Summary. At the Lamto Savanna Ecological Station (Côte d’Ivoire), Odontotermes nr. pauperans (Termitidae, Macro- termitinae) was observed to build mounds enriched with fine particles. Using laboratory experiments we studied the selection of building materials by worker termites offered soil from two contrasting horizons: superficial soil (15 –20 cm) and a deeper layer (70–80 cm). The physical and chemical properties of the unused soil and subsequent termite constructions (foraging galleries and fungus-comb chambers) were compared in each case. When presented with a single soil type, the termites modified soil texture for different parts of their structure. Termite building activity increased when presented with both soil types and a notable selection was observed in the use of a given soil type for a specific part of the structure built. We conclude that termites utilise soil particles selectively, favouring finer particles and making constructions which match ecological, physiological, and behavioural needs. Compared with material from deeper horizons, less energy was expended when surface soil was used as a resource for gallery building and less C and N supplementation was needed. In contrast, termites preferred deeper soil for con- structing fungus-comb chamber walls because this material has greater water-holding capacity. Key words: Termites, Macrotermitinae, building, soil pre- ferences. Introduction Among the soil macrofauna, fungus-growing termites play a primary role in savanna ecosystem functioning. They signifi- cantly modify their environment by increasing the content of fine soil particles, thereby stimulating microbial activity and raising the amount of soil water available to plants (Lee and Wood, 1971; Abbadie and Lepage, 1989; Ouedraogo, 1997; Konaté et al., 1999; Holt and Lepage, 2000). A purely descriptive approach to termite effects on soil properties, however is not enough to understand why ter- mites have to modify the material for their own needs (Black and Okwakol, 1997). The Macrotermitinae subfamily, for example, is characterised by exosymbiosis with a fungus (Termitomyces). This relationship has well-defined tem- perature and humidity optima (Collins, 1977), and hence a complex nest structure is necessary within which conditions remain constant (Grassé, 1986) or are actively controlled via mound architecture (Korb and Linsenmair, 1998). The success of the symbiosis, together with the ability to retrieve water from moist horizons deep in the soil, allows many Macrotermitinae to dominate semi-arid and arid environ- ments (Lepage et al., 1974). In the Lamto Savanna ecosystem (Côte d’Ivoire), Odon- totermes nr. pauperans (Silvestri) cultivates its fungus in interconnected chambers, concentrated in space and cons- tituting a termitarium (Josens, 1972; Konaté, 1998). These termites also construct covered runways (sheetings or gal- leries) on the soil surface in order to collect plant litter (Bagine, 1984). The termitaria density varies from 8.1 to 12.2 ha –1 (Konaté, 1998), occupying up to 9% of the soil surface and representing a soil volume of 300 m 2 in some biotopes (Abbadie et al., 1992; Konaté et al., 1999). Termite workers bring and handle fine particles to the soil surface, which enriches the nest surroundings with fine particles and other incorporated substrates. One of the consequences is an increase in the amount of the water available to plants (Konaté et al., 1999). The impact of termites on soils is closely related to their construction activities and these, in turn, reflect their ecolo- gical needs. Most published work on soil modification, how- ever, is descriptive, consisting of comparisons of in situ con- structions and putative parent soils without an experimental component. In a new approach to study and understand the Insectes soc. 49 (2002) 1 – 7 0020-1812/02/010001-07 $ 1.50+0.20/0 © Birkhäuser Verlag, Basel, 2002 Insectes Sociaux Research article Termite soil preferences and particle selections: strategies related to ecological requirements P. Jouquet 1 , M. Lepage 1 and B.Velde 2 1 Laboratoire d’Ecologie, UMR 7625, Ecole Normale Supérieure, 46 rue d’Ulm, F-75230 Paris cedex 05, France, e-mail: jouquet@biologie.ens.fr 2 Laboratoire de Géologie, Ecole Normale Supérieure, 24 rue Lhomond, F-75230 Paris cedex 05, France Received 23 November 2000; revised 27 April 2001; accepted 21 August 2001.