Disentangling the divergence of morphology and life historySubmitted by editor on 24 February 2020.Get the paper!
Stalactite blue hole from above; Photo credit: R Brian Langerhans.
Environments are complex, with a plethora of different abiotic and biotic components. We inevitably try and simplify this complexity in an attempt to understand how environmental factors drive trait evolution. This is certainly true for a lot of our own previous work, where we have, for example, investigated how organisms respond to differences in predation or to the presence/absence of natural toxicants. However, this focus on a single strong environmental factor neglects the multivariate nature of natural environments—and multiple selective agents likely underlie most adaptation. This problem is compounded by the fact that different aspects of an environment often covary among localities. For example, if aquatic predators are differentially sensitive to higher salinities, then differences in salinity between environments are bound to covary with differences in predation pressures. This can make it almost impossible to tease apart the separate roles of different environmental factors in influencing phenotypic variation among populations or species.
To make matters worse, different aspects of an organism’s phenotype are not necessarily independent from each other either. For example, due to genetic (e.g., linkage or pleiotropy), developmental, or architectonic interdependencies, changes in certain life histories might be tied to certain morphological changes, or vice versa. An obvious example would be that increased investment into reproduction (e.g., larger reproductive organs) can have a direct effect on external body shape because more body cavity space is needed to house these larger reproductive tissues. Nonetheless, both of these phenotypic trait suites are traditionally studied independently.
Here, we employ an underutilized multivariate method (partial least squares structural equation modelling) to disentangle the relative contributions of three different environmental drivers (predation, resource availability, and population demographics) on patterns of morphological and life-history divergence in Bahamas mosquitofish (Gambusia hubbsi; Poeciliidae) inhabiting inland blue holes (i.e., flooded vertical caves) in The Bahamas. This system is particular well-suited for this approach, as blue holes with divergent predation regimes do not systematically differ in these or other environmental variables. We further employ a novel approach by using two-block partial least squares analysis to uncover to what extent divergence in life histories and morphology are integrated and independent.
Male Gambusia hubbsi in Stalactite blue hole; Photo credit: R Brian Langerhans.
To conduct this kind of study, however, we first have to go to blue holes and measure the environmental variables, as well as collect the fish for the measurement of phenotypes. For this study, we examined 14 blue holes across the northern section of Andros Island. Getting to many of these blue holes is not exactly a walk in the park. A couple involve only relatively short hikes after 1-2 hours of driving down old logging roads through the unpopulated interior of the island, provided we avoid catching a flat tire (flat tire count since 2004: 21). But many others involve not only long drives, but also long, slow hikes over difficult terrain, machete-chopping our way through an understory of poison wood trees while carrying our gear. Andros Island is a carbonate island with a pronounced karst landscape, resulting in many sharp rocks and rocky outcrops that can tear through trousers and even shoes – at the very least you end a day of field work often with bloody scratches all over your legs and ankles. Thus, it is virtually impossible to escape unscathed from the ire of poison wood or jagged karst under these circumstances, and one of us suffered quite a special case during this field work!
Photo credits: C Branan, R Martin, B Langerhans.
But in the end, we identified the key environmental factors underlying joint and independent trait divergence. Phenotypic divergence in life histories and body shape resulted mainly from differences in predation regimes, but that both demography and resource availability were also important for certain aspects of phenotypic differentiation. Furthermore, phenotypic divergence between predation regimes involved both integrated (i.e., covariation) and independent life-history and morphological responses. In females, for example, a large proportion of one aspect of the morphological shift between predation regimes, an abdominal distension in the presence of predators, was the direct result of changes in fecundity (which was higher in high-predation blue holes).
Our results strengthen our understanding of what selective agents are most important in driving complex trait variation, which specific traits might respond to which specific selective factors and to what extent might different suites of phenotypes respond jointly and separately.
Authors: Rüdiger Riesch, Ryan A. Martin and R. Brian Langerhans