Ecology, 93(5), 2012, pp. 1194–1203 Ó 2012 by the Ecological Society of America Do predators control prey species abundance? An experimental test with brown treesnakes on Guam EARL W. CAMPBELL III, 1,5 AMY A. YACKEL ADAMS, 2,6 SARAH J. CONVERSE, 3 THOMAS H. FRITTS, 4,7 AND GORDON H. RODDA 2 1 Ohio Cooperative Fish and Wildlife Research Unit, 1735 Neil Avenue, Columbus, Ohio 43210 USA 2 U.S. Geological Survey Fort Collins Science Center, 2150 Centre Avenue, Building C, Fort Collins, Colorado 80526 USA 3 U.S. Geological Survey Patuxent Wildlife Research Center, 12100 Beech Forest Road, Laurel, Maryland 20708 USA 4 U.S. Geological Survey, National Museum of Natural History, MRC 11, Washington, D.C. 20560-0001 USA Abstract. The effect of predators on the abundance of prey species is a topic of ongoing debate in ecology; the effect of snake predators on their prey has been less debated, as there exists a general consensus that snakes do not negatively influence the abundance of their prey. However, this viewpoint has not been adequately tested. We quantified the effect of brown treesnake (Boiga irregularis) predation on the abundance and size of lizards on Guam by contrasting lizards in two 1-ha treatment plots of secondary forest from which snakes had been removed and excluded vs. two 1-ha control plots in which snakes were monitored but not removed or excluded. We removed resident snakes from the treatment plots with snake traps and hand capture, and snake immigration into these plots was precluded by electrified snake barriers. Lizards were sampled in all plots quarterly for a year following snake elimination in the treatment plots. Following the completion of this experiment, we used total removal sampling to census lizards on a 100-m 2 subsample of each plot. Results of systematic lizard population monitoring before and after snake removal suggest that the abundance of the skink, Carlia ailanpalai, increased substantially and the abundance of two species of gekkonids, Lepidodactylus lugubris and Hemidactylus frenatus, also increased on snake-free plots. No treatment effect was observed for the skink Emoia caeruleocauda. Mean snout–vent length of all lizard species only increased following snake removal in the treatment plots. The general increase in prey density and mean size was unexpected in light of the literature consensus that snakes do not control the abundance of their prey species. Our findings show that, at least where alternate predators are lacking, snakes may indeed affect prey populations. Key words: abundance; Bayesian model selection; binomial mixture model; Boiga irregularis; Carlia ailanpalai; Emoia caeruleocauda; geckos; Hemidactylus frenatus; Lepidodactylus lugubris; predation; removal experiment; skinks. INTRODUCTION To what extent are terrestrial vertebrate populations depressed by their predators? A limited example is that exhibited by extinctions of prey, usually on islands, caused by invasive introduced predators. The brown treesnake’s (Boiga irregularis) extirpation of the native forest birds of Guam (Savidge 1987) demonstrates that a predator can eliminate vulnerable prey, but what does that nonequilibrium outcome (extinction) suggest about the interactions of coexisting predator and prey popu- lations? Perhaps coexistence reflects a prey species’ relative insensitivity to predation pressure. Under what sustainable conditions, then, are prey populations appreciably depleted by their coexisting predators? This classic question is most convincingly answered by experimental manipulation of predator presence, which is difficult to do. As noted by Sih et al. (1985:289), ‘‘... virtually no studies have manipulated predators of vertebrate prey.’’ Salo et al. (2007) conducted a meta- analysis of replicated terrestrial studies and found only 45; however, none of these involved either ectothermic predators or ectothermic prey (but see Schoener et al. [2002] discussed below in the fourth paragraph of the Introduction). Studies focused on endothermic predators and prey represent a subset of vertebrate predator–prey relation- ships that incorporate low energetic assimilation efficien- cies and relatively high food intake requirements (Pough 1980). Nowak et al. (2008) argued that endo- and ectothermic predation energetics are fundamentally different; endothermic predators’ food needs are much higher, and the incessant need for food drives a much higher level of activity among both predator and prey. Manuscript received 27 July 2011; revised 9 November 2011; accepted 1 December 2011. Corresponding Editor: J. B. Yavitt. 5 Present address: USFWS, Pacific Islands Fish and Wildlife Office, 300 Ala Moana Boulevard, Room 3-122, P.O. Box 50088, Honolulu, Hawaii 96850 USA. 6 Corresponding author. E-mail: yackela@usgs.gov 7 Present address: 24 Madrone Flyway, Belen, New Mexico 87002 USA. 1194