Summary. Few studies have investigated foraging decisions in collectively foraging social insects with no studies in termites. In termites predation is assumed to be a key mortality factor. Therefore, we experimentally investigated the role of predation pressure in foraging decisions of the fungus cultivating, mound building termite Macrotermes bellicosus in two habitats of the Comoé National Park (Ivory Coast). We used the indirect approach of measuring the Giving up Density (GUD), which is the amount of food left when individuals stop foraging in a food patch, whilst exper- imentally varying predation pressure. Three different condi- tions were examined: (a) natural predation, (b) no predation, and (c) experimental predation through artificial removal of termites. In the shrub savanna, foraging termites responded to increasing predation with increasing GUDs. By contrast, in the gallery forest, there was no gradual response. Instead termites abandoned a food patch immediately after an attack by predators. Without predation GUDs were lower in the savanna than in the gallery forest indicating that food had a higher value in the former habitat. This, together with the differential behavioral responses to predation, was in accordance with high availability of food in the gallery forest and a limited supply of food in the savanna. Thus, according to our results termites traded off predation pres- sure differently, according to the availability of food in both habitats. Key words: Giving up density, optimal foraging, depreda- tion, termites. Introduction Factors influencing foraging decisions have been studied in a wide range of solitarily foraging species since the formaliza- tion of optimal foraging theory (e. g., Stephens and Krebs, 1986; Krebs and Kacelnik, 1991). However, for social insects that forage collectively, and combine individual and cooper- ative tasks, fewer data are available (e.g., ants: Breed et al., 1987; Roces, 1990; Beckers et al., 1993; honeybees: Schmid- Hempel et al., 1985; Wolf and Schmid-Hempel, 1990; review: Detrain et al., 1999). Most studies concerned the influence of food quantity and quality on foraging decisions made by individual foragers within a foraging group of ants (e. g. the influence of food quality on the load size of an indi- vidual; Roces, 1990; Beckers et al., 1993). The combined effect of these individual decisions on the exploitation of food patches by a group of collectively foraging individuals is less well known (Detrain et al., 1999). What is the outcome of optimal foraging strategies of individuals on foraging pat- terns of foraging groups, and thus the exploitation of food by a colony? Among social insects, termites are truly collective for- agers. In comparison to ants and bees in which solitary work- ers often carry out food search and recruit nestmates based on an individual evaluation of resource quality, termites appear to forage only in groups (Traniello and Leuthold, 2000). Termites have never been used to test any predictions of models of optimal foraging theory (Traniello and Leuthold, 2000). As predation is often assumed to be a key mortality factor (Lepage, 1981; Abe and Darlington, 1985) in termites, we investigated the influence of predation on forag- ing decisions in collectively foraging termites. To evaluate the influence of predation on foraging deci- sions the indirect experimental approach of the Giving up Density (GUD; e.g., Brown, 1988; Thorson et al., 1998) was used. The GUD measures the amount of food left in a patch when foraging individuals abandon that patch. This method is based on the marginal value theorem of optimal foraging theory (Charnov, 1976; Stephens and Krebs, 1986; Krebs and Kacelnik, 1991). It assumes that a forager should stay in a Insectes soc. 49 (2002) 264– 269 0020-1812/02/030264-06 © Birkhäuser Verlag, Basel, 2002 Insectes Sociaux Research article Evaluation of predation risk in the collectively foraging termite Macrotermes bellicosus J. Korb * and K. E. Linsenmair Theodor Boveri Institut, Lehrstuhl für Tierökologie und Tropenbiologie, Am Hubland, D-97074 Würzburg, Germany, e-mail: Judith.Korb@biologie.uni-regensburg.de; ke_lins@biozentrum.uni-wuerzburg.de * Current address: Judith Korb, Biologie I, Universität Regensburg, D-93040 Regensburg, Germany, Fax: ++49 941 943 3304 Received 11 January 2002; revised 15 March 2002; accepted 10 April 2002.