Hydrobiologia 397: 171–178, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands. 171 Relationships between trematode infection and habitat depth in a freshwater snail, Semisulcospira libertina (Gould) Kazuko Shinagawa 1 , Misako Urabe ∗ & Makoto Nagoshi Department of Biological Science, Faculty of Science, Nara Women’s University, Kitauoyanishi-machi, Nara 630- 8506, Japan) 1 Present address: Tomigaoka-machi 15-18, Tondabayashi, Osaka 584-0094, Japan Fax: [+81]-(0)742-20-3424. E-mail: urabe@cc.nara-wu.ac.jp Received 10 March 1998; in revised form 27 January; accepted 8 February 1999 Key words: parasitism, freshwater gastropod, trematodes, habitat selection, water depth, host size Abstract We found that in natural conditions, the freshwater snail Semisulcospira libertina was distributed in deeper sites when infected with trematode larvae than when uninfected, and examined whether water-depth selection was differ- ent between infected and uninfected snails in the laboratory. We compared the mean size of infected and uninfected snails, and the correlation between mean water depth and shell size in uninfected snails. Small uninfected snails (up to 11 mm in shell width) showed a correlation between shell width and water depth, while large snails (larger than 11 mm) did not, but used deeper sites on average than smaller snails. In the small-size class, snails infected with mature cercariae used significantly deeper sites than uninfected snails. On the other hand, in the large-size class, snails infected with immature cercariae used significantly shallower sites than uninfected snails. In each size class, snails infected with mature cercariae used deeper sites than snails infected with immature cercariae or parthenitae. Thus, the influence of trematode infection to host behavior varies according to the development stage of trematodes. In the field, snails infected with mature and immature cercariae had a tendency to occur in deeper sites than uninfected snails of both small and large size classes. These results suggest that some factor other than active selection of water depth by snails determined the distribution pattern of infected snails in natural conditions. Introduction Many effects of parasites on host ecology have been reported, based on field studies and laboratory exper- iments. In some cases of molluscan first-intermediate hosts of trematodes, infected snails shift their habit- ats. At least two factors can cause a distribution shift in infected snails: distribution of trematodes in stages infectious to snails (eggs or miracidia), and behavioral change of host snails caused by parasites. Jokela & Lively (1995), for example, surveyed the spatial pattern of infection by trematodes in Pot- amopyrgus antipodarum in a lake and showed that snails infected with Microphallus sp., whose final hosts were waterfowl, occured in shallower wa- ∗ Author for correspondence ter, while snails infected with Telogaster opistorchis, whose final hosts were eels, lived in deeper habitats. They explained the vertical distribution of infected P. antipodarum by the distribution of trematode eggs de- posited by the final host. Curtis (1990, 1993) showed that a trematode, Gynaecotyla adunca, changed the habitat selection of its marine host snail, Ilyanassa ob- soleta, during high tide. This facilitated transfer to the second-intermediate host, Talorchestia longicornis, an example of host-behavior manipulation by parasites, of great interest to evolutionary biology. However, the influence of parasites on host be- havior cannot be examined by simple comparison of infected and uninfected snails, because some other factors may mediate between infection and host beha- vior. In the case of bivalve hosts, the high prevalence of Macoma balthica on mud surfaces is a well-known