Oecologia (1993) 94:153-158 Oecologia © Springer-Verlag 1993 Original papers Effects of vertebrate predation on a caviomorph rodent, the degu (Octodon degus), in a semiarid thorn scrub community in Chile P.L. Meserve 1, J.R. Guti6rrez 2, F.M. Jaksie 3 1 Department of Biological Sciences,Northern Illinois University, DeKalb, IL 60115, USA 2 Departamento de Biologia, Universidad de La Serena, Casilla 599, La Serena, Chile 3 Departamento de Ecologia, Universidad Cat61ica de Chile, Casilla 114-D, Santiago, Chile Received: 18 June 1992 / Accepted: 9 February 1993 Abstract. The effects of vertebrate predation have been monitored since 1989 on 16 replicated 0.56 ha study plots in a semiarid thorn scrub community in north- central Chile. Using fences of different heights with and without holes and suspended game netting to alter prin- cipal predator (foxes and raptors) and large rodent her- bivore (Octodon degus) access, four grids each have been assigned to the following treatments: 1) low fencing and holes allowing free access of predators and small mam- mals; 2) low fencing without holes to exclude degus only; 3) high fencing and netting with holes to exclude predators only; and 4) high fencing and netting without holes to exclude predators and degus. Small mammal population censuses are conducted monthly using mark- recapture techniques. Degu population trends during 1989 and 1990 showed strongly but nonsignificantly lower numbers in control plots during months when den- sities were characteristically low (September-November) for this seasonally reproductive species; since March 1991, differences have become persistent and increasing- ly significant. Predators appear to have greater numeri- cal effects when their prey populations are low. Survival times of degus, particularly established adults, were sig- nificantly longer in predator exclusion grids during the 21/2 years of observation; thus, predation also affects prey population structure. Key words: Vertebrate predation - Small mammals - Semiarid zone - Neotropical mammals - Chile Predation has been suggested to be a major biotic inter- action influencing the population biology and assem- blage structure of birds and mammals; yet they have been infrequent subjects of experimental studies. Sih etal. (1985) found only 19/139 (13.7%) experimental studies of predation involved bird or mammals. Most evidence for strong predation effects has therefore come Correspondence to." P.L. Meserve from indirect approaches. Generally these assume three forms; first, the effects of predators are studied in artifi- cial or seminatural enclosures (e.g., Dice 1947; Ambrose 1972; Marti and Hogue 1979; Kotler et al. 1988, 1991; Brown et al. 1988; Longland and Price 1991). Alterna- tively, behavioral responses of prey species to conditions simulating altered predator risk are recorded (e.g., Thompson 1982a, b; Kotler 1984a, b, c; Brown 1989). Finally, comparisons are made between prey numbers, and predator numbers and diet over a period of time in order to infer the strength of the interaction (e.g., Pearson 1964, 1966, 1971 ;Fitzgerald 1977; Erlinge et al. 1983; Keith et al. 1984; Sinclair et al. 1990; Hanski et al. 1991 ; Jaksic et al. 1992). A more direct approach, however, involves experi- mentally altering vertebrate predator densities in the field by removal or exclusion from study areas in order to assess prey responses. Although frequently done with aquatic invertebrates and lower vertebrates (Kerfoot and Sih 1987), there are few examples among birds and mam- mals (i.e., Pearson op.cit.; Marcstr6m et al. 1988, 1989; Desy and Batzli 1989; Desy et al. 1990; Newsome et al. 1989; Pech et al. 1992). Additionally, most experimental field studies with birds and mammals are short-term; yet, responses of vertebrate populations to manipulations may be long- term. For example, smaller Chihuahuan Desert grani- vorous rodents and plants showed protracted responses to exclusion of larger granivores, and consumers, respec- tively (Munger and Brown 1981; Brown and Munger 1985; Brown et al. 1986). Previous work in Chile suggested that competition among numerically dominant small mammal species is relatively unimportant in the mediterranean scrub and semiarid zones (Glanz 1977; Glanz and Meserve 1982; Meserve et al. 1981 a, b; Meserve and Le Bouleng6 1987; Iriarte et al. 1989), and in southern temperate rainforests (Murfia et al. 1987). Conversely, predation seems more important especially in explaining convergent patterns of habitat utilization among small mammals (e.g., Jaksic et al. 1979, 1981; Jaksic and Ostfeld 1983; Jaksic 1986;