In a paper published in Evolution in 1987, Raymond Huey and Albert Bennett presented the results of their comparative analysis of temperature preferences and temperature dependence of running speeds in Australian lygosomid skinks. By chasing lizards down racing tracks at different temperatures, Hue and Bennett showed that coadaptation between thermal preference and thermal dependence of running can be tight, partial to even antagonistic in different parts of the lygosomid clade. Thirty years after the paper was published, I asked Raymond Huey and Albert Bennett about their motivation for doing this study, the origins of their collaboration, and what we have learnt about this topic since the time of this paper.
Citation: Huey, R. B., & Bennett, A. F. (1987). Phylogenetic studies of coadaptation: preferred temperatures versus optimal performance temperatures of lizards. Evolution, 41(5), 1098-1115.
Date of interview: Questions sent by email on 12th September 2016; responses received by email on 19th May 2017
Hari Sridhar: This paper came at a time when each of you had already spent close to 20 years working on lizards (including for your PhDs). What was the motivation to do the specific study described in this paper, in relation to all that you had done earlier on lizard physiology?
Raymond Huey and Albert Bennett: The route to the Australian skink paper was circuitous. We were both physiological ecologists; neither of us was trained in evolutionary biology or in phylogenetics. We shared an interest in comparative biology, temperature, performance, and energetics. Both of us had done simple interspecific comparisons in our Ph.D. research. Later, we both had measured the thermal dependence of sprint speed in lizards and were interested in co-evolution of thermal preferences and optimal temperatures.
The skink project was Al’s idea, and he obtained a NSF grant to fund the research. Allen Greer had recently shown striking variation in critical thermal maxima of Australian skinks (for example, desert species had higher CTmax than species in other regions). Al saw the potential for looking at the radiations of this large and diverse clade, especially as regards the co-evolution of thermal preference and of thermal sensitivity.
To us, studying thermal diversity within a large clade was critical. Most previous studies of physiological evolution involved very few species (often two), and typically the involved species were distantly related. The rationale was presumably the expectation that physiological differences between distant relatives are more likely to be conspicuous than those of close relatives. However, distant relatives may differ for many reasons, not just the “adaptive” traits being investigated. Australian skinks provided an opportunity to study radiation within a single and speciose clade.
The importance of incorporating phylogenetic considerations into comparative analyses was only just being to be recognized at that time. In our previous research on the physiological ecology of foraging mode in Kalahari lacertid lizards, we had emphasized the importance of studying close relatives; and we were also interested in using a phylogeny to evaluate whether wide-ranging or sit-and-waiting foraging mode was ancestral(Huey & Bennett 1986). However, a new methodology for us and for comparative biology in general was provided by the publication of Felsenstein’s paper (1985) on phylogenies and the comparative method.
Al spent most of a sabbatical year at the University of Adelaide in South Australia, working in Roger Seymour’s lab, and Ray joined in the study for a month or so. Ray and Al made the activity measurements in that laboratory.
HS: Stepping back a bit, how did you get interested in lizard ecology in the first place?
RH & AB: Ray originally planned to study birds, but became interested in lizards while on an Organization for Tropical Studies course in Costa Rica 1966. He realized that lizards were much easier to study (and catch) than birds, and so switching organisms was an obvious move. He became interested in thermal biology especially after spending most of a year in the Kalahari Desert, where it was obvious to him that temperature drove many aspects of the behavior andecology of lizards (Huey, Pianka & Hoffmann 1977). His original interest in the thermal dependence of sprint performance was driven by a desire to evaluate whether seasonal, habitat, or daily variation in body temperature influenced relative performance of lizards in nature. But his work with Anolis lizards in Puerto Rico soon led to comparative work on thermal dependence of thermal preference and thermal sensitivity of locomotion.
Al’s primary research interest was the effect of temperature on functional capacities in vertebrates in general. Reptiles were always of special interest because the reptilian suite of physiological capacities twice, independently, gave rise to the major endothermic groups, mammals and birds. The evolution of those functional systems and the evolutionary selection that favored them continues to be one of the most interesting questions in comparative biology. The functional capacities of reptiles were not well understood in regard to metabolism and activity, and lizards were by far the most accessible group for their study.
HS: The fieldwork and experiments for this study were done in Australia in 1983 and 1984. Could you give us a sense of what your daily routine was like, while you were doing this work?
RH & AB: The first step was collecting the lizards in the field. Al and Henry John-Alder collected some animals before Ray arrived, and others were collected by friends and laboratory co-workers, particularly Mike Thompson. The lab work consisted of two complementary projects. First, Al and Henry measured thermal preferences, critical thermal maxima, and critical thermal minima of several species (Bennett & John-Alder 1986). Next, Al and Ray raced lizards on Ray’s racetrack and measured the thermal dependence of speed (Huey & Bennett 1987). The daily routine involved racing many lizards many times, extracting the fastest speed of each run. It was not very exciting, but the Talking Heads and the Rolling Stones eased the tedium.
HS: How did the collaboration with Al Bennett come about? What did each of you bring to this study? Did both of you go to Australia for the field work and lab experiments?
RH & AB: In the early to mid-1970s, we began to follow each other’s published work, but we didn’t meet until ~ 1976. At that time, Al was an Assistant Professor at UC Irvine and came up to UC Berkeley, where Ray was a postdoc and where Al had previously been a postdoc. We already knew that we shared many interests in common. At that first meeting, we discovered that we were both (independently) working on ways to measure the thermal dependence of sprint speed in lizards. Al had already built a racetrack with photocells at the start and end, and Ray was in the process of building one using multiple photocells and a timer that used TTL logic (computer systems were not yet available).
Our first real work together involved the Kalahari research. We brought complementary abilities to that project and to our future collaborations. Al had much more grounding in physiology, and Ray had more in field ecology.
We enjoyed working together, and our collaboration was definitely synergistic. Ironically, some years later, Ray gave a lecture at Oxford. He talked about work he was doing with lizards, but he also discussed parallel work Al was doing with snakes. After the seminar, one distinguished Professor told Ray that he couldn’t understand why he and Al weren’t at each other’s throats, given the similarity of their research programs. Yes, we were competitors in some sense, but we felt being collaborators was much more fun and much more productive. If you peruse the publication histories of the two of us, you will find that almost all of our research was done and published in collaboration with a large number of people of very diverse interests and expertise. Both of us feel this is a much more effective way of doing science than being confined to a single laboratory group.
HS: Roger Seymour gave you access to his lab space and facilities? How did you know him?
RH & AB: Roger is a physiological ecologist whose graduate mentor was George Bartholomew, Al’s academic grandfather (via Bill Dawson). Thus, Roger was academic family, and he and Al shared a number of connections and interests prior to this sabbatical interaction. Roger was not involved in this research; but during that year, he and Al undertook a variety of other studies together, including field studies on crocodiles in the Northern Territory of Australia.
HS: How did you change the temperature for the racing experiments?
RH & AB: Seymour’s lab had small environmental chambers, so changing temperature of the lizards was easy, much easier than in the previously mentioned Kalahari study. There we had to improvise ersatz “environmental chambers” using large cardboard boxes and hair dryers coupled to a temperature controller.
HS: Table 4 had data gaps at the time of the study. Would you know if those gaps were filled later on?
RH & AB: We do not know whether field Tbs (mean field body temperatures) are now available for these species.
HS: Do you continue to work in these Australian sites till today? When was the last time you visited them? Would you know if these sites have changed a lot from the time you worked there for this paper?
RH & AB: No, we never went back to these sites, nor did additional work on these lizards, nor do we know whether anyone has gone back to these sites. We moved on.
HS: Would you remember how long the writing of this paper took? When and where did you do most of the writing?
RH & AB: We did the lab work in early 1984, and the paper appeared in late 1987. So the analysis and writing took several years. Both of us were working on other (unrelated) projects, but the main reason for the delay was the challenge figuring out how to think about and analyze comparative data in a phylogenetic context.We had no models to follow, and so we had to figure it out ourselves, with a lot of help from Joe Felsenstein (see below).
On writing: Ray was in Seattle, and Al was in Irvine. Skype and email hadn’t yet evolved, and so we communicated largely by mail and occasional visits.
HS: How did you know J. Felsenstein at that time? Did he provide his inputs on phylogenetic methods and analysis remotely or in-person?
RH & AB: Joe was a colleague of Ray’s at the University of Washington. They ate lunch frequently and often discussed phylogenetic issues and comparative methodologies. Joe was a critical influence on the concept and analysis phases
Here was the issue that interested us most: Al and Henry had found marked interspecific variation in the thermal preferences of these skinks, we wanted to know the ancestral state for thermal preference – was it high or low? Also, we wanted to know whether species with relatively high thermal preferences also ran fastest at high temperatures, implying co-adaptation of behavior and physiology.
The basic challenge was how to reconstruct values of thermal preference and optimal temperature, and how to evaluate evolutionary correlations between these two traits. This challenge drew us into thinking about phylogenies and ancestral states.
Methods were available at that time for analyzing categorical traits (e.g., presence or absence of some trait) but not for quantitative ones (e.g., thermal preference, optimal temperature). Ray asked Joe how to do this; and Joe suggested the minimum-evolution algorithm, which we adopted. Incidentally, we didn’t use Independent Contrasts, which Joe had recently published (1985), because we didn’t have branch lengths for our phylogeny. Within a few years, better phylogenies appeared; and Garland, Huey, and Bennett (1991) then used Independent Contrasts to re-analyze the skink data.
HS: Did this paper have any kind of direct impact on your career? In what way did it influence the future trajectory of your research?
RH & AB: That paper had a big impact on comparative/evolutionary physiology, as it was the first paper to show the power of looking at physiological evolution in a phylogenetic context. Even so, it had little effect on our own subsequent careers simply because both of us soon switched from comparative (historical) approaches to using laboratory natural selection or truncation selection to study – and manipulate — physiological evolution in real time. Here we were influenced by experimental work of Michael Rose, Rich Lenski, and Linda Partridge.
Interestingly, Martin Feder and Ted Garland, Jr. – very close colleagues of ours – made similar career shifts about the same time. These shifts were not independent evolutionary “innovations.” We are all talking together and thinking along the same lines, and the transitions of our thinking is evident in “New Directions in Ecological Physiology” (Feder et al. 1987) and two subsequent papers (Garland, Huey & Bennett 1991; Feder, Bennett & Huey 2000).
HS: Today, 29 years after it was published, would you say that the main conclusions still hold true, more-or-less?
RH & AB: The Garland et al. (1991) paper did alter some of our conclusions. But the conclusion that matters most was the utility of studying the evolution of thermal sensitivity (or any other trait) in an explicitly phylogenetic context.
HS: If you were to redo this study today, would you do anything differently, given the advances in technology, study design and statistical techniques?
RH & AB: Yes, of course. We would certainly add more species to gain statistical power, obtain a much better phylogeny, and use contemporary comparative methods.
HS: This paper has been cited over 400 times. Would you know what it is mostly cited for?
RH & AB: We’ve never done a citation analysis, but we assume this paper is widely cited because it was an early example of using phylogenetic analyses to evaluate physiological trait evolution in a clade. It was also the first to use a minimum-evolution algorithm for quantitative traits. And it is also interesting because it specifically quantified evolutionary co-adaptation between behavior and physiology.
HS: You say “Our phylogenetic analysis thus uncovers something missed by the equilibrium analysis – a possible case of behaviour and performance evolving in opposite directions. These patterns should, however, be further tested, using phylogenies based on molecular data”. Has the further testing with molecular phylogenies happened?
RH & AB: Yes, we reanalyzed the data that way in 1991 (Garland et al.). We suspect better phylogenetic information is available now, but we have not searched for it.
HS: You coin a new phrase in this paper – “antagonistic coadaptation”. Did you consider other phrases to use before deciding on this one? Is that still the term used to describe this phenomenon?
RH & AB: We undoubtedly considered other terms. Antagonistic coadaptation was likely inspired by the phrase “antagonistic pleiotropy,” which was important in evolutionary genetics and in the evolution of aging. “Antagonistic coadaptation” still appears in the literature: however, it is used primarily with regards to sexual conflict, not to physiological coadaptation.
HS: In the Discussion you say a key question for future research is “how might selection favor or perhaps ignore the evolution of reduced locomotor performance”. Could you reflect on what we have learnt about this after 1987?
RH & AB: To our knowledge, not much. Biologists seem to be fascinated with the evolution of high-levels of performance, not of reduced performance. [Ken Dial’s work with birds is exceptional in this regard.] But some lineages (or life stages) of many animal lineages have obviously evolved reduced locomotor performance, probably for a variety of reasons. Sloths are derived, not ancestral.
HS: You say “the genetic parameters for the thermal dependence of sprinting have not yet been examined in reptile, and divergence times are unknown for these genera. Ultimately, breeding experiments may be necessary to evaluate this hypothesis”. Do we know more about divergence times now? Have these breeding experiments been done?
RH & AB: In Garland, Huey, and Bennett (1991), we reported available divergence times. We suspect better information is available now, as phylogenetic research on Australian lizards very active. To our knowledge, the only breeding experiments to date are with Anolis lizards, but these data are unpublished. Genetic studies for Drosophila have been done.
HS: Have you ever read this paper after it was published? If yes, in what context?
RH & AB: We must have re-read it (at least part of it) when writing Garland et al. (1991), and Ray probably re-read it when writing a chapter in “New Directions in Ecological Physiology” (1987).
HS: Would you count this paper as a favorite, among all the papers you have written?
RH & AB: A paper can be a personal favorite for many reasons (for example, because the project was with great colleagues, with interesting animals, in an exciting place, or generated interesting data). But a paper can also be a favorite if one learned a lot about how to do biology in the process of writing the paper. Our own understanding of physiological evolution evolved dramatically during this project. We were exploring how to think about, analyze, and present our data and their derived patterns in an explicitly phylogenetic context. This was not easy, but it was ultimately very satisfying – and the associated excitement makes it a favorite project and paper.
In retrospect, we realize that this project was probably the end of our major focus on comparative approaches. As we thought hard about how to do historical analyses, and in so doing we made ourselves very aware of the limitations of trying to reveal past history. We realized that comparative phylogenetic studies are inevitably descriptive, and thus non-experimental. So ironically, this project probably helped motivate us to stop doing comparative analyses of past evolution and to start exploring experimental studies of evolution in real time. Al soon started doing experimental evolution of thermal sensitivity of E. coli, and Ray started with that of Drosophila. Many of the questions were similar to those in the Australian skink story, but the methods and philosophies were entirely different.
Ray later did do some additional comparative work on latitudinal clines in Drosophila subobscura, an Old World species that was introduced into South and North America. This was a comparative project, but one in which the clines had evolved over just a few decades. And he later – as part of a project on vulnerability of tropical lizards to climate warming (Huey et al. 2009) — did include phylogenetic analyses of thermal sensitivity in lizards.
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 29 years ago? Would you add any caveats?
RH & AB: Keep in mind that this was one of the first phylogenetic studies – and certainly the first in comparative physiology – that made an explicit attempt to look at the evolutionary patterns of multiple quantitative traits for a clade of organisms. In that sense, it was pioneering, but the methods we used certainly don’t approach contemporary standards, and the phylogeny available to us at the time was very crude.
Even so, the fundamental co-evolutionary questions we asked are still valid and are still interesting to those interested in evolutionary physiology. And of course, concerns over climate change have brought the evolution of thermal sensitivity of performance into prominence in biology.