Aggressive Behavior Is Rarely the Result of Circulating Testosterone Levels

Anolis cybotes, one of the species included in Husak and Lovern, still showing its dewlap during copulation.

Anolis cybotes, one of the species included in Husak and Lovern’s study, still showing its dewlap during copulation.

If I were to take survey of Anole Annals readers regarding the factors that regulate aggressive and showy behaviors, I suspect that the vast majority of you would implicate testosterone as the primary culprit. Whether we are discussing humans or nearly any other vertebrate, there is a common societal notion that testosterone fuels these behaviors like oxygen fuels fire. The widespread belief is simple: individuals with more testosterone tend to exhibit more aggressive, ostentatious, and risky behaviors.

For decades researchers have investigated the link between testosterone and behavior in a variety of biological contexts – including different behaviors, experimental manipulations, environmental conditions, and life history parameters – but rarely in wild animals or within an evolutionary context. If the supposed testosterone-behavior correlation is extended to a broader, comparative context, it would suggest that aggressive species should also have higher levels of circulating testosterone than more placid species. But, in an upcoming paper, Husak and Lovern test the testosterone-behavior supposition among Anolis lizards and, quite frankly, turn it right on its head. To give away their conclusion at the outset, three of the four “aggressive” anole lineages examined have evolved this behavior without a clear correlation with circulating levels of testosterone.

Anolis lizards are renowned for their convergent anatomical evolution (reviewed in Lizards in an Evolutionary Tree), but these species have also independently evolved similar behaviors. In a study that was one of the first of its kind, Johnson et al. showed that the Anolis ecomorphs exhibit evolutionary convergence towards similar patterns of aggressive display and territorial behaviors. Trunk-ground anoles tended to be the most “aggressive” ecomorphs, consistently exhibiting higher display rates and  territoriality than the trunk-crown, grass-bush, or twig ecomorphs. Twig species tended to exhibit the least aggressive behavior in the analysis. (Also see Ord et al. 2013 for a more fine-scale dissection of display behavior.) Using this pattern of convergent behavior as a foundation, Husak and Lovern predicted that trunk-ground anoles would have higher levels of circulating testosterone than other ecomorphs from the same island, twig anoles the least. The absolute levels of testosterone might vary depending on the specific lineage in question, but they predicted that the rank-order of testosterone on each island would follow the behavioral continuum described in Johnson et al. In total the authors surveyed circulating levels of testosterone and corticosterone, an adrenal steroid hormone associated with stress, in 18 Anolis species!

Figure 1 from Husak and Lovern 2014: Circulating testosterone levels in 18 species of Caribbean Anolis lizards. Bars group by ecomorph classification (CG= crown giant, GB= grass-bush, T= trunk, TC = trunk-crown, TG = trunk-ground, TW= twig) and color coded by island (white = Bahamas, light gray = Jamaica, dark gray = Dominican Republic, black = Puerto Rico).

Figure 1 from Husak and Lovern 2014: Circulating testosterone levels in 18 species of Caribbean Anolis lizards. Bars group by ecomorph classification (CG= crown giant, GB= grass-bush, T= trunk,
TC = trunk-crown, TG = trunk-ground, TW= twig) and color coded by island (white = Bahamas, light gray = Jamaica, dark gray = Dominican Republic, black = Puerto Rico).

As I already stated, the authors found no support for the idea that elevated levels of circulating testosterone consistently drive aggressive behavior in Anolis lizards. Instead they found that three out of the four clades of trunk-ground anoles had the lowest levels of testosterone, the opposite pattern than would be predicted based on their behavior. Only on Puerto Rico did the trunk-ground species exhibit the highest levels of testosterone. In fact, testosterone levels are evolutionarily quite labile among anoles and also do not correlate with body size (snout-to-vent length), body mass, or gross levels of sexual dimorphism (1, 2). However, testosterone levels are not evolutionarily sloppy, showing rapid changes among closely-related species with similar biology; no differences in testosterone were found in either case where sister species were examined, the two Puerto Rican trunk-ground species and the trunk anoles A. brevirostris and A. distichus. Therefore, circulating levels of testosterone among anoles likely correlate with another behavior, evolve as a byproduct of selection for something unrelated to behavior, or might be functionally neutral  whereby changes in behavior are mediated by other downstream factors (i.e., brain-specific expression of hormonal receptors) or other pathways altogether (i.e., arginine vastocin).

Despite our widespread societal ideas regarding the role of testosterone on behavior, data collected within the past few years is starting to push back against this idea. It is important to point out that many of the studies that previously examined the effects of testosterone compared castrated males with intact males or males with testosterone levels elevated beyond physiological norms. Studies like these may be ideal for understanding the physiological effects of testosterone, but are not well suited for understanding the proximate mechanisms of behavioral variation in a diverse natural population where multiple pathways can yield similar results (referred to as “many-to-one mapping” in Husak and Lovern). Johnson et al. (different paper than that referenced above) also found a negative correlation between plasma testosterone and aggressive displays within a wild population of A. carolinensis. Although additional data is clearly necessary, this observation provides an intriguing connection between intraspecific variation and the interspecific patterns of evolutionary divergence described by Husak and Lovern. Furthermore, the mental relationship between testosterone and aggressive behavior is also weakening in other vertebrate systems. For example, in humans a placebo elicited a greater aggressive response if the participants were told that the placebo was testosterone than in individuals actually administered testosterone. Likewise, following a review of testosterone data from 51 wild bird species, it has also become clear that variation in circulating testosterone levels cannot alone explain differences in mating system or secondary sexual trait expression, traits often strongly linked to testosterone in laboratory-based studies.

It is clear that testosterone is a potent physiological hormone that plays a critical role in sexual differentiation during development and in adulthood. But as we move forward in attempts to tease apart the proximate mechanisms of evolutionary divergence, it is clear that we will need to take a more careful approach to studying the effects of testosterone, embracing both experimental and comparative approaches and surveying circulating levels of hormone, the expression of tissue-specific receptors, and downstream effectors. In regard to anole biology, the forthcoming paper by Husak and Lovern is an important step in this direction.

Husak, JF and Lovern MB. In press Variation in steroid hormone levels among Caribbean Anolis lizards: Endocrine system convergence? Hormones and Behavior.

About Thomas Sanger

Thom Sanger is an Assistant Professor at Loyola University in Chicago. His lab specializes on understanding the developmental bases of Anolis lizard diversity.

9 thoughts on “Aggressive Behavior Is Rarely the Result of Circulating Testosterone Levels

  1. Fascinating topic and excellent study!
    Just from a knuckle dragger’s perspective, it would seem to me that higher testosterone levels would increase the drive to mate, not fight. Have the two been found to be intrinsically linked? Perhaps risky or aggressive behavior is driven by evolution and individual survival success.

  2. Such a cool study! For those of you well-versed in sampling testosterone, how stressful a process is it for the lizard, and how long might an animal take to recover from being sampled? It would be super interesting to get repeated testosterone measurements from individuals engaged in different behaviours in the wild…

  3. Armando,
    I hope that some of the people well-versed in behavior will chime in. In the mean time I will take a shot as answering your question. The relationship between circulating hormones and behavior is a vast area of study that is examined in a wide range of vertebrate species. Without writing another review I think that the most direct way to answer your question is that mating and territoriality/aggression are almost certainly not mutually exclusive. I would assume that the “drive to mate” are probably directly responsible for the aggressive behaviors that have been studied by Husak, Lovern, and Johnson.

    Ambika, the authors of this study collected blood from the postorbital sinus. I might be mistaken, but it seems like that level of intrusion would make repeated measures from a single individual extremely difficult.

  4. Yes, Thom, you’ve nailed it. The mechanisms are likely similar (and are for many species that have been studied), and the brain regions controlling the behavior in those contexts are pretty much the same. There are obvious subtle differences, but many of the same displays are used for both contexts, as everyone who visits this blog knows. How the contexts differ in their regulation might be part of the story – we’ll have to see! Lots of potential future studies with these systems, and the Puerto Rican species are particularly interesting because of how different they seem to be..

    With regards to sampling, yes, we use the post-orbital sinus, but I’ve been using a slightly less invasive way to do that which allows for resampling the same individuals (which I do for some things – contact me if interested). The main constraint is getting enough blood each time and still having enough for the individual to survive AND give more for repeated samples. They are resilient animals! After a sample, they blink it off and go right back to what they were doing, though with much higher levels of stress hormones (corticosterone). I’ve seen lizards display at me, as well as other conspecifics minutes after sampling.

  5. I’m not up to speed on this field of study, but I’m not sure I find the results surprising. Seems like test (and corticosterone) has control over enough central fitness components (growth, age at maturity, etc) that test-mediated regulation of aggression is more likely to occur through changes in test sensitivity (receptor expression, sensitivity, response pathways). Different genetic backgrounds of the different species may also complicate things. Of course, the authors already acknowledged all of these possibilities and more. Such an intricate puzzle, but seems like one that may be just elegant enough to tackle!

  6. That’s the prevailing thought, but nobody ever tests it in comparative studies with an adequate number of species. Until someone can show that receptor expression or signaling pathways do change more easily/more often among closely related species than the signal itself it is hand waving. It’s likely pretty true, but we need data, and those types of data have lots of logistical challenges associated with their collection, especially when animals are in remote regions.

Leave a Reply to Ambika Kamath Cancel reply

Your email address will not be published. Required fields are marked *

Optionally add an image (JPEG only)