Before reviewing the criticisms aimed at the peppered moth case and their validity, two points should be made. First, almost all of the criticisms of the case as an example of evolution in action are aimed at Kettlewell’s work and the role of his mentor Professor E. B. Ford. There is little reference to the independent work on the peppered moth over the last four decades of the twentieth century (Cook 2003; Majerus 1998; Lees 1981; Brakefield 1987 for reviews). Such work is only mentioned when carefully selected passages, often taken out of context, are used to support criticisms of the case. Second, there is no mention of the many other species of moth (over 100 species in Britain alone) that exhibit industrial melanism (Kettlewell 1973; Majerus 1998 for reviews; Fig. 3).
The criticisms of the peppered moth case can be broadly split into four categories:
-
1.
Issues of ignorance of the ecology and behavior of the moth
-
2.
Issues of artificiality or poor procedure in experimental protocols
-
3.
Pseudoscientific criticisms
-
4.
Data manipulation and/or scientific fraud
Issues of Ignorance of the Ecology and Behavior of the Moth
Most of the criticisms in this category have come from scientists working with the peppered moth, including Kettlewell himself. Perhaps preeminent among the problems in this category have been difficulties in understanding the dispersal dynamics of the moth, lack of knowledge of the longevity of the moth, and a paucity of observations of peppered moths at rest in the wild during the day.
The main elements of the dispersal characteristics of the peppered moth are now accepted to comprise the flight of adult male moths, the more limited flight of adult female moths, and wind-assisted movement of newly hatched larvae on silken threads (Liebert and Brakefield 1987; Brakefield and Liebert 1990). These dispersal patterns and particularly the movement of first instar larvae are important in explaining the frequency drift of the melanic allele into rural regions to the northeast of industrial centers: Larvae are blown by southwesterly prevailing winds in the “aerial plankton.” Adult longevity in the peppered moth and the temporal dynamics of reproduction through adult life in this species were assessed by Bishop (1972).
The paucity of observations of peppered moths in their natural resting sites has often been highlighted (e.g., Clarke et al. 1985; Wells 2001; Hooper 2002). For example, Clarke et al. wrote, “all we have observed is where the moths do not spend the day. In 25 years we have only found two betularia on the tree trunks or walls adjacent to our traps and none elsewhere” (Clarke et al. 1985). This is important, because many have assumed that, as Kettlewell released moths onto tree trunks, this is where peppered moths spend the day. In fact, Kettlewell (1958b) suspected that peppered moths did not usually rest by day in exposed positions on tree trunks, for he wrote, “whilst undertaking large-scale releases of both forms in the wild at early dawn, I have on many occasions been able to watch this species taking up its normal resting position which is underneath the larger boughs of trees, less commonly on trunks.” It seems likely, therefore, that the reason that Kettlewell released his moths onto tree trunks was simply experimental expediency: on tree trunks, it would be easier to see what was going on.
There is now considerable circumstantial evidence from cage experiments that peppered moths do not usually rest in exposed positions on tree trunks but prefer horizontal branches (Mikkola 1979, 1984; Liebert and Brakefield 1987). Field observations of peppered moths found serendipitously lead to the same general conclusion (Howlett and Majerus 1987; Majerus 1998). The largest data set of peppered moths found in the wild was accumulated during a predation experiment that involved researchers climbing trees at dusk and dawn during the flight season of the moth (May to August) over 6 years. Of 135 peppered moths found, 50% were on horizontal branches (Fig. 4), 37% on trunks (Fig. 5), and 13% were on smaller twigs or in foliage (Majerus 2007). Therefore, although Kettlewell’s predation experiments have been criticized as being artificial because he released them onto tree trunks, it appears that this element of his protocol was not as flawed as some (e.g., Majerus 1998; Wells 2001) have previously thought.
One question that must be addressed in the light of this finding is, why did Clarke and his colleagues found so few peppered moths at rest during all their years of research? My opinion is that they, like most humans, are simply not very good at seeing peppered moths in their naturally chosen resting positions. There is some circumstantial experimental evidence to support this view. An experiment was conducted to test the efficiency of the technique used in some predation experiments that entailed gluing dead moths onto trees in “life-like” positions (e.g., Bishop 1972; Howlett and Majerus 1987). Peppered moths of each of the different forms, set with their wings in natural resting postures, were glued onto birch trunks with a view to maximizing their camouflage. An equal number of live moths were then released onto the same trunks at dawn and allowed to walk up the trunks until they clamped down. Students were then asked to walk toward the birch trunks from 10 m away, having been told that there were six moths to find. When one meter from the trunk, the subjects had to stop and could continue to search for the moths for one further minute. The experiment produced two major conclusions. First, the live moths were significantly harder to see than the dead glued moths. Second, none of the students found all the moths despite knowing that the moths were present in a very restricted area, indicating that people with little experience of looking for cryptic moths are not very good at spotting them (Majerus et al., unpublished data).
Issues of Artificiality or Poor Procedure in Experimental Protocols
There have been many criticisms of Kettlewell’s experiments beyond that concerned with his releasing moths onto tree trunks:
-
1.
In his predation experiments, he often repeatedly used the same trees to release the moths onto, potentially producing a “bird table effect”.
-
2.
In his mark-release-recapture experiment, he released moths at high frequencies, which may have produced an area of local high prey abundance, and birds may have developed a searching image for peppered moths, leading to abnormally high levels of predation.
-
3.
In both types of experiment, he released moths during daylight hours. Peppered moths are very reluctant to fly in daylight. Those prompted to fly in the day will clasp the first substrate that they contact and within a few centimeters settle fully, “clamping down” against the substrate and then staying still. In consequence, moths released in the day may not select the same sites as those that come to rest at the end of their night activity. It seems unlikely that the level of crypsis that would have been achieved by Kettlewell’s released moths would have been as great as that of wild moths.
-
4.
The forms of the moth were not released at natural frequencies. If predators, such as birds, had already formed a searching image for the form that was most abundant at a site, this might bias the results (although notably in the opposite direction to the results obtained).
-
5.
The moths that Kettlewell used in his experiments included moth trap-caught wild moths and reared moths. These may have different behaviors, and there exists no record to show whether Kettlewell kept track of the origins of his moths so that he could conduct analysis to look for differences between them.
-
6.
Some of the moths that Kettlewell used in his Birmingham and Dorset experiments probably did not originate in areas where they were used. This is thought to be the case because typica was scarce at the Birmingham site (10%) and carbonaria was absent from the Dorset site at the time (Kettlewell 1973). It is possible that moths develop local adaptations (Howlett and Majerus 1987; Majerus 1998), and so translocated moths (likely to be the typica used in Birmingham and the carbonaria used in Dorset) may have both behaved differently and been under increased stress.
Each of these criticisms, with one exception, was addressed in one or more of the subsequent independent experiments (Cook 2003 for review), prior to Coyne’s review in 1998.
The one criticism of Kettlewell’s protocol that was not addressed in any of these experiments was that he released moths in the day and so did not allow live moths to select their daytime resting sites at the end of their night flights. This issue has recently been addressed in a predation experiment near Cambridge. The aim of this experiment was to determine whether differences in the levels of bird predation of the typica and carbonaria forms could explain changes in the frequencies of these forms over a period of years.
From 2001 until 2007, the frequencies of the forms of the peppered moth were monitored by trapping at Madingley Wood, to the west of Cambridge. From 2002–2007, a predation experiment was conducted at a site 1.9 km from Madingley Wood. The experiment was specifically designed to rectify flaws in Kettlewell’s predation experiment protocols (Majerus 2005). Thus, the moths used all originated within five km of the experimental site. Moths were released at low densities and at natural form frequencies for the area. Moths were released onto different parts of trees (103 different release sites) in the ratios that the moths used these parts in the wild. The experiment was conducted each year throughout the moth’s flight season. The origin and sex of each moth was recorded prior to release and data for males and females, lab-bred or wild caught, were analyzed separately. Finally, and most crucially, the moths were released into large cages on the trees at dusk (one moth per cage), with the cages being removed in 40 min before sunrise the following morning. Moths were then observed over a four-hour period and incidences of bird predation recorded (Majerus 2005). After four-hours, any remaining moths were recollected and recorded. The results of this experiment showed that the frequency of carbonaria declined from 12% of the carbonaria + typica population in 2001 to just over 1% in 2007. This is equivalent to a mean selection coefficient of 0.29 against carbonaria over this period. In the predation experiment, proportionately more carbonaria were eaten than typica, the difference being equivalent to a selection coefficient of 0.22 against the black form. The difference between these selection coefficients is not statistically significant. The conclusion from this experiment is that differential bird predation of the forms is sufficient to explain the changes in the frequencies of the forms in Cambridge between 2001 and 2007 (Majerus 2007).
Pseudoscientific Criticisms
Few of those who criticize the peppered moth case as an example of Darwinian evolution in action have ever worked on the moth. Moreover, few have experience as field biologists or training in either evolutionary genetics or ecological entomology. Their criticisms of the case, when erroneous, can thus be excused, at least in part, simply because they have little understanding of the ecology of the moth and its predators, or of how natural selection operates. Yet, although the vacuous nature of some of the criticisms is excusable, they do create a significant problem, because many of the readers of these criticisms, particularly those published in newspapers and on the web, do not have the scientific knowledge or experience to objectively appraise the criticisms.
Take, for example, the first sentence in Chapter 1 of Hooper’s Of Moths and Men (Hooper 2002): “To begin at the beginning, the Lepidoptera are divided into two orders: butterflies (Rhopalocera) and moths (Heterocera).” Those who have no experience of entomological classification may not realize that this first sentence is simply wrong. There is a single order of insects called the Lepidoptera (meaning scale wings) to which the butterflies and moths all belong. Within this, those that we call butterflies comprise three superfamilies (Papilionoidea, Hesperioidea, and Hedyloidea) within the order (e.g., Scoble 1992). They are not the most primitive, nor the most advanced in the order. This is a trivial point but illustrates well that, without the necessary background, it is difficult to cogently evaluate the critics of this case. In addition, many of the critics are good writers and are well practiced in arguing their case persuasively.
An example from Hooper (2002), concerning peppered moths and bats, illustrates the problems that nonspecialists have in commenting sensibly on a case such as that of the peppered moth. Prior to publishing Of Moths and Men, Hooper e-mailed me (November 16, 2000) to ask questions about the peppered moth case. One set of questions concerned the role of bat predation and was asked in the context of Hooper’s view that Kettlewell thought that 90% of the predation of adult moths was caused by bats. These questions have previously been published verbatim (Majerus 2005).
I explained my view in a telephone conversation on November 19, 2000. I cannot remember exactly what I said, but my notes on Hooper’s e-mail were that Kettlewell was right that the different forms were unlikely to differ in taste or smell but that scale types and pigments might affect sonar. I also explained why Kettlewell’s reasoning was logical and pointed out the flaws in Hooper’s argument by theoretical example. Hooper (2002) cites this example (accurately I believe); “Say three hundred eggs are originally laid. Once you get to the adult stage, maybe you have ten left. Of these more than half are killed by things not hunting by sight, so say you have four moths left—two typical and two carbonaria. You must be prepared to say that none of the mortality prior to this is due to selection on colour pattern, no pleiotropic effects of alleles, no differences in palatability, no greater energetic costs in producing black pigment and so on. If so, then despite 296 moths being killed up to that point, if those two typicals are eaten by birds, you’ve increased carbonaria by a hundred per cent at one go.”
This simple example was intended to clarify Hooper’s obvious misunderstanding of how selection works, as manifest in part of her e-mail, where, in the two questions she asks, she couples and muddles percentage mortalities (10% selectively by birds, 90% randomly by bats), with the 2:1 and 3:1 selective advantages to one form or the other in Kettlewell’s experiments. Yet, my intent was obviously not achieved: I had overestimated Hooper’s grasp of how natural selection works. To clarify my clarification, I here spell out my reasoning in more detail.
The female moth that laid 300 eggs was either carbonaria or typica and had mated with a male of the other phenotype. As the difference between carbonaria and typica is due to a pair of alleles of a single gene, with the carbonaria allele being fully genetically dominant to the typica allele (Bowater 1914), the carbonaria moth in this pairing would have to be heterozygous to produce equal numbers of carbonaria and typica offspring. This means that of the 600 alleles present in the 300 eggs laid by this female, three quarters (i.e., 450) would have been typica (two in each of the 150 typica eggs, plus one in each of the 150 carbonaria eggs) and one quarter (i.e., 150) carbonaria (one in each of the 150 carbonaria eggs). By the time we reach the final four moths (two typica and two carbonaria), 444 typica and 148 carbonaria alleles will have been randomly eliminated, leaving just six typica alleles (two in each of the two typica and one in each of the two carbonaria moths) and two carbonaria alleles (one in each of the two carbonaria). If the two typica moths are eliminated before they have reproduced, by birds preying selectively, the remaining two moths that breed and pass on their genes will both be heterozygous for the carbonaria and typica alleles. Thus, now, half the alleles that are passed on are carbonaria. So the final clause in my example: “you’ve increased carbonaria by a hundred per cent in one go”, is realized (from 25% to 50% for this family).
Hooper (2002) wrote that Kettlewell’s view was that if bat predation did account for 90% of the predation of adult moths, it did not matter, “because bat predation wasn’t differential predation; evolution was driven by the small percentage of moths that are eaten selectively by birds hunting visually”. So, although Hooper failed to understand my explanatory example, Kettlewell certainly would have.
Hooper’s e-mailed questions, “Wouldn’t it be wrong to assume that bat predation was totally random?” and “Would a good scientist need to do an experiment to rule out selective predation by bats?” (Majerus 2005), caused me to design just such an experiment. In four tests in which equal numbers of flying peppered moths of the two forms were made available to bats, bats did not favor either form over the other. In total, bats were seen to eat 211 typica and 208 carbonaria (Majerus 2008). The results support Kettlewell’s view that although bats may cause significant mortality in B. betularia, they do not prey selectively with respect to B. betularia forms.
It is worth noting that although Hooper (2002) asserts that “Kettlewell himself admitted that they {bats} probably accounted for 90 per cent of the predation of adult moths.”, this was almost certainly not the case. Hooper’s reference for this assertion is a letter from Kettlewell to B.J. Lempke, in June 1959. In this letter, Kettlewell wrote: “No one would be foolish enough to argue that your statement ‘The greatest enemies of moths are not the birds but the bats’ is untrue, but... their predation is not selective. It does not matter the slightest if bats take 90% of a species population at random on the wing but if birds...account for the other 10%, but do so selectively...”. Hooper interprets these figures of 90% bat predation and 10% bird predation—used as explanatory example to illustrate a specific point—to be what Kettlewell believed the actual levels of predation by bats and birds were. Yet, it is clear from Kettlewell’s writings on the peppered moth that he had a very thorough knowledge of the behavior of this species and did not believe that bats, which feed mainly on flying insects, were responsible for 90% of the predation of B. betularia adults. This is, for example, manifest when he notes that, in a resting site selection experiment, he and Conn used only females, “which do not normally fly” (Kettlewell 1973, p. 88, footnote). Here, Kettlewell preempts the findings of Liebert and Brakefield, who demonstrated that female B. betularia rarely fly, except for a single dispersal flight following mating, usually on the second night after eclosion (Liebert and Brakefield 1987). That Hooper gets so much wrong in this small facet of the peppered moth story may be a function of the agenda that she was writing to, or, more probably, was a result of Hooper understanding neither how natural selection operates nor how peppered moths behave and interact with their predators.
My own view, based on both literature and field observation, is that while bat predation of day-cryptic night-flying moths in the summer is considerable, it will differ in level between the sexes of a species because of the variation in the amount that the sexes fly. In the peppered moth, I doubt that bat predation accounts for as much as 90% of total predation on adult males, and I am certain that it accounts for only a small proportion of the total predation of females.
Data Manipulation and/or Scientific Fraud
Finally, we come to suggestions that Kettlewell designed his experiments, “in order to come up with the right answer” (Matthews 1999), or changed his experimental protocols or data in a deceptive manner (Hooper 2002). Rudge (2005) considers Hooper’s accusation of scientific fraud in considerable detail. While not wishing to repeat his deft, surgical dissection of Hooper’s flawed agenda (“She decided in advance that she wanted to tell an entomological whodunnit.”), three major points in Rudge’s arguments are worth detailing. First, he notes that, “among the many scientists who have worked on the phenomenon over the last 50 years, neither Kettlewell’s colleagues, nor his severest critics, nor researchers since have ever alleged that he committed fraud: nor has any historian of biology.”
Second, he points out that one of Hooper’s strongest accusations was that Kettlewell changed his release procedure (increasing the number of moths released) to try to increase the numbers of marked moths that were recaptured and ties this to her own unsubstantiated interpretation of a letter that Ford wrote to Kettlewell on July 1, 1953. She wrote, referring to Kettlewell’s 1955 paper, “there is a table, Table 5, listing releases, catches and recaptures for each day. Squinting at the columns of numbers, we notice a strange thing: from 1 July on, after the letter from his boss, the recaptures suddenly soar.” Both Rudge (2005) and Majerus (2005) note that the timings of the changes in procedure do not fit with this interpretation. Kettlewell changed his protocol on June 30, presumably in response to the low recapture rates that he had informed Ford of. Moreover, the increase in recapture rates was first recorded on July 1 and would represent the moths found in his traps early in the morning following the night of June 30/July 1, i.e., before Ford had written his letter. Third, from his comprehensive review of Hooper’s accusation that Kettlewell committed fraud, in which he examined her book, the sources that she cites and many other sources of the peppered moth that Hooper fails to cite, Rudge (2005) concludes, “that Hooper (2002) does not provide one shred of evidence to support this serious allegation.”
In conclusion, to answer the question of whether the criticisms of the peppered moth story are justified, each of the classes of criticism should be considered separately. Certainly, the criticisms in the first two classes, pertaining to lack of knowledge of some aspects of the behavior and ecology of the moth and to the artificiality in some of the experiments, were justified. Indeed, many of these criticisms have been accepted by the scientists they were aimed at and in some cases the criticisms were self-criticisms. Moreover, most of these weaknesses in the case have been addressed and answered over the last half century.
The pseudoscientific criticisms are themselves so flawed as a result of lack of either objectivity or understanding, or both, that they are clearly unjustified and, as such, need no further response. Finally, the accusations of fraud have been found to be vacuous, unsubstantiated, and unsustainable. While scientifically they require no further consideration, as they denigrate the reputations of two dead scientists, it is to be hoped that those who invented them and others who have repeated them will retract and apologize for these accusations in print.