The vast array of signals used in animal communication is a continuous source of awe and a hot topic in evolutionary and behavioral research. One important factor contributing to the signal diversity we witness today is ‘signal efficacy’: the ability of a signal to travel efficiently through the environment and attract the receiver’s attention. With this in mind, natural selection is expected to mold signal design for maximum efficacy of information transmission and detectability, leading to signal variation among populations/species living in different environments. To illustrate, a recent study by Tess Driessens and colleagues assessed the degree of variation in the dewlap design of Anolis sagrei by comparing 17 populations distributed across the Caribbean (Fig. 1).
Their findings showed large interpopulational variation in dewlap size, pattern, and color, and more interesting, they established a link between the dewlap design of brown anoles and the environment they live in. Lizards occurring in more ‘xeric’ environments had a higher proportion of solid dewlaps with a higher UV reflectance; lizards inhabiting ‘mesic’ environments had predominantly marginal dewlaps showing high reflectance in red. This was true for both males and females. Like Ng et al. (2011) and their observations on dewlap variation in A. distichus across an environmental gradient, Driessens et al. (2017) interpret their findings as evidence for adaptive divergence of a signaling apparatus.
Surprisingly though, while there are numerous great examples of comparative studies finding support for convergent evolution in visual and acoustic signaling systems, (e.g. Endler 1992; Fleishman 1992; Nicholls & Goldizen 2006, to say a few), similar (comparative) studies, but then, on the phenotype of chemical signals are almost entirely lacking. This is probably due to the combination of only very recent developments in chemical analytical and statistical comparative tools, the time researchers need to assemble a large-scale multi-species chemical dataset, and perhaps due to our own predisposition to visual and auditory signals. Currently, the proper analytical tools for studying natural products chemistry are available and affordable, permitting comprehensive taxon-wide research on the evolution of chemical signal diversity and design. Ultimately, there has never been a better time as now to be a comparative chemical ecologist.
Finally, three Belgians, two Spaniards and one Greek (sounds like the start of a joke with ample potential) took up the challenge to examine variation in the chemical signal design of lizards. Although underrepresented in studies on chemical signal diversity, lizards are an excellent group for investigating chemical signal evolution, as many of them they bear numerous glands on their thighs that secrete waxy substances, which they deposit while moving through their habitat. These secretions are often considered the leading source of chemical signals involved in lizard communication.
The study started with a quest. A quest to collect gland secretions of as many species as possible (within a PhD timeframe). Luckily, we were fortunate enough to be able to count on the help of many collaborators (Shai Meiri, Chris Broeckhoven, …). We focussed on lacertid lizards, as they are a species-rich family distributed over a wide geographical area, and known to rely strongly on chemical communication in several contexts.
In total, we sampled secretions from 64 species throughout, Europe, Africa, and Asia, covering a wide array of habitats and climate regions: from the Mediterranean maquis over the alpine meadows in the Pyrenees Mountains, to the sandy Israeli dunes and the Kalahari Desert of South Africa (Fig. 2). Back in the lab, we determined the chemical composition and chemical ‘richness’ (number of different chemical compounds) of the secretions using GC-MS, and obtained climate data for all catch-localities from online databases.
Our gathered data showed considerable variation in the chemical richness and composition of lacertid secretion. Shared-ancestry failed to explain among-species patterns of variation, hinting that chemical signals may change relative rapidly. Most interestingly, our findings revealed a strong relationship between the environmental conditions species live in and the chemical composition of their glandular secretions. On the one hand, lizards living in ‘xeric’ environments, characterized by high temperatures and arid conditions, contained higher proportions of stable and heavy-weight compounds in their secretions. Hot and dry conditions increase the evaporation rate of chemicals, and so, decreasing the longevity of a signal. Stable and heavy-weight compounds most likely reduce evaporation rate and counteract the rapid signal fade-out through evaporation, generating a highly persistent scent-mark. On the other hand, species inhabiting wet, humid conditions produced highly aromatic and low-weight secretions containing numerous different compounds. This chemical mix probably creates a volatile-rich signal that can be used for long-distance airborne communication.
While we cannot deny that these findings of convergent evolution in the design of chemicals signals are fascinating, some would say this outcome is not unexpected.
“[…] a cadre of scientists has taken the […] view, that convergence is the expectation, that it is pervasive, and that we should not be surprised to discover that multiple species […] have evolved the same features to adapt to similar environmental circumstances. From this perceived ubiquity, the scientists draw a broader conclusion: evolution is deterministic, driven by natural selection to repeatedly evolve the same adaptive solutions to problems posed buy the environment. — J. Losos (Improbable Destinies, p. 33)
Nonetheless, I am confident to state that using by far the largest comparative dataset amassed to-date to examine patterns of chemical signal divergence, we have provided strong evidence for a significant relationship between chemical signal design and prevailing environmental conditions, which may results from differential selection on signaling efficacy (Baeckens et al. 2017).