« There are none so blind... | Main | Evolution 2008: sexy plants, battling bacteria, durable cooperation »

Guest blog: The Peacock's Tale

This week's post is by Dave Wisker, a graduate student in Molecular Ecology at the University of Central Missouri.

It's the creationist's dream. If actual evidence of creation is too much to hope for, how about a peer-reviewed paper in a respected journal overturning one of the icons supporting a major element of Darwin's theory?. Sexual selection in peafowl is definitely one of those icons. There appeared to be ample empirical evidence that peahen's preference for more elaborate trains on their mates has led to the spectacular male tail displays we see today. A series of papers in the 1990's by behavioral ecologist Marion Petrie and others seemed to solidly support this, and there is also evidence that elaborate tails may indicate good genes (Petrie et al, 1991; Petrie and Williams, 1993;.Loyau et al, 2005a). This, in itself, is a challenge to an older idea that the peacock's tail shows how arbitrary female preferences can be amplified to extremes by a "runaway"? process (Fisher, 1958). But, whatever their evolutionary origin, the preference itself has rarely been questioned.

However, a recent paper published in Animal Behaviour (http://tinyurl.com/4t69v5), "Peahens do not prefer males with more elaborate trains"?, challenges the conventional wisdom.

Mariko Takahashi and her colleagues studied a feral population of peafowl in Japan for seven years, and report finding "no evidence that peahens expressed any preference for peacocks with more elaborate trains (i.e., trains having more ocelli, a more symmetrical arrangement or a greater length), similar to other studies of galliforms showing that females disregard male plumage."? Intelligent Design and creationist groups have already praised this paper, saying it has demonstrated that the conventional evolutionary explanation for the peacock's tail is little more than a "just so"? story. A look at this paper and those from some representative studies that came to a different conclusion seems in order. Note: all of the studies looked at additional aspects of tail morphology besides eyespot number/symmetry, but this essay will focus only on this particular feature, for space reasons and because it is the heart of the 'controversy'. In addition, all of the biologists involved agree that tail ornamentation is only one of several display cues that a female uses in selecting a mate.

Takahashi's group noted that the number of eyespots, or ocelli, and their symmetry varies during an individual male's lifetime. So they photographed every male's train each season over the seven years, and devised a "Fluctuating Assymetry "index (FA) for each bird. The FA is defined as "a percentage of the number of eyespot pairs having lost one of the pair from a symmetrical linear position divided by the total number of eyespot pairs on the train."? The lower the FA value, the higher the symmetry, with a value of zero indicating perfect symmetry. They also used the number of eyespots as a different metric. Studies by Petrie & Halliday (1994) and Loyau et al (2005b), which did find that females prefer males with more elaborate trains, used eyespot number only. Petrie and Halliday removed 20 of the outermost eyespots from their treatment males using scissors (they did this without disturbing the length of the train), and used eyespot number as their metric. Loyau et al did not experimentally alter the eyespot number, but used the number of ocelli as their metric as well. So it appears all three studies used the same elaboration metric in at least part of their analysis.

Mating success was another metric. Petrie and Halliday used number of copulations for each male (after displaying and being accepted by a female), and the change in number of copulations for each male, if any, after eyespot removal. Loyau et al also recorded number of copulations after being accepted by a female. Takahashi et al took a different approach. They estimated male mating status "using female courtship behaviours. This was done to decrease the likelihood of excluding possible mates from the analyses because of limited observations."? Females show three categories of behavior when in a male territory: passing by the male, passively receiving his display, or actively soliciting a display by "run-around"? behavior. Takahashi's group noted that females who perform multiple "run-around"? behavior for a male often accept that male for copulation later, so they considered any female who performed more than two successive run-arounds for a male as having made a "preferred visit"?, and counted that as a successful copulation. This could be a crucial source of error, if a significant number of females performed multiple run-arounds but did not accept the male for copulation later, or if they copulated after less than three run-arounds. It's certainly worth investigating separately.

Takahashi et al point out that some of the studies which found a positive correlation between male mating success and train elaboration may be flawed because they only observed the birds in the morning, missing the crucial evening period of sexual activity. The creationists have especially harped on this. While it is true that Petrie and Halliday only observed in the morning, Loyau et al's paper specifically states they made observations from 0500 to 1700 hours. So that factor may not be significant in explaining the discordant results.

The Japanese authors also discuss other factors that may explain the different conclusions, namely sample size, length of study, and the age of the males involved. The advantage their study has over the others is the number of individuals (around twice that of the others), length of the study and the range of ages of the males. It is not clear, however, if those factors contributed significantly to the results.

Finally, this paper makes a very interesting observation: all of the studies of peafowl have shown very little natural variation in male train elaboration, regardless of the metric used. This suggests--to me at least--that one has to very careful in evaluating reproductive success, because the positive correlations were not particularly strong. If the males lack enough variation for females to make a choice, then it will be exceedingly difficult to discern female preference for elaborate trains in the wild. The fact that Petrie and others have added experimental variation and detected apparent selection gives one reason to think that, at this point, the evidence still suggests that females do prefer males with more elaborate trains. Suppose your hypothesis was that women prefer men with cars. If every man in your study group had a car, it would be hard to tell, from simple observation, whether female preferences were the cause of car ownership or not. You would need to take away some cars and see whether that has any effect. So the creationist's dream remains just that. However, these studies do inspire more research questions. How strong is the selection, if it exists? Is the low natural variation a direct result of female preference, that is, has the population reached its physiological limit for train elaboration? Does predation control the further runaway growth in train size/elaboration? What effect will this low natural variation have on the future of female preferential behavior? For example, Takahashi et al. found that a behavior known as "shivering"? be important. It looks like a fertile field for further investigation.


Fisher RA (1958). The Genetical Theory of Natural Selection, 2nd Revised ed. Dover. 291 pp.

Loyau A, M Saint Jalme, C Cagniant and G Sorci (2005a). Multiple sexual advertisements honestly reflect health status in peacocks (Pavo cristatus). Behav. Ecol. Sociobiol. 58: 552-557

Loyau A, M Saint Jalme, C Cagniant and G Sorci (2005b). Intra-and intersexual selectiojn for multiple traits in the peacock (Pavo cristatus). Ethology 111: 810-820

Petrie M, TR Halliday and C Saunders (1991). Peahens prefer peacocks with more elaborate trains. Anim. Behav. 44: 585-586

Petrie M and A Williams (1993). Peahens lay more eggs for peacocks with larger trains. Proc. R. Soc. Lond. B 251: 127-131.

Petrie M and TR Halliday (1994). Experimental and natural changes in the peacock's (Pavo cristatus) train can affect mating success. Behav. Ecol. Sociobiol. 35: 213-217

Takahashi M, H Aritat, M Hiraira-Hasegawa, and T Hasegawa (2008). Peahens do not prefer peacocks with more elaborate trains. Anim. Behav. 75: 1209-1219


Excellent write up, very interesting read!

The entire evolutionary position is a just story story and the fossil record will continue to dog evolutionists because it does not support the evolutionary position. And of course, when it comes to the fossil record unfortunately for the evolutionists the burden of proof is upon the claimant.

I cite the following:

Even evolutionist Mark Ridley, who currently serves as a professor of zoology at Oxford, stated the following: "In any case, no real evolutionist, whether gradualist or punctuationist, uses the fossil record as evidence in favour of the theory of evolution as opposed to special creation."

I reserve the right to delete comments that have no connection to the particular post, which, in this case, doesn't mention fossils. But what the heck. The full quote is:
"In any case, no real evolutionist, whether gradualist or
punctuationist, uses the fossil record as evidence in favor of
evolution as opposed to special creation. The does not mean that
the theory of evolution is unproven.
`So what is the evidence that species have evolved? There have
traditionally been three kinds of evidence, and it is these, not the
"fossil evidence", that the critics should be thinking about. The
three arguments are from the observed evolution of species, from
biogeography, and from the hierarchical structure of taxonomy.'

The evolution theory is hurt this way: The new study’s empirically-derived data diminishes (not destroys) the idea that peahens prefer peacocks with more elaborate trains. Therefore, sexual selection is a less satisfactory explanation for the peacock’s extravagant plumage.

[Ford here. Thanks for admitting that, at most, this paper raises doubts about peahen preferences. Yet everything you wrote below assumes sexual selection has been completely disproved, at least for peafowl. Given the criticality of that assumption to the rest of your comment, how about discussing the key difference between this study and others that have shown peahen preferences for more elaborate tails? Since you apparently missed it the first time, I've bolded it.]

Is survival selection a sufficiently powerful mechanism to explain the plumage? The peacock's overly-extravagant tail would hinder its survivability. (reduced maneuverability, unable to fly, more conspicuous to predators, harder to maintain, energy for growth is expended). [Agreed.]

The theory of evolution is plagued with another problem. It has to explain how one of the most elaborate and priceless art masterpieces in the world was painted by a cumbersome “hit and miss” process of natural selection. This would be likened to Da Vinci painting a Mona Lisa by making one brush stroke every 3 years (peacock sexual maturity). After making only one stroke, the unfinished painting is presented to art critics. After heeding the consensus of the critics, Da Vinci adds another stroke.

[Since each peacock is descended from parents that were themselves functional birds, single brush strokes aren't a good analogy for the evolution of tail ornamentation. A better one would be revision of a mediocre rough draft of a poem by making thousands of copies with different random changes (including duplications and deletions of parts), letting critics vote on their favorites, and repeating millions of times. -- Ford]

This process is repeated every 3 years. [i.e., millions of times over the course of peacock evolution]

If a paint bottle is accidentally spilled on the canvas (lethal mutation), the painting is destroyed.

[Fortunately, each generation has thousands of copies without lethal mutations, and one or two of them may be seen as improvements by the critics.]

Does anyone really believe that evolution not only painted the feathers, but designed their mechanical structure and the bird’s entire system?

[Just biologists and those members of the public who respect our expertise. Do you have an alternative hypothesis? If so, is it consistent with the molecular data, such as: http://www.gbwf.org/pheasants/polyplectron.pdf ? ]

Before answering, click here and do some serious looking. No artist can duplicate the necessary iridescent paint let alone all the ideas for the patterns, textures, shapes, and dimensions. Human artists only "copy".

Nice to see someone who really knows their subject. I wrote a similar article for a squidoo lens that was closely related to this topic.

Very nice post. Thanks for this.

What troubles me in most discussions of sexual selection theory is that there is an emphasis on demonstrating the impact of a female (or, in some cases male) preference on the evolutions of the species.

It seems to me there is insufficient attention to the question of how the preference itself evolved in the first place. Conceptually, before a peahen can positively select for a peacock with extensive plumage, the peahen's own preference had to first be positively selected by natural selection. This is not only a chicken-and-hen problem, though it is that, but it raises the fundamental question as to how the pre-existing preference could possibly have been naturally selected in the first place. I don't understand how the female preference could evolve, and be positively selected as beneficial to the female's survival, before the males even came along with the attribute sought. To fudge this issue by saying that the process must have occurred simultaneously will not work.

In order for the existence of the preference to drive the evolution of the plumage, or the existence of the plumage to drive the preference, one must pre-exist the other, raising the question of how the pre-existing characteristic could possibly have been selected. (It is inconceivable that by luck both the preference and the feature would pop up in a pair of mutations, and these two mutants would find themselves and mate offspring with both characteristics.) And I don't understand how, if for example it was the preference that first popped up in a mutation or series of mutations, this variation could possibly be positively selected and spread across the peahen population when there was no positive survival influence on the population--not just one we can't identify, but not one conceivable, as far as I can tell.

Thanks for your blog, and for the opportunity to comment.

[Thanks for your comment. Although it may indeed have been neglected, there has been some attention to this problem. See Alexandra Basolo's "Female preference predates the evolution of the sword in swordtail fish" for an early example. More recently, I've seen seminars showing male fish color or frog call frequency are correlated with female ability to see that color or hear that frequency, which in turn are related to light environment (dependent on water depth) or background noise. I'll keep an eye out for papers on this. -- Ford]

A religious friend had handed me an article from some faith based magazine citing this study and declaring that Darwin has been proven wrong yet again.

I somehow doubt that he'll read your post when I send the link to him- the faithful thrive on simple answers to sooth their simple minds.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Type the characters you see in the picture above.