Predator cannibalism can shift prey community composition toward dominance by small prey species

Abstract

Cannibalism among predators is a key intraspecific interaction affecting their density and foraging behavior, eventually modifying the strength of predation on heterospecific prey. Interestingly, previous studies showed that cannibalism among predators can increase or reduce predation on heterospecific prey; however, we know less about the factors that lead to these outcomes. Using a simple pond community consisting of Hynobius retardatus salamander larvae and their associated prey, I report empirical evidence that cannibalism among predators can increase predation on large heterospecific prey but reduce that on small heterospecific prey. In a field-enclosure experiment in which I manipulated the occurrence of salamander cannibalism, I found that salamander cannibalism increased predation on frog tadpoles but reduced that on aquatic insects simultaneously. The contrasting effects are most likely to be explained by prey body size. In the study system, frog tadpoles were too large for non-cannibal salamanders to consume, while aquatic insects were within the non-cannibals' consumable prey size range. However, when cannibalism occurred, a few individuals that succeeded in cannibalizing reached large enough size to consume frog tadpoles. Consequently, although cannibalism among salamanders reduced their density, salamander cannibalism increased predation on large prey frog tadpoles. Meanwhile, salamander cannibalism reduced predation on small prey aquatic insects probably because of a density reduction of non-cannibals primarily consuming aquatic insects. Body size is often correlated with various ecological traits, for instance, diet width, consumption, and excretion rates, and is thus considered a good indicator of species' effects on ecosystem function. All this considered, cannibalism among predators could eventually affect ecosystem function by shifting the size composition of the prey community.

Keywords: Hynobius retardatus; cannibalism; gigantism; predator–prey interaction; top‐down effects.

FIGURE 1

Size differences between salamander larvae, frog tadpoles, and aquatic insects. (a) Photo of salamander larvae, frog tadpoles, and Ephemeroptera larvae. (b) Relationship between body length and gape width of largest salamander in the enclosures on day 34 and relationships between body length and width of prey items. Filled circles represent the largest salamanders in the enclosures. The lines represent the standard deviation of body length and width of prey items

FIGURE 2

(a) Photograph of the 36 enclosures in the field pond in Teshio Experimental Forest of Hokkaido University. (b) Schematic diagram of the experiment. Canni, No‐canni, and No‐sal are abbreviations of Cannibalism, No‐cannibalism, and No‐salamander treatments, respectively

FIGURE 3

Survivorship of the frog tadpoles during the experimental period. The thick horizontal bars represent the median, the box contains 50% of the data, and the whiskers indicate the range. Treatments not sharing the same lowercase letter were significantly different from each other (Tukey's HSD: p < .05). See Figure 2 for abbreviations

FIGURE 4

The numbers of (a) Ephemeroptera, (b) Diptera, and (c) Hemiptera remaining in the enclosures over the experimental period. Error bars denote standard error. The shaded area in the figures represents the period when cannibals, non‐cannibals, frog tadpoles, and aquatic insects coexisted within enclosures. The total number of (d) emerging Ephemeroptera and (e) emerging Diptera was collected during the experiment. The thick horizontal bars represent the median, the box contains 50% of the data, and the whiskers indicate the range. Treatments not sharing the same lowercase letter were significantly different from each other (Tukey's HSD: p < .05). See Figure 2 for abbreviations