Lizards reduce spider species richness by excluding rare species

Ecology, March, 1998 by David A. Spiller, Thomas W. Schoener

INTRODUCTION

The relationship between predation and prey species diversity has been a major issue in ecology for several decades. Understanding the relationship is necessary to forecast the community-wide impact of eliminating or introducing predator species. Predictions can be difficult, however, because in addition to a direct negative effect, predators may have a variety of indirect positive and negative effects on prey populations (reviews in Sih et al. 1985, Holt and Lawton 1994, Menge 1995, Abrams et al. 1996). In some systems a single predator species determines most patterns of prey community structure, but in other systems prey communities are controlled by several predator species with each alone having little or no impact (Hixon 1991, Menge et al. 1994). Physical disturbances can also play a dominant role in structuring communities (Sousa 1984, Wootton et al. 1996). Hence, communities are structured by various blends of biotic and abiotic factors. Nevertheless, different types of systems might exhibit different general patterns.

Field experiments have shown that prey species diversity can respond to increasing predation intensity in three different ways (Hixon 1986): (1) no change, (2) a monotonic decrease, and (3) a unimodal response. The first response may occur when the predators are inefficient or are swamped by high prey recruitment (Underwood and Denley 1984, Thorp 1986). When predation is important, indirect effects by predators on prey may determine whether the second or third response occurs. In the simplest case with no indirect effect, the predator has only a direct negative effect on prey species; therefore, diversity can only decline monotonically with predation intensity (e.g., Hixon and Beets 1993).

In the unimodal response, diversity is highest at some intermediate level of predation (Connell 1978, Lubchenco 1978, Hixon and Brostoff 1983). Along a gradient from low to intermediate levels of predation intensity, predators reduce competitive dominants, allowing more prey species to coexist (e.g., Paine 1966). From intermediate to high levels of predation intensity, prey species are eliminated by predation, and diversity declines (e.g., Paine and Vadas 1969). The ascending phase of the unimodal response has been termed "predator-mediated coexistence" (Caswell 1978, Hastings 1978), a type of indirect effect found to be very common in rocky intertidal systems (Menge 1995).

In contrast to predator-mediated coexistence, other types of indirect effects may cause prey species diversity to decrease with predation intensity (review in Holt and Lawton 1994). First, predator-avoidance behavior may cause prey species to compete for a limited number of refuges (Holt 1987) or for a limited food supply within refuges (Mittelbach and Chesson 1987, Persson 1993). This indirect effect intensifies prey competition and promotes exclusion, whereas predator-mediated coexistence reduces competition. Second, "apparent competition" (Holt 1977, 1984) occurs when predators respond numerically to an increase of one prey species and subsequently have a stronger impact on a second prey species. This effect may lead to exclusion of the second prey species. Evidence of apparent competition was common in rocky intertidal studies (Menge 1995). Third, in "intraguild predation" (Polis et al. 1989, Polis and Holt 1992), a predator that consumes species in more than one trophic level may both eat and compete with intermediate-trophic-level prey species (Pimm and Lawton 1978). The combination of these direct and indirect effects, both negative, may be especially deleterious to prey species.

Experimental evidence of predator-mediated coexistence comes almost exclusively from studies in which the prey are either sessile aquatic organisms or terrestrial plants, although a few examples with mobile aquatic prey exist (Sih et al. 1985, Hixon 1986, Menge 1995). Fuentes and Jaksic (1988) suggested that predator-mediated coexistence in terrestrial animal communities is rare because they are relatively closed systems, and because in the absence of predation strong competition among prey species leading to rapid exclusion of most of them is unlikely. Along the same lines, Roughgarden (1986) proposed that the processes structuring marine, space-limited animals and terrestrial, food-limited animals are fundamentally different (also see Schoener 1986).

Experimental introductions of lizards onto small Bahamian islands resulted in rapid and permanent extinction of most rare web-spider species (Schoener and Spiller 1996). Only the single commonest species ever persisted continuously on lizard-introduction islands. Concordantly, manipulation of lizards within enclosures on a mainland showed that lizards reduced the number of web-spider species (Spiller and Schoener 1988, 1994). However, because enclosures with lizards usually contained more than one web-spider species and data on rare species were very sparse, we could not ascertain whether rare species were differentially excluded by lizards. The island-introduction and mainland-enclosure experiments were conducted in very different habitats and species composition of the web-spider communities differed markedly. Does the same pattern of lizard predation - exclusion of rare species - occur under these different conditions? If so, the phenomenon is not dependent on a particular habitat or composition of prey species. We here report the effects on web spiders of manipulating lizards within enclosures continuously for 54 mo. The larger data set on rare species enabled us to ascertain the pattern of predation. Additionally, because in theory indirect effects are expected to become established more slowly than direct effects (Bender et. al. 1984, Yodzis 1989), the present results are a better assessment of the total impact of lizards on the web-spider community. In the Discussion, we compare in detail the results of the enclosure experiment with the island-introduction experiment. To assay the generality of the pattern, we compare our studies to four other recent experimental field studies in terrestrial systems of the impact of vertebrate predators on arthropod species (Joern 1986, Parmenter and MacMahon 1988, Fowler et al. 1991, Bock et al. 1992).