had switched into an almost complete disregard for the general importance of drift.  We may ask:  why?  In short, the answer is that a)  some of the standard examples of drift-based variation (including some he had worked on himself in Drosophila) came to be shown to be of clear selective importance, after all, and b)  Dobzhansky found a way that continuing variation (his beloved variation!) could be reconciled with selection, albeit selection of a particular kind.

Consideration of this problem involves a major debate:  the "classic - balance" argument

The empirical finding that selection was, after all, involved in some well-studied systems previously considered as showcases of neutrality, though of course interesting and important, could hardly provide the basis for a complete shift of emphasis in interpretation of evolutionary processes.  Beatty (see bibliography) provides a persuasive case that it was not simply the empirical information so much as the changing theoretical conceptions of how selection might work (and lead to sustained population variation) that provided the primary explanation of Dobzhansky's major change of heart.

Beatty points out that Dobzhansky had a long-standing rivalry with H.J. Muller, who was the architect and champion of the "classic view" of selection and population structure.  This was the variation-destroying view of selection, remember, for which Dobzhansky had little sympathy - specifically, the notion of a single optimal ("wild-type") allele at each locus, giving a single optimal state in each character, giving a single optimal phenotype in the species as a whole.  Muller's ideas on selection's role came from his work (for which he got a Nobel Prize) on mutations using X-rays (such mutants are indeed likely to be deleterious;  but given our curiosity about the origins of ideas used in evolutionary science, we might wonder what else in Muller's life suggested his view that selection functioned primarily in purifying the species, so as to produce a standard, superior, phenotype - one is tempted to speculate about his background in German-style biological philosophy, which was strongly influenced by the idea of ideal forms, and other Platonic notions. )

Dobzhansky's mature view, on the other hand, was that selection favoured genotypes that were highly heterozygous;  this he referred to as the "balance view" under which there was no single set of optimal alleles.  Clearly, such a heterozygote-superiority system would actively maintain allelic variability within the population (and was thus superior to the 'drift-solution,' which only tended to keep variation around.)  It was the 'discovery' of this variation-maintaining style of selection that, according to Beatty, converted Dobzhansky from a proponent of drift as an important agent in evolutionary populations to a strong enthusiast for selectionist scenarios:  in Gould's terms, he 'hardened' from a pluralist to a firm selectionist, albeit an enthusiast for a particular type of selection.

The origins of the "balance" view are somewhat murky

Now, where did Dobzhansky come by his Balance views?  whence the conviction of widespread heterozygote superiority of fitness?  By now, we may be less than astonished to discover that the empirical base for the change of view was slim.  There was a theoretical basis for the expectation of heterozygote superiority in the work of Fisher on the evolution of dominance, articulated in the 1920s and 30s (note 5).  Briefly, the notion is that deleterious mutants will necessarily exist primarily in the heterozygous state;  there, any variation in the rest of the genome which causes the deleterious effects of the allele to be lessened will be selected positively;  thus, given the existence of such "modifying" genetic variance, the deleterious allele will evolve into recessiveness.  Fisher came by this idea in attempting to explain the general recessiveness of deleterious alleles.  Applying similar reasoning to an allele's beneficial pleiotropic effects leads to the overall suggestion that the beneficial effects of alleles should evolve into dominance, while their deleterious effects should evolve into recessiveness.  Thus, heterozygotes would manifest the beneficial effects of both alleles at the locus, but none of their deleterious effects, i.e. heterozygotes would have a fitness superiority over homozygotes, on average.  This theory concerning the evolution of dominance was, incidentally, another large source of disagreement between Fisher and Wright.

But just as there was no compelling empirical evidence for widespread heterozygote advantage itself in nature, there was equally slim empirical support for the evolution of dominance modification.  So whence did Dobzhansky's conviction derive?  Beatty believes that it is primarily on the basis of his being convinced that variation was near-ubiquitous, and seeing balancing selection as the best bet for generating it.  Moreover, it seems clear that Dobzhansky saw variation not only as there, but as good - he not only liked variation because it permitted a continuing adaptive response to the environment, but because variation had an intrinsic value to him - and especially in Homo sapiens.  Let's hear his own words on this theme, in a passage explicitly considering the fate of human populations facing rapid environmental change:

As Beatty says at this point:  "Assuming that humans, like ladybug and fruitfly species, had sufficient intraspecific variation to ensure their adaptability, Dobzhansky judged the future of the human species to be quite rosy.  His world was thus a happy one."  It is amusing to note here that, while Fisher and Dobzhansky were thus both preoccupied by the evolutionary fortunes of humanity, their views of what properly constituted that humanity, and what was of value in it, were radically different and so, perhaps not surprisingly, were their "solutions" to the "problem."

More immediately, we may thus see that the Classic view was a direct threat to Dobzhansky's conception of what was valuable in humanity (in any population of course, but, for Dobzhansky, especially in human ones.)  In particular we may note that a more serious threat still came from the eugenic policies which are based on a "classic" view of selection.  And on this particular theme, we must take especial note that Muller, who, as we have seen, was Dobzhansky's long-time arch-enemy in evolutionary theorizing, was an ardent eugenecist (albeit a Marxist one!) who felt that civilization had permitted the accumulation of a potentially insupportable "genetic load" (he coined this term) of mutations.  Dobzhansky had long been an enemy of eugenic attitudes;  as an example of this explicit posture, here are his thoughts, from the first edition of Genetics and the Origin of Species:

Later, in his 1962 book, Mankind Evolving, in confronting Muller's concepts of "genetic load" and his eugenic solutions, Dobzhansky says: Here, perhaps, we approach a fundamental animator of Dobzhansky's particular enthusiasms, first for drift, turning eventually into (balance) selection:  they both promote continuing variability, which was a population's best guarantee against extinction through a failed race with the environment;  thus, on his view, genetic variability is to be seen as a benefit, not as a "load."  But let's not forget that, from his naturalist's vantage point, he really knew that such variation was indeed there, and he did have ample empirical evidence that selection was often involved.  However, one wonders if perhaps there are not other factors, more deeply-buried - for he clearly believed that variation was not only a useful attribute in the evolutionary drama, but that it was a good thing.

The period of ideas we have just been surveying, then, (mostly from the 30s to the 50s and early 60s) saw the general decline in favour of drift as a major player in evolutionary processes.  As noted, this was in part due to some signal successes in showing selection in natural polymorphisms, but also in no small part due to the enormous influence of the writings of Dobzhansky, who came himself to adopt enthusiastically a particular view of how selection worked, and what it produced.  It is a considerable irony that, a few years after Dobzhansky's fights with Muller and the Classic view, two things happened:  Lewontin & Hubby (note 6) began to open the floodgates of data on molecular variation in natural populations which demonstrated unequivocally that enormous quantities of genetic variability indeed do exist in nature;  but it also became clear, and acknowledged, that simple heterozygote advantage could not be sustained as a global explanation of it.  Other explanations were needed;  one of them, ironically perhaps, was - guess what? - neutral evolution by random drift;  but there were other types of selection to be found, too, such as niche-dependent and frequency-dependent fitness.  As we shall see, most of the old arguments continued (and continue still):   about the prevalence of neutral variation, about the heritability and adaptiveness of individual and population-level variation, about the long-term evolutionary importance of population differentiation, about the adaptiveness of species differences and the speciation process.


Note 1  Dobzhansky, T.  1937  Genetics and the Origin of Species  Columbia Univ. Press.
                  ___________   1951  ditto  3rd. edition.   Return to text.

Note 2  In his monographic treatment of Darwin's Finches (published in 1945, but written in 1940), David Lack argued that most of the specific and subspecific differences were non-adaptive, while in his popular  account, published in 1947, that they were mostly adaptive!  It is especially noteworthy that just prior to his first work, Lack had extensive discussions on speciation with Ernst Mayr, and concluded non-adaptiveness for his finches.  Mayr had at the time (and for long after) an enthusiasm for "the founder principal"  (a non-adaptive mechanism based in sampling error) as an explanation for the often-divergent appearance of island taxa, and no doubt influenced Lack.  But  during the war, Lack reconsidered, changed his mind, and said so in his 1947 book.  Return

Note 3  Gould, S.J.  1983  The Hardening of the Synthesis.  in Grene, M. 1983  Dimensions of Darwinism.  Return

Note 4  Dobzhansky, T.  1937  Genetics and the Origin of Species  Columbia Univ. Press.   Return 

Note 5  see for example:
Fisher, R.A.  1928.  American Naturalist  62:  115-126;  571-574.       1929.  American Naturalist  63:  553-556.    1930  American Naturalist  64:  385-406.      1934.  American Naturalist  68:  370-374.  Return 

Note 6   Hubby, J.S. & R.C. Lewontin  1966  Genetics  54:  577-594;  Lewontin & Hubby  1966  Genetics  54:  595-609.  Return

Go to next section