Predicting species distribution across time and space is a major challenge for ecologists. As one of the few general ecological laws, Taylor’s powder law states that the log(variance) of the abundance of populations increases linearly with the log(mean) of these populations. Although this law has been confirmed by observational studies on many populations, the log(mean)-log(variance) relationships, particularly the slope, vary considerably among different populations. We know little about the effect of trophic interactions on the manifestation of Taylor’s law.
Here, we present an experimental study to examine the effect of predation and trophic polymorphism on the log(mean)-log(variance) relationships in laboratory protist communities.
We used two closely related bacterivorous protists, Tetrahymena pyriformis and Tetrahymena vorax, as the prey, and Euplotes sp. as the predator. T. vorax develops three morphologically distinct phenotypes that prefer different food resources, whereas T. pyriformis individuals remain uniform. We manipulated the presence/absence of trophic polymorphism by using either T. vorax or T. pyriformis as the prey, and the presence/absence of predator Euplotes sp. We varied the media dilution rate to manipulate the mean population abundance. We monitored the population dynamics of these species/phenotypes for ten weeks. Furthermore, we developed mathematical models to simulate the population dynamics.
Results/Conclusions
Major axis regressions show that the log(variance) increased with the log(mean) in both prey and predator populations in all treatments. Predation and trophic polymorphism, however, influenced the slopes of the log(mean)-log(variance) regressions. When T. pyriformis served as the prey, predation significantly decreased the slope of the log(mean)-log(variance) regressions in the prey populations. This contrasted with a complete lack of effect of predation on the slope of T. vorax populations, in which trophic polymorphism developed. The presence of trophic polymorphism also reduced the slope of the log(mean)-log(variance) regressions in the predator Euplotes sp. populations. Interestingly, residuals from linear regressions of T. pyriformis populations were significantly greater than that of T. vorax populations. Mathematical models suggested that T. vorax populations showed stronger density dependence than T. pyriformis populations. Such higher density dependence in T. vorax populations resulted in a more reliable log(mean)-log(variance) relationship in these populations.
Our results altogether demonstrate the importance of trophic interactions for species abundance and temporal variation.