Oecologia (1990) 83:512-518 Oecologia 9 Springer-Verlag 1990 Testing values of crested porcupine habitats by experimental food patches Joel S. Brown * and Philip U. Alkon Mitrani Center for Desert Ecology, Jacob BlausteinInstitute for Desert Research, Ben-GurionUniversityof the Negev, Sede Boqer Campus 84993, Israel Received July 25, 1989/ Acceptedin revisedform February 22, 1990 Summary. We established depletable, artificial food patches in three habitats used by Indian crested porcu- pines (Hystrix indica) in a desert biome, and measured the number of food items remaining (i.e., "giving up density"= GUD) following nightly foraging bouts. Por- cupines discriminated between resource types (peanuts vs. garbanzo beans), and exhibited clear habitat prefer- ences in the face of uniform resource availability in time and space. Lowest GUD's (= lowest foraging costs) were in the habitat of densest cover, and during dark (little or no moon) nights. The results indicated a high sensitivi- ty to predation risk. Crested porcupines behaved as ex- pected of optimal foragers, and appear to be excellent subjects for further field experiments using the GUD approach. Key words: Habitat selection - Foraging behavior - Pre- dation costs - Desert porcupines The value of a habitat for a forager is influenced by the available food resources it contains, by the effort required to exploit those resources, by the risk of preda- tion while foraging, and by the foraging opportunities available in other habitats. An optimal forager should weigh these disparate inputs to fitness (costs and benefits) in shaping its pattern of habitat utilization. From the investigator's standpoint, simply measuring the time al- located by a forager among habitats will not reveal the importance of specific costs and benefits. By experimen- tally controlling these inputs, however, it may be possible to separate their effects, and to delineate salient environ- mental influences on a forager's behavior (e.g., Wilson 1976; Werner et al. 1983; Lima et al. 1985; Pimm et al. 1985). * Present address and address for offprint requests: Department of Biological Sciences, Box 4348, University of Illinois, Chicago, IL 60680, USA One such experimental approach uses depletable arti- ficial food patches to determine the food density at which a forager ceases to harvest additional items (Brown 1988; also see Wilson 1976; Hodges and Wolf 1981). This "giv- ing up density" (GUD) may be a powerful index to habi- tat preferences and foraging costs. Among small grani- vorous rodents, GUD's in artificial seed trays have proved useful for: delineating mechanisms of species coexistence (Brown 1989), measuring costs of predation on foraging (Brown et al. 1988; Kotler et al. 1988), study- ing diet selection on patchy resources (Brown and Mit- chell 1989), and defining the animals' patch assessment abilities (Valone and Brown 1989). Here we report a successful application of the GUD method in field experiments on the foraging behavior of a large, nocturnal desert herbivore, the Indian crested porcupine in the Negev highlands of southern Israel. Our field experiments derived from previous observations on habitat selection by porcupines and on environmental influences on their temporal and spatial activity (e.g., Alkon and Saltz 1988; Saltz and Alkon 1989). The pres- ent study revealed that porcupines behaved in a manner consistent with fitness maximization, and that predation risk exerted a strong influence on porcupine foraging decisions. Study area and animals The study was made on a portion of the Zin plain, a large loessial plateau (560 m elevation) in the Sede Boqer region of the Negev Desert. Landscapes, climate, and vegetation of the region have been previously described (e.g., Saltz and Alkon 1989). We focused on three adja- cent habitats within a relatively small part of the plateau (Fig. 1): (1) the wadi edge, a zone of deep moderately disturbed loess, abutting a steep tributary wadi in which a porcupine den complex was located. The area supports patches of Erodium hirtum, tubers of which are an impor- tant porcupine food; (2) the "barrens', a flat zone heavily