Spatial distribution of Psammomys obesus (Rodentia, Gerbillinae) in relation to vegetation in the Negev desert of Israel by A. V. TCHABOVSKY1 and B. R. KRASNOV2 'Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences, Leninsky prosp., 33, Moscow 119071, Russia 2Ramon Science Center, Jacob Blaustein Institute for Desert Research, Ben-Gurion University of the Negev, P. O. Box 194, Mizpe Ramon 80600, Israel E-mail: krasnov@bgumail. bgu. ac. il Summary. - Psammomys obesus in the Negev desert constructs burrows in the beds of dry rivers (wadi) densely covered with large shrubs of Atriplex halimus and on the first fluvial ter- race that is covered with scattered and low shrubs of Anabasis articulata. We suggested that the distribution of P. obesus depends on the intensity and the asynchrony pattern of vegetation of these two chenopod species. To test this suggestion, we studied the spatio-temporal distribution of P. obesus in relation to vegetation cover across these two habitats. The percentage of vegetating shrubs of A. halimus increased steadily from 1999 to 2001, while shrubs of A. articulata were much more verdant in 1999 and 2001 than in 2000. In gene- ral, intensity of vegetation of A. articulata varied more among years than that of A. halimus. In 1999 and 2001, burrows used by P. obesus were distributed almost equally between habitats, whereas in 2000 there were more inhabited burrows in the "wadi" habitat, especially those of females that were confined in this habitat. Number of inhabited burrows on the terrace varied strongly between years, while the variation in the number of inhabited burrows in the wadi bed was much less pronounced. Wadi and terrace habitats are considered as a "key" and temporary habitats, respectively. KEY WORDS: Psammomys obesus, Atriplex halimus, Anabasis articulata, habitat selec- tion, vegetation cover. Mammalia, t. 66, n0 3, 2002: 361-368. 362 MAMMALIA INTRODUCTION The reasons why a species prefers a particular habitat (or a set of habitats) and avoids other habitats may vary. Most important, a species is usually adapted to the physical characteristics of its preferred habitat (e. g., Shenbrot et al. 1999). Another rea- son for preference of a specific habitat may be the consequence of specialization to a type of food resource found there. The fat sand rat, Psammomys obesus Cretzschmar, 1828, a diurnal large solitary gerbillid inhabiting deserts of North Africa and the Middle East, demonstrates distinct habitat selectivity that differs in different regions within its geographic range (Daly and Daly 1973; Shenbrot et al. 1999). For example, in the Negev Highlands it occurs mainly in dry riverbeds (wadis) among loess hills, whereas in the Arava valley it can only be found in regions of sands (Shenbrot et al. 1999). The reason for this pattern of habitat distribution is that the species is a strict folivore and needs large amounts of succulent green vegetation all the year round (Daly and Daly 1973; Kam and Degen 1989; Zaime and Gautier 1989). The areas that satisfy its foraging requirements in the Negev Highlands are patches of Atriplex hali- mus and Anabasis articulata on the flood plains of wadis among the loess hills while in the Arava valley it forages on dense stands of Haloxylon persicum that grow on sand. Thus, P. obesus has been considered as a locally-restricted habitat specialist (Shenbrot et al. 1999). Daly and Daly (1973, 1974, 1975) showed that patchy distribution of P. obesus in Algeria was determined by the patchy distribution of chenopods. Adult females established their burrows preferentially in areas with higher abundance of Suaeda mollis bushes than in areas with fewer bushes (Daly and Daly 1975). In Tuni- sia, density of P. obesus appeared to be highly correlated with the abundance of fresh green vegetation (chenopods) (Fichet-Calvet et al. 1999, 2000). Zaime and Pascal (1989) reported that the abundance of P. obesus in Morocco was significantly higher in locations with larger and more voluminous bushes than where bushes were small. This strong relationship between P. obesus and the food plants has been used to explain why P. obesus construct burrows under their food shrubs and avoid open habi- tats (Daly and Daly 1975; Harrison and Bates 1991). Nevertheless, in Syria, Rioux et al. (1990) found P. obesus in relatively dry habitats with poor vegetation. Similarly, in the Negev desert, we observed that P. obesus made burrows not only under large shrubs of A. halimus at the wadi bottoms but also on the first fluvial terrace (Tcha- bovsky et al., 2001a, 2001b). A. halimus is the main component of the wadi vegetation, attains large size (in terms of both height and crown diameter), and, thus, provides P. obesus with protection from avian predators. In contrast, the "terrace" habitat is covered mainly with scattered and low shrubs of A. articulata and can, thus, be consi- dered to be much less protected than the wadi beds. Indeed, solitary P. obesus perceive themselves much more vulnerable in this open habitat. They spend less time above- ground, rest less, are more vigilant and move faster on the terrace than under the dense cover in the wadi bed (Tchabovsky et al., 2001 a). We suggested that the pattern of the distribution of P. obesus between these two habitats (highly protected and poorly pro- tected) depends on the intensity of vegetation of two chenopod species, namely A. halimus in the "wadi bed" habitat and A. articulata in the "terrace" habitat, and is affected by the asynchrony in the vegetation of these shrubs. To test this suggestion, we studied the spatio-temporal distribution of P. obesus in relation to vegetation cover across two types of habitats, namely «open terrace» and «closed wadi bed» in the Negev Highlands, Israel. DISTRIBUTION OF PSAMMOMYS IN THE NEGEV 363 MATERIAL AND METHODS Study site The study was carried out on the northern rim of the Ramon erosion cirque, Negev Highlands, Israel (30°35'N, 34°45'E) between 1999 and 2001. The main land- scape of the area is a complex of hills with a deep loess layer (up to 1 m) and wide dry riverbeds covered with shrubs of Anabasis articulata, (Forssk. ) Moq. -Tan., Atriplex halimus, L., Artemisia sieberi, Besser, Salsola schweinfurthii, Solms-Laub. and Noaea mucronata, (Forssk. ) Aschers. et Schweinf. The region is characterized by hot dry sum- mers (mean daily air temperature in July is 34°C) and relatively cool winters (mean daily temperature in January is 12. 5°C). Data were collected during March and April when mean daily temperature ranged between 14°C and 29°C. The mean annual rainfall in the area is near 100 mm and occurs in winter. Annual precipitation varied between years and was 54, 62 and 74 mm in 1999, 2000 and 2001, respectively. The study site (800 m by 100 m) was established along the valley of the epheme- ral river bed (wadi) Nizzana and included both the bottom of the wadi and the first flu- vial terrace. The wadi bed (wadi bottom and wadi banks) was covered densely with A. halimus shrubs ranging in size between 2 and 8 m in diameter and up to 2 m in height. There were also a few shrubs of A. articulata, A. sieberi and Salsola spp. in this habitat. The first fluvial terrace bordering the rocky second terrace was covered with the scattered shrubs of A. articulata that were 0, 5-1. 0 m in diameter and up to 0. 6 m in height. In addition, there were also a few small shrubs of A. halimus, Artemisia spp. and Salsola spp. Rodent trapping and survey Sand rats were live-trapped using Havahart two-door cage traps (model 1025) bai- ted with fresh A. halimus leaves. Traps were placed near the entrances of active bur- rows and were checked continuously during the day, so that each trapped individual stayed in a trap no more than 15 min. Fat sand rats do not need free water, so the bait satisfied their water requirements. Each trapped animal was sexed, aged, weighed and marked individually by cutting its fur in a symmetric recognizable pattern. To estimate P. obesus distribution, we conducted continuous inspection routes across the study site one to three times a day. We mapped and numbered all burrows with open entrances within the study site, and whenever a sand rat was encountered, it was identified according to its mark and its location was recorded. Thus, we obtained a list of burrows used by the sand rats of different sex and age categories, i. e. adult males, adult females, subadult individuals and juvenile individuals. Animals with body mass less than 70 g were classified as juveniles; animals with body mass of 70-120 g were considered subadults; while animals with body mass 120-170 g were considered subadults if they were sexually immature or adults if they were in reproductive condi- tion. Animals heavier than 170 g were regarded as adults. We analyzed the distribution of the entire population of sand rats within the study site as well as that of adult females. The latter were the most sedentary animals among all sex and age classes. Adult males, subadults and juveniles ranged more widely than adult females and used some burrows as temporary shelters. This could potentially bias a pattern of habitat dis- tribution and, thus, we did not conduct separate analyzes of the distribution of these 364 MAMMALIA categories of animal. In total, we captured and marked 71 P. obesus, including 29 adult females, that were located at 104 burrows. Habitat description Each burrow was classified according to its location as either «wadi burrow », i. e. constructed on the bank or in the bottom of the wadi bed, or «terrace burrow», i.