Local Adaptation in Trinidadian Guppies Alters Stream Ecosystem Structure at Landscape Scales despite High Environmental Variability
While previous studies have shown that evolutionary divergence alters ecological processes in small-scale experiments, a major challenge is to assess whether such evolutionary effects are important in natural ecosystems at larger spatial scales. At the landscape scale, across eight streams in the Caroni drainage, we found that the presence of locally adapted populations of guppies (Poecilia reticulata) is associated with reduced algal biomass and increased invertebrate biomass, while the opposite trends were true in streams with experimentally introduced populations of non-locally adapted guppies. Exclusion experiments conducted in two separate reaches of a single stream showed that guppies with locally adapted phenotypes significantly reduced algae with no effect on invertebrates, while non-adapted guppies had no effect on algae but significantly reduced invertebrates. These divergent effects of phenotype on stream ecosystems are comparable in strength to the effects of abiotic factors (e.g., light) known to be important drivers of ecosystem condition. They also corroborate the results of previous experiments conducted in artificial streams. Our results demonstrate that local adaptation can produce phenotypes with significantly different effects in natural ecosystems at a landscape scale, within a tropical watershed, despite high variability in abiotic factors: five of the seven physical and chemical parameters measured across the eight study streams varied by more than one order of magnitude. Our findings suggest that ecosystem structure is, in part, an evolutionary product and not simply an ecological pattern.
Map showing subwatersheds sampled for landscape scale (denoted by outlined watersheds, where either HP [1–4] or LP [5–8] guppies are located) and exclosure experiments (denoted by a dashed subwatershed) across the Caroni drainage in Trinidad's Northern Range mountains. The Caroni drainage contains large sections of high-density urban areas (diamond shaded area) and rural areas with mixed agriculture (hatch shaded area). Subwatersheds chosen for landscape study were located in watersheds with the largest tracts of intact-forested headwaters. Inset (B) illustrates paired reach design (above and below waterfalls) of the landscape experiment, where each stream contains reaches below barrier waterfalls containing killifish and guppies and reaches above waterfalls that are guppy-free.
Least square means (±SE) of (A) algal biomass and (B) invertebrate biomass from the landscape scale experiment. Symbols show streams with locally adapted low predation (LP) guppies compared with introduced high predation (HP) guppies. We found a significant interaction (P < 0.05) between community and guppy phenotype for both (A) algal biomass and (B) invertebrate biomass, indicating that the effects of guppies depend on the phenotype of the population.
Mean (±SE) response ratios for ecosystem variables measured in two macroconsumer exclusion experiments: low predation phenotype (LP) guppies and killifish present (control) or excluded; high predation phenotype (HP) guppies and killifish present (control) or excluded. Response variables are algal accrual rate (mg m−2 day−1), algal biomass (mg m−2), and invertebrate biomass (mg m−2). Positive and negative response ratios represent positive and negative top-down effects of macroconsumers, respectively. Statistical significance is signified by * (P < 0.05).
Contributor Notes
Associate Editor: D. M. Green.
From “Eco-Evolutionary Dynamics in Cold Blood,” an ASIH-sponsored symposium at the 2016 Joint Meeting of Ichthyologists and Herpetologists in New Orleans, Louisiana.