In 1991, Geoffrey Hill published a paper in Nature describing the results of his field experiments with house finches which showed that: 1. females choose to mate with brightly-coloured males; 2. males with brightly-coloured plumages tended to contribute more to parental care and were also likely to be of better “genotypic quality”. Twenty-five years after the paper was published, I asked Geoffrey Hill about why he chose this topic for his PhD, the challenges of carrying out these experiments, and what we have learnt since about the role of colour in mate choice in house finches.
Citation: Hill, G. E. (1991). Plumage coloration is a sexually selected indicator of male quality. Nature, 350(6316), 337.
Date of interview: Questions sent by email on 5th November 2016; responses received by email on 18th November 2016.
Hari Sridhar: Correct me if I’m wrong, but this paper seems to have formed part of your PhD thesis. Could you tell us what was the motivation behind this specific piece of work, in relation to the rest of your PhD?
Geoffrey Hill: My 1991 Nature paper was the centerpiece of my Ph.D. dissertation at the University of Michigan. My goal in conducting my dissertation research was to explain why birds have specific ornamental colors and patterns. Darwin’s hypothesis that choosing females had mating preferences for ornamental coloration had not been tested in birds. Work on mate choice for coloration in guppies had just been published by John Endler and Astrid Kodric-Brown, and I was anxious to try to conduct “definitive” experimental tests of choice for color in birds. The Nature paper presented the results of a field experiment, and I published the results of lab experiments with caged birds in a companion paper in Animal Behaviour.
HS: Stepping back a bit, could you tell us what got you interested in the topic of bird coloration in the first place? Also, how did you pick house finches as your study species?
GH: I completed a master’s degree at the University of New Mexico before entering a doctoral program at Michigan in 1986. My master’s research focused on explaining delayed plumage maturation in songbirds. I picked that topic because it was a topic of interest among behavioral ecologists in the early 1980s. In bird species with delayed plumage maturation, males do not attain full-blown breeding plumage in their first year even though they are sexually mature; they have drabber plumage that often makes them look like females. Several hypotheses had been proposed to explain delayed plumage maturation, and I tried to test these competing hypotheses in my master’s research. I soon realized, however, that it was rather pointless to try to explain age-specific plumage coloration when we had no real understanding of coloration in any context. I came to the University of Michigan determined to “figure out” coloration in birds and by extension ornamentation in animals generally.
I decided to focus on carotenoid coloration, which was almost unstudied in birds at the time, because of an obscure statement in an ornithology text by Joel Welty: “Canaries will, in successive molts, gradually change from yellow to intense orange if fed red peppers.” Condition dependent sexual signaling was a new and exciting topic in behavioral ecology when I read that statement about canary coloration in the early 1980s. The prospect of diet-dependent coloration sounded intriguing to me. Then, when I went to the literature to find out more about carotenoid coloration in birds, I found a paper published in 1976 in The Auk on pigmentation in House Finches. In that paper, Alan Brush and Dennis Power used thin layer chromatography to identify the carotenoid pigments in House Finches, one of the first characterizations of carotenoid pigments in feathers in any bird. House Finches were an invasive species in Michigan, having just colonized the area a few years before I arrived in Ann Arbor, and they seemed to be a perfect study bird. So, in 1987, I started capturing wild House Finches and visually quantifying their feather coloration.
HS: If you don’t mind my asking, how come your PhD supervisor(s) weren’t authors on this paper?
GH: In the 1980s, the evolutionary biology program at Michigan was not what we might call a nurturing environment for grad students. Students were expected to devise their own research projects, get their own funding, and prove they had potential as research biologists. The attrition rate was high. I designed, funded, and executed my thesis research without help from my advisor, so there was never any consideration of putting his name on the paper. My dissertation had six chapters, and I published all six as single-authored papers. I’m proud to say that all six of my single-authored dissertation chapters have each been cited more than 100 times. It was a swim-or-drown environment at Michigan in the 1980s, but once I started swimming, it did make the transition to independent faculty member relatively easy.
HS: Could you give us a sense of what your daily routine was like during the fieldwork for this study – How did you commute to your field site? Who were the people who helped you with fieldwork? etc.
GH: In choosing the House Finch as my study organism and the University of Michigan campus as my field site, I consciously traded the joys of working in a wild setting for the practicality of a tractable system in which I could conduct experiments. My plan was to study House Finches both in the wild and in captivity. The “wild” for House Finches was the main university campus – finches nested in shrubs and in ivy on the sides of buildings and they fed on weed seeds in the lawns. I started my finch project in the summer of 1987, capturing wild birds in backyards around the area to create a captive flock in the rooftop aviary at Michigan. I didn’t take birds from the campus population because I wanted those for my field study.
Once the breeding season started in March, my days were frantic. I would typically ride a bus to campus before first light, set up a mate choice trial that ran for 3 hours, then grab binoculars and head for fieldwork (out the front doors of the museum). About two days per week, I would trap and color band the local population of finches. In the second year of the field study, I used hair dyes to change the coloration of the males that I caught. That color-change experiment was the basis for my 1991 Nature paper. On other days, I would walk around campus with my binoculars recording the behavior of banded birds with a special focus on pairing. After about two and a half hours of field work, I’d run back into the museum, literally run up 5 flights of stairs to the roof, set up birds for the second mate choice trial of the morning, and then go back out for more field work. That field season kept me in great physical shape. When birds started nesting, I was often grabbing a ladder (a really big and heavy ladder) to go out and climb to nests to check contents or band chicks. This routine didn’t stop until the end of July when the finches had finished nesting.
HS: You say you used hair dyes, hair lighteners and colour developer for your experiments. Could you tell us a little more about this – Who came up with the idea of using these? Did you try any other options before deciding on this? Do you remember what particular brands of these chemicals you used? etc. Do you continue to use the same in experiments even today?
GH: Devising a means to change the coloration of birds was one of the most challenging parts of my research. Making birds redder was not so hard. There were red art markers that would increase plumage redness. But I also wanted to decrease redness, which was not easy and for which there were no published methods. I tried calling experts in bird coloration around the country like Sievert Rohwer at the University of Washington and Ken Parkes at the University of Pittsburgh. All of the senior zoologists who I contacted encouraged me to do the experiments, but they had no useful advice to offer regarding how to lighten the red coloration of finches.
I was starting to think that I would have to abandon the experiment or do it only by adding red to feathers when it suddenly occurred to me who I needed to consult: the real experts in manipulating the color of integument, beauticians. I looked in the phone book, found the address for a beauty supply shop, and drove over. It was both a salon and store, and those women really were experts on manipulating coloration. And when I explained my research project, they were extremely sharp and insightful consultants. In retrospect, I should have taken down their names and acknowledged them in the paper. They set me up with a strong hair lightener that was designed to remove melanin pigmentation from hair and that I hoped would remove carotenoids from feathers. Fortuitously, non-traditional hair colors were popular by the 1980s in Ann Arbor, so they had permanent, scarlet red hair dyes.
It all worked amazingly well, but it was a rather harsh experimental treatment for the birds. First, because birds coat their feather with oil from their uropygial gland, I knew I would have to strip oil off the feathers for dye to adhere to the feather substrate. So, I had to lightly shampoo the feathers of the birds before the color treatment. Getting birds wet to the skin is dangerous because they lose all insulation and can go hypothermic very fast, so I used warm water and worked fast. Immediately after the shampoo treatment, the birds either had red dye or hair lightener worked into their feathers. Both of these products had to be left on the feathers for 15 minutes to set, so I would wrap the bodies of the birds from the neck down in plastic wrap and then set the birds on a heat pad to keep them warm. The heat also helped the dyes and lighteners work. Finally, I had to rinse out the dye or lightener and then let the birds dry, fluff, and preen. Again, wet birds are at serious risk of hypothermia even in a relatively warm room, so I put the birds in cages in front of heat lamps. Heat lamps were dangerous for birds because a small bird can go from warm to over-heated very fast. To avoid that, I shielded half of the cage from the lamp so birds could move out of the heat if they got uncomfortable. The whole process worked really well and I didn’t lose a single bird during this dying process. I held the birds overnight to recover and within a few hours of the last rinse, they were fluffed and preened and looked like normal house finches again.
Since that experiment, I have run several experiments with my students in which we increased the coloration of red, yellow, and blue species with art markers, but I have never gone back to the dye treatment.
HS: Do you continue to work in this site? When was the last time you visited this place? Has it changed in any way from the time you worked there? Are house finches still common there?
GH: I did a last field season on the site in the summer of 1991 after I defended my dissertation in April, but when I drove away in June 1991 I was done forever with the House Finch population on the campus of the University of Michigan. Five years after that 1991 field season, a new passerine disease, Mycoplasma gallisepticum, swept through the eastern population of finches killing about half of the population. The eastern population has not rebounded from the effects of this disease, and finches in the east remain about half as abundant as they were in the early 1990s. I assume that the Ann Arbor population is about half as large as it was when I studied it.
HS: How long did the writing of this paper take? When and where did you do most of the writing? You acknowledge a number of people who commented on the manuscript – could you tell us who these people were and how you knew them at that time?
GH: I wrote this paper quickly, perhaps in two weeks for the first submitted draft. I’ve always been efficient at getting ideas down on paper. Also, as solo author, I wasn’t delayed waiting for anyone else. I worked at my own pace.
Then and now, I do almost all of my writing at my desk at the university. I find home too distracting. At the University of Michigan, I was in an office suite with two desks in my room and two desks in the attached room. Phil Chu was my roommate, while in the room next door was Rick Prum and Rebecca Irwin. Anyone interested in bird coloration will know the name Prum (Richard O Prum) who was not working on coloration in the 1980s when we were office mates but who started working on coloration in 1990s and did many important and groundbreaking studies. It was a great group of young scientists and down the hall was Michael Nachman (now director of the MVZ [Museum of Vertebrate Zoology]), Steve Dobson (now on faculty with me at Auburn), Kyle Summers, Steve Frank, Lyle Gibbs and many other students who are now prominent professors. It was my graduate student and postdoc colleagues who were my advisors and sounding boards, much more so than the professors. I cannot remember every discussing my dissertation ideas with any faculty member, but I discussed my ideas incessantly with my grad student and postdoc colleagues. As I go back and look at the acknowledgements in the paper, I listed my doctoral committee and no one else. In retrospect, I should have acknowledged my grad student and postdoctoral colleagues who helped me think through the design of that experiment.
HS: How did you draw the figures for this paper?
GH: There is only one figure, and it is really pretty terrible. I used to generate all of my graphics using a statistics program called Statview on an early generation Macintosh computer. I would create graphical summaries of data in PICT format that I could then import into a draw program. I really miss those old stats and drawing programs.
HS: Did this paper have a relatively smooth ride through peer-review? Was Nature the first place this was submitted to?
GH: I was a very young and naïve grad student when I sent that paper off to Nature. I just wrote it up and put it in the mail. (Yes, you still actually photocopied and mailed manuscripts in the early 1990s.) I don’t know what the rate of rejection without review was in 1990 (it is about 60% now), but my manuscript received a full review. In the decision letter, the editor, Rory Howlett, used the word “reject”, and I thought my paper was rejected. About a week after I got my decision letter from Nature and when I was getting ready to start to rework it for a new journal, I mentioned to my grad student friend and colleague, Michael Nachman (now director of MVZ at Berkeley), that I had had a paper rejected by Nature. He had never seen a decision letter from Nature and asked if he could see it. I’ll never forget his face as he read the letter. He looked up from the letter and said “Geoff, this is not a rejection letter. This is a tentative acceptance letter. They want you to revise the paper and resubmit!”. I was completely clueless. I had seen the word “reject” and assumed Nature didn’t want it. Thank goodness Michael was there to coach me. So I revised the paper, wrote a response to the reviewer’s critiques, and the next letter I got was “you will receive proofs in a month”.
HS: What kind of attention did this paper receive when it was published?
GH: The Nature paper got no special attention at all, so far as I could tell. I submitted House Finch photos for the cover, but Nature used a map of Antarctica on the cover instead. My house finch study seemed buried amongst the cell biology and physics. There was zero press about the paper. No one even mentioned it to me the week it came out.
HS: What kind of impact did this paper have on your career and the future course of your research?
GH: It is hard to assess the impact on my career. I was struggling to get a postdoctoral position toward the end of my graduate program at Michigan, and according to my eventual postdoc advisor, Bob Montgomerie, the Nature paper made a big difference in earning an NSERC postdoc in Canada. So, it likely got me a postdoctoral fellowship when I really needed one, and it was likely a factor in getting me an interview and eventually a faculty position at Auburn University.
HS: Today, 25 years after it was published, would you say that the main findings still hold true, more-or-less?
GH: I would say that the central finding holds up better than more-or-less. The take-home message from the Nature paper is that female House Finches show a mating preference for red carotenoid coloration, which is a condition-dependent signal of male quality. That core finding has been corroborated with other studies of House Finches both by my lab group and by other lab groups, and the observation that red coloration is an important criterion in female mate choice has been demonstrated in other species of birds as well.
HS: If you were to redo this study today, would you do anything differently, given the advances in technology, theory, statistical and lab techniques etc.?
GH: Looking back 25 years later, I think that I got substantially off track in my early studies of condition-dependent signaling by focusing so much on carotenoid as a limiting resource and by pushing resource limitation as the basis for honest signaling. Ironically, the resource limitation idea caught on so well that it is now dogma in the field of sexual signaling even as I have concluded that it is, for the most part, incorrect. I now have my own studies from 20 years ago cited against my new ideas. But, none of this rethinking concerns female mate choice for carotenoid coloration, which is the focus of the Nature paper. I would repeat the experiment in the same way, with the same basic experimental design.
HS: This paper has been cited close to 700 times. At the time of the study, did you anticipate at all that it would have such a big impact? Would you know what it is mostly cited for?
GH: I conducted this study because I knew there was a critical need for a test of female mate preference for ornamental coloration. Moreover, I knew that tying such a mating preference to condition-dependent signaling would draw people’s attention. I can’t say that I am really surprised by the modest success of the paper.
HS: You say “In all comparisons, returning males were significantly more colourful than males that did not return , suggesting that colourful plumage is an indicator of male viability and perhaps genetic quality , although the relationship between colour and viability might be due to phenotypic effects.” Today, do we know more about the relationship between colour and viability?
GS: We are just starting to understand how coloration connects to organismal performance, including viability. Earlier this year (2016), I was part of a team that discovered the gene that enables House Finches and other songbirds to convert yellow dietary pigments to red pigments. This is the key gene whose activity determines whether an individual finch will be more or less red. Now that we know the identity of this key redness gene, we can investigate how its function links individual carotenoid coloration to individual performance and ultimately viability. We have a grant pending to fund such a study and my grad students are already collecting data to test hypotheses for how this might work.
HS: You found a significant correlation between plumage of fathers and sons and say “Although this correlation fits predictions of the ‘good genes’ models, it could also be due to maternal or paternal effects as colourful males tend to provide more food and probably pair with higher quality females than drab males.” Do we know more about the reason for this correlation today?
GH: We really do not know any more about heritability of red coloration in male House Finches now than we did in 1991. We are further along in building fundamental understanding of the genetic basis for coloration now, and so I think we are closer to an answer. A major focus of my lab going forward is to use functional genomics techniques to understand how gene regulation and inherited genetic variation might contribute to color variation among males.
HS: Have you ever read this paper after it was published? If yes, in what context?
GH: Yes, in the early 1990s I ran a graduate class “Advanced Ornithology” in which we focused on feather coloration and ideas related to honest signaling. The paper was assigned reading and I re-read it in preparation for class. I’m not like an actress or actor who states that they never watch their own movies. I re-read my own papers all the time, often because I have to remember what I wrote.
HS: Would you count this paper as a favorite, among all the papers you have written?
GH: A few years ago, on my lab web page, I started a list of the “25 most impactful Hill lab papers”. I currently [18th November 2016] have the 1991 Nature paper listed as the 8th most impactful paper out of about 200 that I have published. Among the seven papers that I rated above it, the oldest is 2011. So, in my mind, it was 20 years before I wrote a more impactful paper than the central paper in my doctoral dissertation. I’ve made a conscious effort over recent years to write fewer but more meaningful papers like my 1991 Nature paper, so I would definitely say that it is one of the papers of which I am most proud.
HS: What would you say to a student who is about to read this paper today? What should he or she take away from this paper written 25 years ago? Would you add any caveats?
GH: I would advise a student who was about to read the paper to consider that this Nature paper was the summary of a simple experiment conceived of and executed by a young graduate student. Inexperience is an obvious hindrance to a young person creating cutting-edge science, but a fresh perspective, freedom from accumulated biases and skepticisms, and naivety that enables “impossible” tasks to be attempted can be a huge counter to the inexperience. I would encourage the student that there are innumerable simple, high-impact studies waiting to be conducted.