The Beauty of Kettlewell's Classic Experimental Demonstration of Natural Selection (2024)

Abstract

What does it mean for a scientific experiment to be referred to as “beautiful”? In a thoughtful recent analysis, Robert Crease (2003) presents a gallery of the 10 most beautiful experiments in science, carefully chosen from a set of more than 300 candidates submitted by the readership of an international magazine, Physics World. The task of his book is to analyze what it means for an experiment to be regarded as beautiful, both with reference to these particular illustrations (drawn from physics) and in general. Crease points out when experiments may appropriately be called “beautiful” by drawing attention to how the term is used in other contexts:

An experiment, unlike a painting or a sculpture, is dynamic. It is more like a dramatic performance, for it is something that people plan, stage and observe in order to produce something they are critically interested in. How can we know the circumference of the earth without stretching a measuring tape around the equator? How can we tell that the earth rotates without flying into outer space and watching, or know what's inside an atom? By carefully staging an event in the laboratory—sometimes with simple objects such as prisms and pendulums—we can make the answers appear right before our eyes. Form emerges out of chaos—not magically, like a prestidigitator pulling a rabbit out of a hat, but because of events we ourselves orchestrate. We make the world's mysteries speak.

The beauty of an experiment lies in how it makes its elements speak. (p. xviii)

While freely admitting that aesthetic judgments contain a subjective element, Crease nevertheless identifies the beauty of an experiment primarily with reference to three common elements: the significance of the results (depth), the overall conceptual simplicity and elegance of the experimental design (efficiency), and the conclusiveness of the results (definitiveness). His study of the most beautiful experiments reveals that they have an element of human drama: they are often the achievements of individual scientists undertaking painstakingly tedious work under difficult circ*mstances for years in isolation. They also include an element of luck: an accident or fortuitous choice of technique/phenomenon for study that in retrospect is recognized as especially helpful. As will be shown below, Crease's analysis makes sense of why H. B. D. Kettlewell's classic mark–release–recapture experiments are often regarded as among the most beautiful experiments in evolutionary biology.

Kettlewell's investigations

Kettlewell's life work centered on the phenomenon of industrial melanism: the rapid rise in the frequency of heretofore rare dark forms of many moth species in the vicinity of manufacturing centers throughout Britain and continental Europe, which appeared to be a consequence of the first large-scale air pollution associated with the Industrial Revolution. The phenomenon was widely recognized as an example of natural selection, owing to breeding experiments establishing that the dark form in these species was inherited and survey data documenting that the rise had occurred primarily in the vicinity of industrial areas. The rapidity of the spread, in particular, made it an ideal candidate for experimental study.

Kettlewell is famous for a set of investigations he conducted in the early 1950s to clarify why the dark forms were at an advantage relative to pale forms in soot-darkened areas. E. B. Ford, Kettlewell's mentor at Oxford University, thought the spread of the dark form reflected a physiological advantage conferred by the gene responsible for the dark form. Ford (1940) reached this conclusion on the basis of the collective wisdom of breeders who studied the genetics of melanism in the affected species, and his own backcross breeding experiments. To account for why the spread was limited to manufacturing areas, Ford drew attention to the obvious disadvantage of dark coloration against avian visual predators when the moth rests on pale lichen-covered trees in unpolluted environments (figure 1). Kettlewell's goal was to provide evidence for this second part of Ford's theory, using the best-known example of a species exhibiting the trend, the peppered moth (Biston betularia).

It is important to recognize that at the time of Kettlewell's study, no field experiments on bird predation had been conducted on industrial melanic species; indeed, Kettlewell was initially unable to find any documented evidence that birds searched for and ate cryptic insects at all. Some controversial evidence provided by J. W. Heslop Harrison suggested that melanism might be due to the mutagenic properties of magnesium salts present in the soot. (Neither Ford nor Kettlewell took Harrison's suggestion particularly seriously, although Kettlewell did attempt to induce melanic coloration in the moths by raising them on polluted foliage.)

Kettlewell began by developing and testing a scoring procedure to assess just how conspicuous the pale (typical) and dark (carbonaria) forms of the moth were when they rested on pale lichen-covered and soot-darkened backgrounds. He then conducted a preliminary experiment using captive birds in an aviary to determine whether birds ate moths at all and, if so, whether they had the same difficulty humans do when trying to spot moths at rest on their “correct”background (i.e., the background against which they are the least conspicuous). He placed moths representing both forms on pieces of pale lichen-covered bark taken from rural (unpolluted) forests, and also on soot-darkened bark taken from a wood that had been heavily polluted downwind of an industrial setting. He then released two captive Great tit* into the enclosure and monitored the results by periodic censusing and observation from a distance using binoculars. During the first two hours, the birds ignored the moths entirely, and then proceeded to take nearly all of the moths, regardless of the background against which they rested. Kettlewell conjectured that participation in the experiment might have conditioned the birds to become specialists on moths, and introduced a broad spectrum of other endemic insects to widen their feeding interests. This alteration “proved successful,” at least to Kettlewell's satisfaction, in that from this point onward he was able to establish that the birds preferentially took the more conspicuous form first.

The third and most celebrated part of Kettlewell's investigations was a mark–release–recapture experiment that was conducted first in a polluted wood near Birmingham in July 1953. In 1955, he repeated this experiment on a smaller scale in the Birmingham wood, and did a comparable experiment in an unpolluted wood in Dorset. The experiment involved three steps (figure 2). First, Kettlewell marked a population of male peppered moths containing both the dark and the pale forms, using a dab of cellulose paint on the underside of their wings. Second, he released these moths (447 carbonaria, 137 typical) on trees in an area of wood chosen specifically to reduce the possibility of migration from the test site. Third, he attempted to recapture the marked moths over several nights using a combination of a mercury vapor light trap and a series of assembling traps containing virgin females. Kettlewell predicted that, all other things being equal, if the recapture rates for the two forms were the same, this would indicate that neither form was at an advantage. However, if the recapture rates were different (e.g., if the recapture rate for the dark form was higher in the polluted wood), this was presumptive evidence that one form was better able to survive during the interval between release and recapture. During the 1953 experiment in the polluted Birmingham wood, Kettlewell recaptured 123 of the released carbonaria moths, but only 18 of the released typical moths. Thus he concluded that the recapture rate of the carbonaria form in the polluted wood (123/447 = 27.5 percent) was twice as large as the recapture rate for the typical form (18/137 = 13.1 percent). He considered and rejected several possible explanations for what might account for the difference, such as differential migration from the test site and differential longevity. This led him to conclude that the difference must be due to bird predation, and indeed on one occasion his wife actually saw a bird preying on a resting moth. The 1955 repeat of this experiment in the Birmingham wood yielded similar results, while the comparable experiment in the unpolluted wood (in which 393 typical moths and 406 carbonaria moths were marked and released) revealed a recapture rate for the typical form (54/393 = 13.7 percent) that was three times that of the carbonaria form (19/406 = 4.7 percent; figure 3). Kettlewell is widely regarded as having clinched the argument by having his associate, the distinguished ethologist Niko Tinbergen, create a film record of bird predation during the 1955 experiments (Tinbergen 1961). This film conclusively demonstrated that when presented a choice, birds in nature representing several distinct species with different foraging techniques preferentially preyed upon the more conspicuous form of the moth in both polluted and unpolluted contexts.

The beauty of the mark–release–recapture experiment

The foregoing description reveals many of the features identified by Crease as the hallmarks of beautiful experiments in science (depth, efficiency, and definitiveness), while also drawing attention to discipline-specific differences between biology and physics.

Depth

Kettlewell's experiments are significant both empirically, as an early demonstration of natural selection, and methodologically, as a demonstration of how the mark–release–recapture technique pioneered by Fisher and Ford (1947) for estimating population size might be used to estimate selective advantages in field populations. Kettlewell's work was a pioneering study in an emerging field devoted to the study of the origin and distribution of melanism throughout the animal kingdom. It is, however, striking that, in contrast to Crease's analysis of experiments in physics, this experiment was not a test of a fundamental theory. (Michael Freemantle [2003] also observed this distinction between physics and other disciplines in his survey of the 10 most beautiful experiments in chemistry.) Kettlewell's experiment was a test of whether a particular example of natural selection could be accounted for with reference to a specific mechanism or agent, namely, bird predation. If it had failed, biologists would not have rejected the theory of natural selection. (Nor would they have rejected the extensive population genetics data that provide evidence that the increase in frequency of dark forms in the vicinity of manufacturing sites is an example of natural selection.) If Kettlewell's experiment had failed, they would simply have concluded that the selective advantage might not be due to bird predation. This reflects the fact that biologists do not test universal laws; rather, they attempt to determine whether phenomena being tested fit within the range of particular models (e.g., purported forces of evolutionary change).

Efficiency

In a recent careful analysis of the Meselson-Stahl experiment, Frederic Holmes (2001) draws attention to the central role played by considerations of simplicity and elegance in accounting for why it is considered among the most beautiful experiments in biology.

The beauty of the Meselson-Stahl experiment is invariably connected with its simplicity. When reduced to its essential features, it is readily understood even by beginning students of the life sciences. Teachers look on it with fondness for the ease with which its message can be conveyed. Scientists throughout history have extolled the simplicity of nature and have admired theories and other discoveries that seem to reveal aspects of that simplicity. (p. ix)

Holmes cautions, however, that this simplicity reflects our need to make sense of the world within the limits of our cognitive capacities. His careful analysis of events leading up to the Meselson-Stahl experiment and its subsequent portrayal draws attention to how very complicated the experiment actually was, and also how the process of disseminating it to scientists and the lay public can be understood in terms of the systematic removal of complexity. As I will show, much the same can be said about the beauty of Kettlewell's mark–release–recapture experiments.

Kettlewell's experiments are conceptually simple and elegant in their design. They rest upon a straightforward, symmetrical comparison of the fates of two forms of the moth, coupled with a brilliant piece of logical reasoning by which Kettlewell used a series of simple observations in an attempt to rule out several possible alternative explanations. He repeatedly contrasted the results of his experiments in the two settings as “opposite” (figure 3), which, although rhetorically very effective, misleadingly suggests that the experiments in the unpolluted wood served as a control for the studies done in the polluted setting (Rudge 2003).

Kettlewell also drew great attention to the exquisite match between the pale moth and its “correct” pale lichen-covered background, and between the dark moth and the soot-darkened tree trunks (figure 1), both of which he often described as “beautiful.”This observation might lead one to wonder whether some aspect of the beauty of Kettlewell's experiment is simply a reflection of the beauty of the phenomenon being studied. In his study of experiments in physics, Crease denies this possibility. For instance, he quotes Michael Faraday as describing candles as “beautiful,” not with reference to their shape or color, but rather to the way our understanding of the burning of a candle calls attention to the interplay of a wide range of universal laws (Crease 2003). Freemantle likewise points out that, although elemental white phosphorus is quite beautiful, Hennig Brandt's discovery of phosphorus is not an example of a beautiful experiment: Brandt had no coherent experimental intent or expectation of what would happen, and the experiment he used had several repulsive features, including the putrefaction and distillation of urine (Freemantle 2003). This being said, with regard to Kettlewell's experiment, it seems quite clear that the reason his elegantly simple experiment yielded results is precisely because several unique features of the phenomenon of industrial melanism are illustrated by the peppered moth. The dark form is due to a single dominant gene. The difference between the two forms is obvious. The rapidity of the rise in frequency of the dark form over the space of only 50 years suggested that it reflected a rapid change in the environment. It is precisely because the phenomenon itself was apparently simple that it lent itself to a simple demonstration.

Definitiveness

Kettlewell's experiments were initially hailed as definitive demonstrations of natural selection. The magnitude of the selection advantages he observed, not to mention the film record of birds preferentially taking moths in the order of their relative conspicuousness, effectively removed all doubt. Crease's analysis draws attention to how truly important experiments in physics have played a role in illuminating something that is not obvious, in a manner that is both compelling and direct. In Kettlewell's case, previous largely theoretical work on natural selection implied that the incremental nature of the process over large periods of time would preclude direct human observation. Kettlewell and Ford recognized that the speed of the phenomenon of industrial melanism, unlike other examples of natural selection, would permit direct observations to be made. And indeed, Kettlewell's experiments were successful at providing among the first direct observations of a proximal mechanism of natural selection ever recorded.

Finally, there is an element of drama to Kettlewell's work, elaborated in great detail in Hooper (2002). Her book-length popularization illustrates the enormous work and self-sacrifice Kettlewell made in his epic, nearly single-handed attempt to demonstrate the protective value of camouflage for the moth in the two environments. She also draws attention to how Kettlewell's experimental work fit into his personal quest to establish himself as a scientist after leaving medical practice in the face of numerous personal hardships. Although Hooper does not interpret it as such, the story she shares is an archetypal example of a heroic quest, as the term is used in literature and literary criticism.

The classic demonstration of natural selection

Experiments become classics according to Crease's account once they enter the science curriculum, as indicated by their regular inclusion in textbooks. In physics, discussion of the first or clearest demonstration of an important empirical result that bears on some theoretical concept or law can help students appreciate an entirely different way of looking at the world. It provides a context against which further developments in the science, based in part on the insights of crucial groundbreaking experiments, can be understood. In short, when a physics experiment achieves textbook status, this is a direct reflection of its theoretical or empirical importance to the scientific community.

In biology, there are examples that certainly conform to this analysis. Sickle-cell anemia, for instance, has achieved textbook status as a direct result of the theoretical importance of the empirical work of Anthony C. Allison and others, which is widely acknowledged as establishing definitively that the continuing prevalence of the disease in malaria-infested regions of the world was due to heterozygote advantage. It is one of the very few cases of empirical validation for the theoretical possibility that a heterozygote might, in certain environments, be at an advantage compared to a hom*ozygote. It also has enormous potential from the standpoint of biology pedagogy, as an illustration of how a single biological phenomenon can be understood with reference to multiple sub-disciplines in biology (e.g., molecular biology, genetics, physiology, ecology, and evolution).

These considerations do not explain why the phenomenon of industrial melanism and Kettlewell's work on it have become ubiquitous in biology textbooks. Although it was the first experiment to identify a specific agent of natural selection, many other examples of natural selection existed before Kettlewell's work (Dobzhansky 1951), and better, more fully documented examples are now available (e.g., studies of the evolution of pesticide resistance in insects and antibiotics; see also Endler [1986] for numerous other documented examples of natural selection). The reason why the phenomenon of industrial melanism is regularly recounted in textbooks is not its theoretical or even empirical importance to biology, but rather several features that make it particularly conducive for the teaching of biology (Rudge 2000). These include the relative simplicity of the phenomenon—that is, it illustrates the simplest form of selection (directional), the advantage is conferred by one effect of a single gene, and the mechanism is intuitively obvious (birds will have the same difficulty spotting inconspicuous moths as humans do). Students can readily be assumed to be familiar with the major elements of the story (pollution, moths, birds). From seeing the amount of pollution deposited on trees downwind of industrial sites, to viewing the correlation between manufacturing centers and frequencies of the melanic forms, to noticing how conspicuous each form of the moth is against different backgrounds, to observing birds differentially prey upon the more conspicuous form of the moth, each of the important details of the story lends itself to a visual representation (Rudge 2003). And, perhaps just as important, each of the images in question is easily interpreted by laypeople without strong backgrounds in math and science. These points were not lost on Kettlewell, who created a movie from film clips taken by his colleague Tinbergen (1961) during the 1955 experiments, and also disseminated his investigations in numerous popular articles (e.g., Kettlewell 1959).

Problems with the textbook depiction

The brief descriptions of Kettlewell's work on industrial melanism contained in popular articles, encyclopedia entries, textbook accounts, and other venues are often profoundly misleading. In a recent book-length review of research on melanism, Michael Majerus, a well-known researcher on the phenomenon, points out multiple discrepancies between the simplified account we associate with Kettlewell and what is known today about the phenomenon. Majerus analyzes the textbook story of industrial melanism in terms of seven claims, and draws attention to how very misleading each of these claims is from the standpoint of experts (Majerus 1998; but see Grant 1999). For instance, simplified descriptions such as the one provided above often imply that the peppered moth has only two forms, a pale (light-colored) form and a dark (melanic) form. But in point of fact, as was well known to Kettlewell, the peppered moth has a second melanic form, f. insularia, which is actually regarded as a complex of forms that range in appearance from some that strongly resemble the pale typical form to those that are nearly as dark as f. carbonaria.

Kettlewell's experiment was based on the assumption that peppered moths spend the day motionless on tree trunks in plain sight, but there is relatively little consensus regarding precisely where the moths spend the day. Indeed, there is some evidence to suggest that they might rest on the undersides of branches of trees higher in the canopy, rather than on the trunks where Kettlewell put them. This consideration has made photographs of the two forms on soot-darkened and lichen-covered tree trunks particularly controversial (figure 1). Some have objected to the fact that Kettlewell staged these photographs by placing examples of the typical and carbonaria forms in proximity to one another, because, in doing so, the photograph implies that they have been captured in their natural resting positions. Indeed, some discussions of industrial melanism are illustrated with photos of dead moths pinned to trees (Ford 1981). It is quite clear from the texts in question that Kettlewell and Ford merely intended the photographs to illustrate the relative camouflage value of pale and dark coloration against lichen-covered and soot-darkened backgrounds (Rudge 2003).

Some have also objected to the large concentrations of moths used during the experiment, which in Kettlewell's opinion were necessary to ensure that differences in recapture rates, if detected, would be statistically significant. This same procedure, unfortunately, opens up the possibility that local birds are simply reacting to the elevated densities and behaving in a way that is not natural. Perhaps most problematic, Kettlewell's experiment relied on mixtures of laboratory-bred and wild-caught endemic insects, which confounded the results of his study by opening the possibility that the differences in recapture rates he found were due to differences in survivorship between lab-reared and wild-caught moths. Like almost all who have done research on the phenomenon, Majerus nevertheless concludes that the phenomenon of industrial melanism is, despite these problems, still an excellent example of natural selection and, further, that bird predation is the most important selective factor.

As one might imagine, advocates of creation and intelligent design have misinterpreted these long-recognized problems as undermining the continued use of the example in science textbooks (Wells 2000; but see also Rudge 2002). The motive in these attacks is transparent: if one can knock down the poster child of evolution, the entire theory will thereby become suspect. As noted above, the continued inclusion of the phenomenon of industrial melanism in textbooks should not be interpreted as an indication of its theoretical importance to evolutionary biology. Moreover, no one should be surprised that there are differences between the explanations given in popularizations and introductory textbooks, which are intended for children and the lay public, and what is actually known about the phenomenon as it is discussed in journal articles intended for scientists. The existence of ongoing questions about the phenomenon is indicative of an active area of research. That scientists continue to believe the explanation we associate with Kettlewell does not imply they are being dogmatic in the face of overwhelming contradictory evidence. Biologists don't abandon well-supported theories simply because they run into problems applying an idealized explanation to a particular example, particularly when no better alternative theory is available. Moreover, even if we threw out Kettlewell's experimental work on bird predation entirely, it wouldn't threaten in any way the enormous genetic evidence for natural selection in the evolution of melanism, nor the eight or so follow-up predation studies on the mechanism that qualitatively support Kettlewell's original conclusions (Cook 2000, 2003).

Similar discrepancies surround our current understanding of Kettlewell's work on the phenomenon. In the first book-length popularization of this episode, Judith Hooper (2002) portrays Kettlewell as a bumbling amateur and all but accuses him of committing fraud. The latter charge is entirely baseless (Grant 2002, Rudge 2005). No one would deny that, judged by contemporary standards, the design and conduct of Kettlewell's experiments are problematic (Shapiro 2002). Nor would even Kettlewell deny that his work on the subject involved a number of false steps and errors of interpretation: his first scientific publication of the results of his study is remarkably candid about the numerous problems he encountered in the field (Kettlewell 1955). But do these considerations imply that Kettlewell's experiments should no longer be used in science classrooms? The answer, I believe, is a decided no. Discrepancies between the textbook accounts of Kettlewell's work and what scientists now know about it represent a beautiful opportunity for science instructors to help students appreciate a variety of issues associated with the nature of science, such as the tentative nature of scientific conclusions (Rudge 2000, 2004a, 2004b). The same may also be said for other examples in the history of biology, such as A. C. Allison's work on sickle-cell anemia (Howe and Rudge 2005).

Conclusions

Current scientific controversies surrounding industrial melanism, and Kettlewell's work on it, do not call into question the fact that it is an example of natural selection. Nor do they support the contention that it should be removed from the science classroom. The simplified account we associate with Kettlewell is a particularly lucid illustration of natural selection and how scientists study it. Comparing and contrasting this simplified account with what is actually known represents a wonderful opportunity for students to learn about the nature of science.

This being said, it is fascinating to consider the implications of these reflections on our evaluation of the beauty of Kettlewell's experiments. As mentioned above, Crease's analysis identifies the beauty of an experiment with the significance of the results, the overall conceptual simplicity and elegance of the design, and whether the results are regarded as conclusive. But as we've seen, the many features that recommend Kettlewell's experiments as among the most beautiful of experiments in evolutionary biology reflect how Kettlewell and others since have depicted them in popularizations and textbook accounts. Inclusion of the phenomenon of industrial melanism in textbooks is generally misinterpreted as indicative of its theoretical importance to biology. The elegant design we admire is only apparent, an artifact of the systematic removal of detail surrounding the historical context of the experiments and the problems Kettlewell encountered in trying to interpret his results. Portraying the experiments as definitive (aside from the incontrovertible evidence furnished by the film record) ignores the important role a host of follow-up studies have played in securing the basic conclusions Kettlewell drew from his results. These considerations don't undermine Crease's basic analysis. Rather, they point out that at a fundamental level, the beauty of an experiment may not rest so much with the experiment itself as with how it is subsequently portrayed.

Acknowledgements

The author gratefully acknowledges the helpful comments of Richard M. Burian, Uric C. Geer, Bruce Grant, Eric M. Howe, J. R. G. Turner, and an anonymous reviewer on an earlier version of this manuscript. The author thanks Geoffrey O'Brien of the British Universities Film and Video Council for helping him secure a copy of the film Evolution in Progress for research purposes. Attempts to locate a survivor of Kettlewell's family who might be able to grant permission for use of still photos from this film were unsuccessful.

References cited

Figure 1. Biston betularia: one typical (pale) and one carbonaria (dark) form resting on a lichen-covered tree in unpolluted country (Dorset; left) and on blackened, lichen-free bark in an industrial area (the Birmingham district; right). Photographs originally appeared separately as plates 14 and 15 in Ford (1975).

Figure 2. Images of H. B. D. Kettlewell marking, releasing, and recapturing moths in connection with his 1955 experiment in Deanend Wood, Dorset, taken from a film record of his experiments entitled Evolution in Progress (Tinbergen 1961).

Figure 3. Bar graphs depicting the results of Kettlewell's 1955 and 1956 mark–release–recapture experiments in polluted and unpolluted country. Originally published as figure 1a in Kettlewell (1956, p. 289) and used with permission from Nature.

Author notes

1

David W. Rudge (david.rudge@wmich.edu) is an assistant professor in the Department of Biological Sciences at Western Michigan University, Kalamazoo, MI 49008. He is currently writing a book on H. B. D. Kettlewell's famous experiments on industrial melanism, with special attention to how the episode has been portrayed in science textbooks.

© 2005 American Institute of Biological Sciences