A Long-Term Study of Vertebrate Predator Responses to an El Niño (ENSO) Disturbance in Western South America Fabian M. Jaksic; Sergio I. Silva; Peter L. Meserve; Julio R. Gutiérrez Oikos, Vol. 78, No. 2. (Mar., 1997), pp. 341-354. Stable URL: http://links.jstor.org/sici?sici=0030-1299%28199703%2978%3A2%3C341%3AALSOVP%3E2.0.CO%3B2-%23 Oikos is currently published by Nordic Society Oikos. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/oikos.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org. http://www.jstor.org Fri Nov 9 14:04:54 2007 OIKOS 78: 341-354. Copenhagen 1997 A long-term study of vertebrate predator responses to an El Nifio (ENSO) disturbance in western South America Fabian M. Jaksic, Sergio I. Silva, Peter L. Meserve and Julio R. Gutierrez Jaksic, F. M., Silva, S. I., Meserve, P. L. and Gutierrez, J. R. 1997. A long-term study of vertebrate predator responses to an El NiBo (ENSO) disturbance in western South America. - Oikos 78: 341-354. We analyzed the putative effects of the El Niiio Southern Oscillation (ENSO) of 1991-92 in a semi-arid locality of northern Chile. We obtained 30 months of pre-ENS0 data, followed by 36 months of peak and post-ENS0 data (total = 5.5 yr). The rainy winter of 1991 resulted in a three-fold increase in total seed bank (perennial and ephemerals pooled) and in ephemeral (but not perennial) herb cover. Seed and herbage eaters (rodents) irrupted to population levels ca 20 times higher during the breeding season of 1991 than the preceding wintering season. Diurnal carnivorous predators (hawks, owls, and foxes) showed a delayed response to the irruption, increasing from seven individuals sighted during the wintering season of 1991 to 13 during the wintering season of 1992. A seemingly counterclockwise trajectory of predator abundance versus prey levels suggested a pattern of prey-driven dynamics, but confidence intervals were likely broad. In this semiarid locality, it appears that ENS0 effects did not cascade down from higher to lower trophic levels, but rather the opposite. In this bottom-up scenario, we predict that as primary productivity varies with rainfall, so should secondary (mammal prey densities), and tertiary productivity (vertebrate predators). Long-term monitoring of this terrestrial ecosystem is needed to test this prediction. F. M. Jaksic and S. I. Silva, Departamento de Ecologia, P. Universidad Catdlica de Chile, Casilla 114-0, Santiago, Chile. - P. L. Meserve, Dept of Biological Sciences, Northern Illinois Univ., DeKalb, IL 60115, USA. - J. R. Gutiirrez, Departamento de Biologia, Universidad de La Serena, Casilla 599, La Serena, Chile. The El Niiio Southern Oscillation (ENSO) is an hemispheric-scale phenomenon that originates off the northeastern coast of Australia, and causes multiple effects across the Pacific Ocean on the western coastline of the Americas (Diaz and Markgraf 1992). Although generally studied in atmospheric and oceanographic contexts, some associated biological phenomena have been reported (Barber and Chavez 1983), particularly geographic and bathymetric shifts in marine life, including marine birds (e.g., Guerra et al. 1988), and more sporadically, changes in terrestrial floral composition (e.g., Dillon and Rundel 1990, Villagran 1993). ENS0 effects on terrestrial fauna have long been suspected, but poorly documented. Accepted 3 July 1996 Copyright O OIKOS 1997 ISSN 0030-1299 Printed in Ireland - all rights reserved Among those putative effects, outbreaks (or irruptions, "ratadas") of small mammals have frequently been reported in association with ENS0 disturbances in semi-arid regions of western South America (Hershkovitz 1962). Several researchers noted the mammal outbreaks that occurred in southern Peru and northcentral Chile (Table 1). All of them concurred that these outbreaks were triggered by unusually high rainfall (Table 1). Indeed, Fuentes and Campusano (1985) found an overall significant association between high rainfall years and mammal outbreaks throughout north-central Chile. The mechanism linking these two phenomena appears to be increased primary productivity, both as herbage production and seed Table 1. Mammal outbreaks and rainfall patterns in western South America. Place and date of outbreak Coastal Peru, 1972 Fray Jorge, Chile, 1972 La Serena, Chile, 1972 Auco, Chile, 1987 Fray Jorge, Chile, 1991 Chief mammal respondent Year's rainfall Mean yearly rainfall Previous year rainfall Authority Phyllotis darwini PhyNotis darwini, Akodon olivaceus Oligoryzomys longicaudatus Phyllotis darwini, Akodon olivaceus, Thylamys elegans Ph.vllotis darwini, Akodon olivaceus, Octodo~zdegus 81 mm 255 mm 32 mm 69 mm no data 72 mm 174 mm 82 mm 80 mm Pearson (1975) Fulk (1975), Meserve and Le Boulenge (1987) PCfaur et al. (1979) 513 mm 206 mm 158 mm Jimenez et al. (1992) 233 mm 85 mm 32 mm Meserve et al. (1995) bank storage (Armesto et al. 1993, Gutierrez et al. 1993a, b). Here we scrutinize the long-term database accumulated in Fray Jorge (north-central Chile), for primary production and small mammal abundance, and analyze the putative effects of the 1991-92 E N S 0 disturbance on the higher trophic level, that of vertebrate predators. The theoretical framework of our study refers to three major components of vertebrate predation: numerical, functional, and guild responses shown by predators when faced with fluctuating mammalian prey populations. When prey abundance increases, predators may respond numerically by immigrating from peripheral areas and (or) reproducing more successfully in situ (Solomon 1949). When prey abundance declines, predators may decrease by emigrating, failing to reproduce, or dying. Functional responses, on the other hand, involve changes in the relative numbers of prey eaten by individual predators as prey densities vary (Holling 1959). Here, we are interested in a particular expression of functional response, the prey switching behavior of predators (Murdoch 1969, Murdoch and Oaten 1975), whereby they change their selection of prey depending on the relative frequency that prey are encountered in the field. The third component refers to the trophic guild structure of predators (Marti et al. 1993). In this case, we are interested in assessing whether guild composition and packing differs between pre- and post-ENS0 changes in prey levels. We test whether numerical, functional, and/or guild responses are displayed by predators faced with fluctuating prey resources brought by the 1991-92 E N S 0 disturbance at Fray Jorge. Study site and methods Study site The study site is located in a semi-arid mediterranean scrub community in Quebrada de las Vacas (240 m elevation), an interior valley in Fray Jorge National Park (30°38'S, 71°40'W), north-central Chile. This low elevation scrub zone contains a homogeneous cover of 342 spiny drought-deciduous and evergreen shrubs, and a herbaceous understory, on a predominantly sandy substratum. The most characteristic plant association is that of Porlieria chilensis - Proustia pungens - Adesmia bedwellii (Muiioz and Pisano 1947), with mean shrub cover averaging 59% (Meserve 1981a, b). The semi-arid mediterranean climate of this site has 90% of the mean 85 mm annual precipitation falling between May and September (from late autumn to early spring in the southern hemisphere), and warm, dry summers. Prey abundance Since May 1989, a large scale experimental manipulation has been in progress at the study area, involving the selective exclusion of vertebrate predators and (or) large-sized small mammal herbivores from 16 fenced 75 by 75 m (0.56 ha) plots. Small mammal populations are being monitored on all plots with grids consisting of 5 rows by 5 columns (i.e., 25 stations at 15-m intervals) trapped for 4 dlmonth with two Sherman-type live traps/station. Standard mark and release techniques (Meserve and Le Boulenge 1987) are used with data taken on small mammal species, number, trap station, sex, reproductive condition, and weight. Meserve et al. (1993a, 1996) provide full details on the rationale and layout of the experimental setup. For the purposes of this report, small mammal abundance was determined from monthly live-trapping in the four control grids open to both predators and small mammal species (potential prey). Determinations of minimum number known alive (MNKA) were obtained using the CMR capture-recapture program (Le Boulenge 1985, Meserve and Le Boulenge 1987). Trappability was high (> 90%) for most species, enabling use of enumeratlon techniques. Monthly MNKA estimates on the four control grids were averaged together and then averaged again for the six-month periods corresponding to two major biological seasons: breeding season (SeptemberFebruary) and wintering season (March-August), for generation of "expected" frequencies of small mammals in predator diets (see Jaksic et al. 1993a and below). OIKOS 7 8 9 (1997) Predator abundance Starting March 1989 (wintering season), a quantitative estimate of the abundance of diurnal predators (seven hawks, one diurnal owl, one fox species) was obtained by recording the number of different individuals observed daily at the study site. Three nocturnal owl species were monitored by collecting their regurgitated pellets and - starting in early 1990 - by their calls, which only allowed records of their presence/absence. Approximately 750 ha in the vicinity of the trapping grids in Quebrada de las Vacas were covered by this tally, conducted simultaneously with the trapping sessions (i.e., 4 d/month). Predator food habits Fresh raptor pellets and fox feces were collected on a monthly basis starting March 1989 (wintering season). Regurgitated pellets of both diurnal and nocturnal raptors were collected under known roosts, nests, or perches in a 750-ha area approximately centered in the trapping grid area. Feces of foxes were collected in and around trapping grids (see below). Prey were identified with a dissecting scope, usually to the species level, using keys (Reise 1973) and locally collected voucher specimens. The minimum number of individual prey items in pellets and feces was estimated on the basis of paired or unique anatomical elements such as crania, mandibles, teeth rows, wings, elytra, antennae, stings, etc. (Marti 1987). Because all local predators fed on vertebrates and large arthropods (see below), we assume that our tabulations are essentially complete representations of predator diets. Predator numerical response Because we obtained data on the temporal course of small mammal population numbers, we can assess the numerical response of local mammal-eating predators to changes in those prey levels. The non-parametric Spearman coefficient was used to assess the degree of correlation between those two variables (Siegel and Castellan 1988). Predator selectivity Chi-square goodness-of-fit tests were run between observed frequencies of small mammals in predator diets, and the expected frequencies generated from trapping data (cf. Pearre 1982, Jaksic 1989a). Small mammal species were pooled whenever necessary to obtain expected frequencies > 5, as required by the Chi-square test (Sokal and Rohlf 1981). When a given mammal species appears in a predator's diet more than expected from its field abundance as estimated by trapping, we say that this prey species is "selected, meaning that it is overrepresented in the predator's diet. Similarly, a prey is "underconsumed" (we prefer not to imply that it is "avoided") when it is found underrepresented in a given predator's diet. Predator selectivity was assessed on a biological-season basis (i.e., 6-month periods). To take advantage of the long-term aspect of our study, we analyzed these selectivity data as a time series. Departures between prey frequencies in diets and in the field could go either way, from overrepresentation (positive departure = selectivity) to underrepresentation (negative departure). The time series was analyzed in terms of the sign of the departures by means of the binomial test (Siegel and Castellan 1988). The series was examined even when yielding non-significant values, searching for interesting patterns such as selectivity before or after the small mammal irruption that occurred during the breeding season of 1991. We did this because a short series of positive values followed by one of negative values (or the reverse) yields a statistically non-significant result with the binomial test, but is suggestive of a prey selectivity shift on the part of the predators. Predator guild structure We computed diet overlap between any two predator species (cf. Marti 1987) as: where p, and q, are the relative occurrences of prey category i in the diets of the two predators, respectively. Ov ranges from 0 to 1 (0-100% overlap). We applied this equation to diet data using the highest possible taxonomic resolution of prey (plant foods were not considered): species for vertebrates and orders for invertebrates (cf. Greene and Jaksic 1983). Using entries in the diet similarity matrices constructed for calculating Ov, we applied the unweighted pair-group clustering method with arithmetic averaging (UPGMA, Sneath and Sokal 1973). Following Marti et al. (1993) and in order to compare results from throughout the study period, we set the threshold for assigning predators to trophic guilds at 50% diet similarity. Less arbitrary means of defining guilds (e.g., Jaksic and Medel 1990) would result in thresholds for guild designations that varied seasonally, rendering comparisons of guild structure between seasons and among years much more difficult (Jaksic et al. 1993b). Statistical comparisons between the levels at which guild nodes (branches) formed during pre- and post-ENS0 years were made with the Mann-Whitney U test (Siegel and Castellan 1988, see Jaksic et al. 1993b for a similar approach). .;ij C 2 5 Fig. 1. Seasonal precipitation and small mammal density (n = 6 months/season) at Fray Jorge National Park, north-central Chile, throughout 11 successive biological seasons (W = wintering season, March through August; B = Breeding season, September through February). The horizontal line indicates the most recent 10-yr average (85 mm). U) C 8 Results Predator abundance Prey abundance The most abundant diurnal raptors were the blackchested eagle (Geranoaetus melanoleucus) and the Chimango caracara (Milvago chimango). These two species were observed flying over (but infrequently perching) during virtually all seasons (Table 3). Harris hawks (Parabuteo unicinctus), red-backed hawks (Buteo polyosoma), and American kestrels (Falco sparverius) were less frequently observed flying over the study site, and seldom perching. They were absent from the study site during whole seasons. The remaining diurnal raptors, peregrine falcon (Falco peregrinus) and Aplomado falcon (Falco femoralis), were of sporadic occurrence at the study site, each sighted only during a few seasons. Four species of owls were detected at the study site, but only the diurnal burrowing owl (Speotyto cunicularia) could be censused (Table 3). The nocturnal great horned owl (Bubo virginianus) and barn owl (Tyto alba) were heard, and their pellets were collected through all seasons in the field (see below). The pygmy owl (Glaucidium nanum) was frequently heard but their pellets were not found in all seasons (see below). Despite Fulk's (1976) report, we did not see or hear short-eared owls (Asioflammeus) at the study site. Only one species of mammalian carnivore was observed at the study site (Table 3), the Culpeo fox (Pseudalopex culpaeus). Reptilian predators were sporadically seen and consisted of only the long-tailed snake (Philodryas chamissonis) and the Chilean racerunner (Callopistes palluma), both diurnally active during summers. Total abundance of small mammals was positively associated with rainfall (Fig. 1). They were apparently declining from the wintering season of 1989 and reached their lowest abundance in the wintering season of 1991. That season, rainfall exceeded 200 mm, and small mammal abundance during the breeding season of 1991 increased to almost 20 times their previous abundance. They continued increasing until the wintering season of 1993, and started declining again, reaching during the wintering season of 1994 an abundance level similar as that observed during the breeding season of 1991. That a lag existed between winter rainfall pulses and population responses may be related to the fact that local small mammals start reproducing during early spring (Fulk 1976, Meserve and Le Boulenge 1987). Seven species of small mammals (six rodents, one marsupial) were captured throughout the study period (Table 2). Octodon degus, Phyllotis darwini, and Akodon olivaceus were the most abundant species at the site. They accounted for an overall 36%, 24%, and 23% of small mammal captures, respectively, and were present throughout the study (Table 2). The remaining four species together accounted for only 17% of total captures. Of these, Abrocoma bennetti, Abrothrix longipilis, and Oligoryzomys longicaudatus were not captured at all during entire seasons. The only marsupial, Thylamys elegans, although never abundant, was present throughout the study period. Table 2. Small mammals trapped at Fray Jorge National Park, north-central Chile. Density (numberlha) and percentage of total sample that season came from estimates of minimum number known to be alive. Tabular entries are means of six monthly assessments made in four 0.56-ha live-trapping grids. W = wintering season (March through August), B = Breeding season (September through February). Species W89 B89 W90 B90 W91 B91 W92 B92 W93 B93 W94 Mean ( n = 11) Abrocoma bennetti Abrothrix longipilis Akodon olivaceus Octodon degus Oligoryzornyslongicaudatus Phyllotis darwini Thylamys elegans 1.1 8.7 14.5 53.0 0.2 20.3 2.2 18.7 0.0 4.9 14.0 37.1 5.3 26.5 12.1 11.0 1.3 2.6 11.6 36.7 10.6 29.5 7.8 16.1 4.8 0.0 4.8 38.5 0.0 44.4 7.5 7.8 1.5 0.0 6.1 47.0 0.0 33.3 12.1 2.8 0.6 4.2 60.6 3.2 8.9 19.2 3.4 49.1 0.1 8.1 44.1 5.9 18.8 19.3 3.6 102.5 1.1 8.5 45.6 20.3 1.3 21.9 1.3 132.3 1.3 9.6 28.1 34.5 1.0 24.2 1.2 133.8 2.6 11.3 14.4 46.2 0.8 22.6 2.1 88.6 1.2 11.1 9.7 70.5 0.0 3.9 3.7 43.0 1.4 6.3 23.0 35.7 4.3 24.1 5.2 55.1 Density (No./ha) Predator numerical response As shown the four owl and One fox 'pecies studied at the site are generally mammal eaters. Although we did not collect data On the remaining predators, obtained Over seven years at the nearby site of Auco indicate that the Harris hawk (63-lo0% of numerica1 prey hawk (41-70%), black-chested eagle (33-63%)3 and American kestrel (0- 17%) are mostly to partially mammal-eaters (Jaksic et al. 1996). Chilean racerunners in Auco are essentially insectivorous (84% of its diet by numbers), with mammals comprising only 9% of prey occurrences over one year of study (Castro et al. 1991). However, they may sometimes prey extensively on small rodents (Mellado 1982). Also at Auc6, Aplomado falcon was reported to prey mostly on birds and not at all on mammals over one breeding season (Jimenez 1993). Chimango caracaras from Pudahuel preyed mostly on insects (87% of its diet by number) and rarely (2%) on mammals (Yaiiez et al. 1982), but their biomass contribution was not that negligible. Longtailed snakes from all over Chile have been reported (Greene and Jaksic 1992) prey On and (80% number), and less On ('2%). The only reda at or Present in Fray Jorge for which there are no dietary data is the peregrine falcon, likely a bird-eater (Jaksic, pers. obs.). In summary, all predators present at our study site (except for two of the three falcons) depend to some extent on mammal Prey. Therefore, disregarding both peregrine and A ~ l o m a d O we pooled our seasonal censuses of predators, and evaluated whether they collectively varied in numbers together with their presumed prey resources. Fig. 2 indicates that this was the case. Early during the study, predators were declining, in phase with decreasing mammal densities. The irruption during the breeding season of 1991 apparently fueled an increase in the number of predator individuals sighted at the site (Fig. 2). Their increase lasted until the wintering season of 1993, when they abruptly declined together with their main prey. At the end of our study (wintering season of 1994), predator abundance was similar to that observed in the beginning. A seemingly counterclockwise trajectory of predator abundance in association to varying prey levels suggests a prey-driven dynamics (Jaksic et al. 1996). Nevertheless, the confidence intervals associated with estimates of both predator and prey abundance are unknown and likely broad, thus rendering this conclusion rather speculative, Spearman's r was 0,738 (n = 11, P = 0.0095). Predator food habits Overall3 4672 pellets and 2180 feces were c011ected5and 16 660 prey items were identified. The presence of plant fruit seeds) in fox feces was observed (cf. Castro et al. 1994). Pellets of the four owls were collected, but only those of great horned owls, barn owls, and burrowing owls were found in all seasons. Pellets of pygmy owls were less frequent, and those of diurnal raptors were found only exceptionally. Our observations indicate that the owls exhibited a much higher site fidelity than the wide ranging diurnal raptors. The roosting and nesting sites of local owls were all within 4 km of the trapping grids, Roosting and nesting sites of diurnal raptors were located at longer distances from the study site, thus accounting for the scarcity of pellets found, Foxes provided consistent seasonal numbers of feces. N o effort was made to collect feces of either long-tailed snakes or Chilean racerunners. Great horned owls ate primarily mammals (63- 100% of prey occurrences throughout the study period; Table 4). Insects and arachnids occurred at widely fluctuating levels in the diet (combined figures ranged between 0-36% by numbers); much of the invertebrate consumption may be due to juvenile owls (Jaksic, pers. obs.), which appear to be indiscriminate predators as they develop their hunting skills. Birds were consistently consumed at low levels, whereas reptiles and amphibians were only rarely taken. Barn owls preyed mainly on small mammals (56100n/~of their prey throughout the study; Table 4). Table 3. Mean number/season (n = 6 months) of vertebrate predator individuals sighted at Fray Jorge National Park, north-central Chile, throughout 11 biological seasons (wintering and breeding). Area surveyed was 750 ha. ns =not seen but heard, nh = neither seen or heard. Species W89 B89 W90 B90 W91 B91 W92 B92 W93 B93 W94 Mean(n=ll) Falconifoms (hawks) Buteo polyosoma Falco femoralis* Falco peregrinus* Falco sparverius Geranoaetus melanoleucus Milaago chimango Parabuteo unicintus Strigifoms (owls) Speotyto cunicularia Bubo airginianus Glaucidium nanum Tvto alba Carnivores (foxes) Pseudalopex culpaeus Total (No.1750 ha) *Not considered in Fig. 2 because it is a bird-eater. Insects and arachnids fluctuated widely in the diet (combined figures 0-40%). Birds were consistently taken at low numbers, but not reptiles or amphibians. Burrowing owls ate primarily insects and arachnids (combined figures 51-96% throughout the study period; Table 4), but also mammals (2-49%). The biomass contribution of mammals, however, was much higher than that of invertebrates (Silva et al. 1995). Birds and reptiles were rarely preyed upon, whereas amphibians were more prevalent in their diet. These amphibians were all terrestrial toads (Bufo chilensis), and were only partly consumed, as burrowing owls discarded the intestines and the apparently distasteful integument. Pygmy owls generally ate more mammals (13-100% of prey occurrences throughout the study; Table 4) and fewer insects (0-76%) than did burrowing owls. The former also ate more non-mammalian vertebrates, specifically birds. Neither amphibians nor reptiles (except for a single snake) were found in the diet of pygmy owls. Culpeo foxes ate primarily mammals, but with large fluctuations (12-98% of prey occurrences throughout the study; Table 4). Insects and arachnids together comprised from 0 to 79% of prey occurrences, in keeping with reverse trends of mammal consumption. Birds and reptiles were consistently preyed upon (though not amphibians), but their combined occurrences never exceeded 19%. Fruit seeds were frequent in the feces only during the breeding season of 1990 and the wintering season of 1991. Castro et al. (1994) reported that Culpeo foxes in Fray Jorge ate fruits only when mammal abundances were very low, which was clearly the case during those two seasons (Fig. 1, Table 4). Predator selectivity Great horned owls significantly selected 0. degus (i.e., took it in higher proportion than expected) and underconsumed both A. longipilis and A. olivaceus (Table 5). A selectivity shift away from P. darwini became evident after the breeding season of 1991. Barn owls selected P. darwini with a marginally non-significant probability ( P = 0.066) and significantly underconsumed A. longipilis, A. olivaceus, and 0. degus (Table 5). They apparently selected A. bennetti early in the study but generally ignored it after the breeding season of 1991. Burrowing owls significantly selected P. darwini and underconsumed 0. degus throughout the study (Table 5). Both A. longipilis and A. oliuaceus appeared to be selected prior to the breeding season of 1991 and were consistently ignored afterwards. Pygmy owls showed no significant preference for any small mammal, but they significantly underconsumed both T. elegans and 0. degus (Table 5). It should be noted that in the case of this owl, the time series was shorter than for all other owls (six vs eleven seasons, respectively). This renders it more difficult to detect significant departures. Likely, pygmy owls did select P. darwini and underconsumed A. bennetti. Culpeo foxes significantly selected A. bennetti (Table 5), whereas they underconsumed A. longipilis, A. olivaceus, and T. elegans throughout the study (note that they also underconsumed P . darwini at a marginally non-significant level; P = 0.066). Culpeo foxes appeared to select 0. degus after the breeding season of 1991, when small mammals irrupted in association with the E N S 0 disturbance. Fig. 2. Numerical response of mammal-eating vertebrate predators to small mammal densities at Fray Jorge National Park, north-central Chile, throughout 11 successive biological seasons (wintering and breeding). 0 20 40 60 80 100 120 140 Mammal density (Noha) Predator guild structure Before the E N S 0 disturbance (up to the wintering season of 1991), the following trophic guild structure was observed. Great horned and barn owls formed a tight mammal-eating guild during the first five seasons of the study (Fig. 3). These two owls clustered together because of their high and shared consumption of P. darwini and A. bennetti. The burrowing owl clustered with the Culpeo fox alone during one season, with both the Culpeo fox and pygmy owl during one, with all three predators during another, and kept separate during the remaining two seasons. Shared insectivory (in addition to carnivory) among the burrowing owl, Culpeo fox, and pygmy owl accounted for their frequent association in an omnivorous guild. The breeding season of 1991, with the onset of the mammal irruption brought together again the burrowing owl, Culpeo fox, and pygmy owl into a close omnivorous guild, and once again, great horned and barn owls remained in a tight mammal-eating guild which concentrated on P. darwini and A. bennetti as Prey. From the wintering season of 1992 on, an interesting reshuffling of guild membership occurred (Fig. 3). The great horned owl clustered with the Culpeo fox during all of the five seasons considered, owing to their high and shared consumption of A. bennetti and P. darwini. The barn owl continued preying substantially on these same two prey species during the wintering season of 1992 and thus joined both the great horned owl and the Culpeo fox. Staring in the breeding season of 1992, however, the barn owl shifted from preying on A. bennetti toward A. oiivaceus, and disengaged from those two predators. Because the burrowing and pygmy owls, in addition to insects, also preyed on P. darwini and A. olivaceus, they now clustered more closely with the barn owl. Only during the last season of the study (wintering season of 1994) did the barn owl converge again toward the great horned owl and Culpeo fox, because of its increased consumption of A. bennetti. The levels of similarity at which guilds formed did not vary pre- or post-ENSO. The 16 guild nodes observed up to the wintering season of 1991 did not significantly differ from the 23 nodes observed later on (Mann-Whitney U test, Normal approximation with continuity correction, i = 0.071, P = 0.943). ENS0 effects on trophic level productivity The following aspects of trophic level productivity were observed before, during, and after the 1991-92 E N S 0 disturbance in Fray Jorge (Table 6). The two preE N S 0 years (1989 and 1990) had average or low rainfall; during 1991 and 1992 rainfall was almost three times the average (85 mm); the two post-ENS0 years were similar to 1989 and 1990, respectively. Primary productivity (Gutierrez et al. unpubl.) showed the following associations with the rainfall pattern just described (Table 6): The number of perennial species increased only slightly during and after the E N S 0 disturbance, and their percentage ground cover remained almost constant throughout six years. In contrast, the number of ephemeral species doubled during the E N S 0 years, and their cover peaked dramatically during 1991, thereafter decreasing to figures similar to pre-ENS0 years. The seed bank of perennials peaked in 1991 at about four times the pre-ENS0 densities, and thereafter has remained relatively constant at 14 000 seeds/m2, still about three times higher than during pre-ENS0 years. The seed bank of ephemerals peaked in 1992 but only to 50% of previous levels, and then quickly decreased to seed densities even lower than pre-ENS0 years. Secondary productivity, at least that of mammalian seed and herbage eaters (Table 6) in general showed a 1-yr delayed response to the above pulses in primary productivity. Densities of small mammals peaked in 1992 and 1993 at about 6 times those recorded in the previous three years, and then quickly declined to a third that peak in 1994. Table 4. Diets of predators at Fray Jorge National Park, north-central Chile, by percent of total prey found in pellets or feces throughout 11 biological seasons (wintering and breeding). np = no pellets were found. Species Great homed owl Mammals Other vertebrates Invertebrates Number of prey Number of pellets Barn owl Mammals Other vertebrates Invertebrates Number of prey Number of pellets Burrowing owl Mammals Other vertebrates Invertebrates Number of prey Number of pellets Pygmy owl Mammals Othervertebrates Invertebrates Number of prey Number of pellets Culpeo fox Mammals Other vertebrates Invertebrates Number of prey Numberoffeces Feces with fruits W89 B89 W90 B90 np np np 0 np np np np 0 np np np np 0 np np np np 0 83.6 7.6 8.8 171 209 0.0 12.4 8.6 79.0 614 145 0.0 56.4 9.5 34.1 305 235 0.0 54.0 11.2 34.7 659 293 15.0 np W91 13.2 10.5 76.3 38 8 B91 W92 B92 W93 B93 W94 Total 100 0 0 15 nP nP nP "P 0 79 6 53 138 79 93.6 3.7 2.8 109 96 0.0 2152 405 1453 4010 2180 81 28.6 7.1 64.3 14 5 58.3 0.0 41.7 24 14 66.7 0.0 33.3 9 6 92.1 2.6 5.3 38 35 43.6 33.2 10.7 18.6 45.7 48.2 692 585 378 178 64.4 1.8 87.4 5.5 7.1 183 126 0.0 90.6 8.3 1.1 372 258 0.0 97.9 2.1 0.0 146 119 0.0 Tertiary productivity, that of vertebrate predators, also showed a somewhat delayed response to mammal prey densities (Table 6), declining from 11 and 121750 ha during pre-ENS0 years to only 7 during 1991, when small mammals began increasing. When the mammal irruption was in full swing (1992 and 1993), predator densities reached a peak of 171750 ha during both years, and decreased slightly as mammal populations crashed the following year. Note that we could not assess the numbers of nocturnal owls at the site, but they likely followed the same pattern. Discussion Changes in mammalian prey abundance With overall densities ranging from a low of 3 individuals/ha during the wintering season of 1991 to a high of 134 during the wintering season of 1993, there were also marked changes in the composition of the mammalian assemblage (see also Meserve et al. 1995). Before the E N S 0 disturbance (up to the wintering season of 1991), 0. degus was proportionally stable, P. darwini was increasing, and A. olivaceus was decreasing. Afterwards, A. olivaceus irrupted markedly but briefly, whereas 0. 11 96.0 2.9 1.1 174 143 0.0 degus picked up slowly but steadily reaching 71% of all captures by the end of the study. P. darwini hovered at about 20%, except at the very end when it reached a low of 4% of all captures. This inconstancy in relative abundance of mammalian prey thus provides a crucial cue for changing the predators' search image, and hence facilitate prey switching (Murdoch 1969, Murdoch and Oaten 1975). Over the entire span of the study and from the viewpoint of predators, 0. degus was increasing in the field, P. darwini was a relatively predictable prey, and A. olivaceus was an irruptive species. Numerical response of predators Ten predator species were year-round residents at the study site (five hawks, four owls, and one fox). They responded numerically to mammalian prey levels, but we could not determine if they did so by immigrating into the area after the E N S 0 disturbance, by reproducing more successfully in situ, or both. The burrowing owl was the predator that most strongly responded to mammalian increases, which indicates that although insects numerically dominate its diet, mammals are its staple prey (Silva et al. 1995). Wintering 0 20 40 60 80 100 100 Diet similarity (%) Wintering Diet similarity (%) Breeding 1989 80 60 40 20 0 Diet similarity (%) 1992 Breeding Diet similarity (%) Fig. 3. Trophic guild structure of predators at Fray Jorge National Park, north-central Chile, throughout 11 biological seasons (wintering and breeding). Diet similarities close to 100% indicate high overlaps in diet; those close to 0% indicate distinct diets. Trophic guilds recognized are those with > 50% diet similarity (species encased in boxes). Species abbreviations are: Bv, Bubo virginianus; Ta, Tyto alba; PC, Pseudalopex culpaeus; Gn, Glaucidium nanum; and Sc, Speotyto cunicularia. Table 6. Association between rainfall and productivity features at Fray Jorge National Park, north-central Chile, throughout six years. Note the heavy rains associated with El Niiio disturbance of 1991-92. Number and cover of plant species were measured in September each year, whereas seed densities were measured in August (n = 4 in all cases; detailed data in Gutierrez et al. unpubl.). Densities of small mammals and of predators are averages of wintering and breeding seasons reported in Tables 2 and 3, respectively. ---- - - Characteristics Rainfall (mm) No. perennial plant species % perennial cover No. ephemeral plant species % ephemeral cover No. perennial seeds/m2 No. ephemeral seeds/m2 No. small mammals/ha No. diurnal predators/750 ha Functional response of predators It appears that all four owls cued in on P. darwini as a staple prey, but the great horned owl shifted away from it and more toward 0. degus as the latter species became abundant (Table 7). The Culpeo fox underconsumed P. darwini, appeared to key on A. bennetti (which no owl did), and shifted toward 0. degus as it became abundant. This large-sized rodent (second only to A. bennetti) was underconsumed by the three smaller owls, likely because it fell out of their killing and handling capabilities (cf. Meserve et al. 1987, Jaksic et al. 1992). The rodents A. longipilis, A. olivaceus, 0 . longicaudatus, and the marsupial T. elegans were generally underconsumed by all five predators. Only the barn owl appeared to select the marsupial. Two features of predator selectivity require further discussion (Table 7). Octodon degus is labelled as a diurnal/crepuscular rodent, thus rendering it puzzling that the evidently nocturnal great horned owl preys on it. However, the only study to date that has quantitatively timed captures of central Chilean small mammals (Iriarte et al. 1989a), reported that 0 . degus has an activity period that extends before dawn, and beyond dusk even to dark night. Therefore, this species is available to strictly nocturnal hunters. The other interesting aspect is that the two largest predators in the area - great horned owl and Culpeo fox - are the only ones that apparently select 0 . degus and A. bennetti, the two largest-sized mammal prey at the study site. Either size-selectivity on the part of the large predators, inability of the smaller predators to kill and handle those prey species, or both, may explain the pattern here reported (cf. Iriarte et al. 1989b, Jaksic 1989b, Jaksic et al. 1992). Guild response of predators Three possible guild responses to the E N S 0 disturbance may have been predicted: changes in guild size (number of species per guild), structure (similarity levels at which guilds form), and membership (species composition of each guild). Throughout the study, two guilds were recognized, a mammalivorous one with generally two species, and an omnivorous one with generally three species. The absence of the pygmy owl during the first four seasons, and during the last season of the study, was the only factor contributing to changes in guild size. We suspect, however, that the absence of this owl was more an artifact of not finding its pellets and feeding roosts than of its not residing at the study site. The statistical distribution of guild nodes did not differ significantly between the first five seasons (Pre-ENSO) and the six latter seasons. Thus, no clear differences were observed in guild size or structure before or after the E N S 0 disturbance. Instead, changes in membership were conspicuous in both the mammalivorous and omnivorous guilds. The mammal-eating guild formed by the great horned and barn owl persisted until the breeding season of 1991, and was characterized by a high, shared consumption of P. darwini and A. bennetti. Thereafter, the great horned owl and Culpeo fox formed a mammalivorous guild that was characterized by the high, shared consumption of A. bennetti. Except for two seasons, the barn owl became more closely associated with their omnivorous relatives, burrowing and pygmy owl, due to its continued predation on P. danvini and a shift from A. bennetti toward A. olivaceus. Predation impact on small mammals Most predation on local small mammals seemed to be exerted by the four owls and one fox whose diets were studied here. Large hawks are important mammal predators elsewhere (e.g., Jaksic et al. 1992), but were mostly seen traversing but not hunting in the study site. Reptilian predators were scarce: Chilean racerunners, in addition to being mainly insectivorous, have a very limited activity season because of their being heliothermic (Castro et al. 1991). This also applies to long-tailed Table 7. Summary of apparent prey preferences for predators at Fray Jorge National Park, north-central Chile. Symbols: + = taken more frequently than expected from trapping data; - = taken less frequently; 0 =taken about as expected; +/- and - I + = apparent preferences shift over time. Activity periods are: D = Diurnal; C = Crepuscular; N = Nocturnal; DCN = Continuously active. Mammal prey A. bennetti A. longipilis A. olivaceus 0. degus 0. longicaudatus P. darwini T. elegans Weight (g)* X+SD (n) 201.2f46.1 54.3 8.8 32.3 5.3 140.9 20.9 24.4 3.0 58.2 13.7 22.6 9.5 (12) (38) (70) (46) (14) (124) (5) + + + + + Activity N N DCN DC N N N B. virginianus 1227 197 6 (N) + 0 - - T. alba + 307 22 (8) CN 0 G. nanum 81 13 (3) DCN 0 - - - - - +/- ++ - + - + - S. cunicularia 247 22 (3) CN 0 + 0 0 - 0 0 +0 + - P. culpaeus + 6520 3019 (5) CN + - -/+ - - * Weights for small mammals from Meserve et al. (1987); owl weights from Jaksic et al. (1992); fox weight from Meserve et al. (unpubl.). snakes, which may need no more than thirty-two 30-g rodents to survive and reproduce successfully over a whole year (Bozinovic and Rosenmann 1988). Based on their food habits and selectivities, it may be predicted that the experimental exclusion of owls and foxes from the study site should result in population increases of chiefly 0. degus, P. darwini, and A. bennetti. Indeed, Meserve et al. (1993b, 1996) demonstrated strong effects of predation on the survivorship and demography of 0. degus and Lagos et al. (1995) documented marked effects on microhabitat use by this rodent between predator open and predator exclusion grids. Meserve et al. (1996) found that P. darwini showed a significant response to the absence of predators, with larger population size and higher survivorship on predator exclusion grids. Unfortunately, A. bennetti has been characterized by low levels of trappability at the study site, such that statistically significant effects of predator exclusion have been difficult to detect (Meserve et al. 1996). On the other hand, we would not expect any major effect of predator exclusions on the remaining mammal species, which were generally taken in proportion or less than expected relative to their field abundances. This was the case for A. olivaceus and 0. longicaudatus, two other species for which Meserve et al. (1996) found no significant predator exclusion responses except for some minor survival differences. sion that predation effects proceed from top to bottom (e.g., Estes and Palmisano 1974, Castilla and Duran 1985, Kitching 1986, Moreno et al. 1986). For the terrestrial ecosystem here analyzed, it appears that E N S 0 effects do not cascade down from higher to lower trophic levels, but the opposite. In this bottom-up scenario, we predict that as primary productivity varies with rainfall, so should secondary (i.e., mammal prey densities), and tertiary productivity (i.e., vertebrate predator densities). Over 5.5 yr we were able to follow both an increase and decrease phase in Fray Jorge, but a new E N S 0 cycle is needed to determine whether the system behaves as postulated. Long-term monitoring of this terrestrial ecosystem should provide the acid test for this prediction. - This study was approved as part of the Program of Sponsored Research in the System of Protected Wildlife Areas operated by the Corporation Nacional Forestal (CONAF) of Chile. Waldo Canto and Juan Cerda kindly allowed us to work in the Park and provided logistic support. This research was supported by grants from the National Science Foundation (BSR 88-06639, BSR 90-20047 and DEB 93-18565), and the Fondo Nacional de Investigacion Cientifica y Tecnologica (FONDECYT 90-0930 and 193-1150). Acknowledgements References Annesto, J. J., Vidiella, P. E. and Gutierrez, J. R. 1993. Plant communities of the fog-free coastal desert of Chile: plant strategies in a fluctuating environment. - Rev. Chil. Hist. Nat. g6: 271-282. Barber. R. T. and Chavez. F. P. 1983. Biological - conseauences of El Niiio. - Science 222: 1203- 1210. Bozinovic, F. and Rosenmann, M. 1988. Energetics and food requirements of the female snake Phillodryas chamissonis during the breeding season. - Oecologia 75: 282-284. Castilla, J. C. and Duran, L. R. 1985. Human exclusion from the rocky intertidal zone of central Chile: the effects on Concholepas concholepas (Gastropoda). - Oikos 45: 391 399. Castro, S. A., Jimenez, J. E. and Jaksic, F, M. 1991. Diet of es in north-central Chile. the racerunner ~ a l l i ~ i s tpalluma J. Herp. 25: 127-129. - , Silva, S. I., Meserve, P. L., Gutierrez, J. R., Contreras, L. C. and Jaksic, F. M. 1994. Frugivoria y dispersion de - ENS0 effects: a bottom-up scenario Since the influential papers of Paine (1966) and Harper (1969), most students of predation and herbivory have focused on the effects of consumers on their prey diversity, be that of animals, plants or algae (see examples in Fuentes and Jaksic 1988). Demonstrations that some top predators or herbivores act as keystone species, whose removal or addition causes dramatic changes in prey species composition and abundance (cascading effects) have conveyed the general impres- semillas de pimiento (Schinus molle) por el zorro culpeo (Pseudalopex culpaeus) en el Parque Nacional Fray Jorge (IV Region, Chile). - Rev. Chil. Hist. Nat. 67: 169-176. Diaz, H. F. and Markgraf, V. (eds) 1992. El Nifio: historical and paleoclimatic aspects of the Southern Oscillation. Cambridge Univ. Press, New York. Dillon, M. 0 . and Rundel, P. W. 1990. The botanical response of the Atacama and Peruvian Desert floras to the 1982-83 El Nifio event. - In: Glynn, P. W. (ed.), Global ecological consequences of the 1982-83 El Nifio event. Elsevier, Amsterdam, pp. 487-504. Estes, J. A. and Palmisano, J. F. 1974. Sea otters: their role in structuring near shore communities. - Science 185: 10581060. Fuentes, E. R. and Campusano, C. 1985. Pest outbreaks and rainfall in the semi-arid region of Chile. - J. Arid Environ. 8: 67-72. - and Jaksic, F. M. 1988. The hump-backed species diversity curve: why has it not been found among land animals? Oikos 53: 139-143. Fulk, G. W. 1975. Population ecology of rodents in the semiarid shrublands of Chile. - Occas. Pap. Mus. Texas Tech Univ. 33: 1-40. - 1976. Owl predation and rodent mortality: a case study. Mammalia 40: 423-427. Greene, H. W. and Jaksic, F. M. 1983. Food-niche relationships among sympatric predators: effects of level of prey identification. Oikos 40: 151- 154. - and Jaksic, F. M. 1992. The feeding behavior and natural history of two Chilean snakes, Philodryas chamissonis and Tachymenis chilensis (Colubridae). Rev. Chil. Hist. Nat. 65: 485-493. Guerra, C. G., Fitzpatrick, L. C., Aguilar, R. E. and Venables, B. J. 1988. Reproductive consequences of El Nifio Southern Oscillation in Gray Gulls (Larus modestus). - Col. Waterbirds 11: 170-175. Gutierrez, J. R., Meserve, P. L., Contreras, L. C., Vasquez, H. and Jaksic, F. M. 1993a. Spatial distribution of soil nutrients and ephemeral plants underneath and outside the canopy of Porlieria chilensis shrubs (Zygophyllaceae) in arid coastal Chile. Oecologia 95: 347-352. - , Meserve, P. L., Jaksic, F. M., Contreras, L. C., Herrera, S. and Vasquez, H. 1993b. Structure and dynamics of vegetation in a Chilean arid thorn scrub community. Acta Oecol. 14: 271-285. Harper, J. L. 1969. The role of predation in vegetational diversity. - Brookhaven Symp. Biol. 22: 48-68. Hershkovitz, P. 1962. Evolution of Neotropical cricetine rodents (Muridae), with special reference to the Phyllotine group. - Fieldiana Zool. 46: 1-524. Holling, C. S. 1959. The components of predation as revealed by a study of small-mammal predation on the European pine sawfly. - Can. Entomol. 91: 293-320. Iriarte, J. A,, Contreras, L. C. and Jaksic, F. M. 1989a. A long-term study of a small-mammal assemblage in the central Chilean matorral. - J. Mammal. 70: 79-87. - , Jimenez, J. E., Contreras, L. C. and Jaksic, F. M. 1989b. Small-mammal availability and consumption by the fox, Dusicyon culpaeus, in central Chilean scrublands. J. Mammal. 70: 641-645. Jaksic, F. M. 1989a. Opportunism vs selectivity among carnivorous predators that eat mammalian prey: a statistical test of hypotheses. - Oikos 56: 427-430. - 1989b. What do carnivorous predators cue in on: size or abundance of mammalian prey? A crucial test in California, Chile, and Spain. - Rev. Chil. Hist. Nat. 62: 237-249. - and Medel, R. G. 1990. Objective recognition of guilds: testing for statistically significant species clusters. - Oecologia 82: 87-92. - , JimCnez, J. E., Castro, S. A. and Feinsinger, P. 1992. Numerical and functional response of predators to a longterm decline in mammalian prey at a semi-arid Neotropical site. - Oecologia 89: 90-101. - - - - , Meserve, P. L., Gutierrez, J. R. and Tabilo, E. 1993a. The components of predation on small mammals in semiarid Chile: vreliminarv results. - Rev. Chil. Hist. Nat. 66: 305-32'1. - , Feinsinger, P, and Jimenez, J. E. 1993b. A long-term study on the dynamics of guild structure among predatory Oikos 67: vertebrates at a semi-arid neotropical site. 87-96. - , Feinsinger, P. and Jimenez, J. E. 1996. Ecological redundancy and long-term dynamics of vertebrate predators in semi-arid Chile. - Conserv. Biol. 10: 252-262. Jimenez, J. E. 1993. Notes on the diet of the Aplomado Falcon (Falco fernoralis) in northcentral Chile. J. Ravtor Res. 27: 16i-163. ' - , Feinsinger, P. and Jaksic, F. M. 1992. Spatiotemporal patterns of an irruption and decline of small mammals in northcentral Chile. - J. Mammal. 73: 356-364. Kitching, R. L. 1986. Prey-predator interactions. - In: Kikkawa, J. and Anderson, D. J (eds), Community ecology: pattern and process. Blackwell, Carlton, Victoria, pp. 214239. Lagos, V. O., Contreras, L. C., Meserve, P. L., Gutierrez, J. R. and Jaksic, F. M. 1995. Effects of predation risk on space use by small mammals: a field experiment with a Neotropical rodent. Oikos 74: 259-264. Le Boulenge, E. 1985. Computer package for the analysis of capture-recapture data. - Acta Zool. Fenn. 173: 69-72. Marti, C. D. 1987. Raptor food habits studies. In: Giron Pendleton, B. A,, Millsap, B. A,, Cline, K. W. and Bird, D. M. (eds), Raptor management techniques manual. Natl. Wildl. Fed., Washington, DC, pp. 67-80. - , Korpimaki, E. and Jaksic, F. M. 1993. Trophic structure of raptor communities: a three-continent comparison and synthesis. - In: Power, D. M. (ed.), Current Ornithology, Vol. 10, Plenum, New York, pp. 47-137. Mellado, J. 1982. Sobre la alimentacion de Callopistes maculatus (Reptilia: Teiidae). Dofiana Acta Vertebrata (Spain) 9: 372-373. Meserve, P. L. 1981a. Trophic relationships among small mammals in a Chilean semiarid thorn scrub community. J. Mammal. 62: 304-314. - 1981b. Resource partitioning in a Chilean semiarid small mammal community. - J. Anim. Ecol. 50: 745-757. - and Le Boulenge, E. 1987. Population dynamics and ecology of small mammals in the northern Chilean semiarid region. - Fieldiana Zool. New Ser. 39: 413-431. - , Shadrick, E. J. and Kelt, D. A. 1987. Diets and selectivity of two Chilean predators in the northern semi-arid zone. Rev. Chil. Hist. Nat. 60: 93-99. , Gutierrez, J. R., Contreras, L. C. and Jaksic, F. M. 1993a. Role of biotic interactions in a semiarid scrub community in northcentral Chile. - Rev. Chil. Hist. Nat. 66: 225-241. - , Gutierrez, J. R. and Jaksic, F. M. 1993b. Effects of vertebrate predation on a caviomorph rodent (Octodon degus), in a semiarid thorn scrub community in Chile. Oecologia 94: 153-158. - , Yunger, J. A,, Gutierrez, J. R., Contreras, L. C., Milstead, W. B., Lang, B. K., Cramer, K. L., Herrera, s., Lagos, V. 0 . : Silva, S. I., Tabilo, E. L., Torrealba, M. A. and Jaksic, F. M. 1995. Heterogeneous responses of small mammals to an El Nifio Southern Oscillation event in northcentral semiarid Chile and the importance of ecological scale. - J. Mammal. 76: 580-595. - , Gutierrez, J. R., Yunger, J. A., Contreras, L. C. and Jaksic, F. M. 1996. Role of biotic interactions in a small Ecology 77: mammal assemblage in semiarid Chile. 133-148. Moreno, C. A,, Lunecke, K. M. and Lepez, I. 1986. The response of an intertidal Concholepas concholepas (Gastropoda) population to protection from Man in southern Chile and the effects on benthic sessile assemblages. Oikos 46: 359-364. - - - - - - - - - Muiioz, C. and Pisano, E. 1947. Estudio de la vegetacibn y flora de 10s parques nacionales de Fray Jorge y Talinay. Agric. TCcnica (Chile) 7: 71 -173. Murdoch, W. W. 1969. Switching in general predators: experiments on predator specificity and stability of prey populations. - Ecol. Monogr. 39: 335-354. - and Oaten, A. 1975. Predation and population stability. Adv. Ecol. Res. 9: 2-132. Paine, R. T. 1966. Food web complexity and species diversity. - Am. Nat. 100: 65-75. Pearre, S. 1982. Estimating prey preference by predators: uses of various indices, and a proposal of another based on Chi-square. - Can. J. Fish. Aquat. Sci. 39: 914-923. Pearson, 0 . P. 1975. An outbreak of mice in the coastal desert of Peru. - Mammalia 39: 375-386. Pefaur, J. E., Yafiez, J. L. and Jaksic, F. M. 1979. Biological and environmental aspects of a mouse outbreak in the semi-arid region of Chile. - Mammalia 43: 313-322. Reise, D. 1973. Clave para la determinacibn de 10s craneos de marsupiales y roedores chilenos. - Gayana: Zool. (Chile) 27: 1-20. Siegel, S. and Castellan, N. J. 1988. Nonparametric statistics for the behavioral sciences. 2nd ed. - McGraw-Hill. New York. Silva, S. I., Lazo, I., Silva-Aranguiz, E., Jaksic, F. M., Meserve, P. L. and Gutierrez, J. R. 1995. Numerical and functional response of burrowing owls to long-term mammal fluctuations in Chile. - J. Raptor Res. 29: 250255. Sneath, P. H. A. and Sokal, R. R. 1973. Numerical taxonomy. - Freeman, San Francisco. Sokal, R. R. and Rohlf, F. J. 1981. Biometry: the principles and practice of statistics in biological research, 2nd ed. Freeman, San Francisco. Solomon, M. E. 1949. The natural control of animal populations. - J. Anim. Ecol. 18: 1-35. Villagran, C. 1993. Una interpretation climatica del registro palinologico del ultimo ciclo glacial-postglacial en Sudamerica. - Bull. Inst. Fr. Etud. Andines 22: 243-258. Yifiez, J. L., Nufiez, H. and Jaksic, F. M. 1982. Food habits and weight of Chimango Caracaras in central Chile. - Auk 99: 170-171. OIKOS 78:2 (1997)