Eco-Evolutionary Effects on Population Recovery Following Catastrophe

As a new PhD student I had read many of the classic studies confirming the role of natural selection in generating adaptive divergence between guppies that coexist with fish predators (high-predation populations) and those that did no (low-predation populations), and I was very excited to experience this system firsthand as I started my dissertation research. However, when I arrived in Trinidad for my first full field season, I was horrified to discover that my study organism was nowhere to be found. Intense and prolonged flooding resulted in near complete decimation of the guppy populations inhabiting the high-predation (downstream) sections of many of the famous guppy rivers in Trinidad. The "millions fish" (a local Trinidadian name for the guppy), seemed like quite a misnomer after it took us the better part of a week to catch a couple dozen guppies from our focal study stream (the Marianne River). This was problematic for me because I had intended to study how geneflow from low-predation populations constrained adaptation of high-predation guppies (of which there now were almost none). After the initial panic subsided, this situation inspired my advisor (Michael Kinnison) and me to think about eco-evolutionary dynamics. How would contemporary evolution (in the form of selection against migrants) influence recovery of the high-predation populations? Would reproduction by the few remaining high-predation guppies (which were locally adapted) drive recovery, or, would the low-predation immigrant guppies make a significant demographic contribution despite lacking many of the anti-predator characteristics of high-predation guppies? In the March 2011 special issue of Evolutionary Applications ("In the light of evolution: interdisciplinary challenges in food, health, and the environment"), myself and several co-authors describe research where we attempt to answer some of these questions.

Intense flooding decimated high-predation guppy populations (notice the scourning on the bank of the Marianne River) and negated my dissertation research proposal.

Metapopulation theory provided some basis for anticipating an important role for contemporary evolution in this circumstance. One important concept within metapopulation ecology is the population rescue effect, where immigration from nearby populations prevents the extinction of a local population subjected to a catastrophic disturbance or chronically harsh environmental conditions. Of course, the ability of migrants to contribute to population growth depends on their ability to survive and reproduce; if the phenotypes of migrants are poorly-suited to local conditions, then any population rescue will be mitigated by selection against migrants. Several studies have attempted to model these eco-evolutionary dynamics and have found that migration can variously facilitate or prevent extinction, depending on (among other things) the fitness and number of migrants. Few studies have actually quantified these eco-evolutionary dynamics in the wild, which is unfortunate because many conservation and restoration programs rely on supplemental stocking from captive populations or natural recolonization from nearby populations (in either case we might expect migrants to have lower fitness than residents).

To assess the eco-evolutionary consequences of selection against migrants we performed an experimental introduction of both high- and low-predation guppies (each individually marked) into a focal high-predation site (where local high-predation guppies had been completely wiped-out). Consistent with our expectations, we found that both male and female low-predation guppies had very poor survival compared to high-predation guppies. We surveyed this experimental population for approximately 3 months which allowed us to sample the offspring of the original experimental fish introduced into our site, and (using population genetic assignment tests) quantify the demographic contribution of each ecotype. We found a large difference in the demographic contribution of each ecotype to population recovery. Compared to a population model based on purely “ecological” expectations (assuming no fitness differences between ecotypes), the demographic cost of selection against migrants was very high (around 45% in two different years). We describe these eco-evolutionary dynamics as “cryptic” because they resulted from a fleeting environment perturbation, and caused no net phenotypic change in the local population. Nonetheless, the consequences of eco-evolutionary interactions in this system were profound for population recovery. Such “cryptic” dynamics suggest that eco-evolutionary interactions may be quite common in nature and generally relevant to conservation and restoration efforts.