Host-tree selection by Florida tree snails, Liguus fasciatus (Muller, 1774), in Big Cypress National Preserve, Florida, USA

It has frequency been suggested that the Florida tree snail, Liguus fasciat us, exhibit preferences for specific host trees, although most accounts of host-tree selection are based on anecdotal observations. We assessed the re lative use of host trees in eight hammocks in the Big Cypress National Pres erve for which the relative availability of each potential host-tree specie s was known. Based on a total of 1,464 tree snail observations on 21 specie s of host tree within our study plots, host trees were not used in proporti on to their availability when the expected values are derived either from t he number of trees or from basal area. Of the most common tree species on o ur study area, wild tamarind (Lysiloma latisiliqum) was consistently used i n excess of its availability. Wild lime (Zanthoxylum fagara), was not a com mon tree species, but was also used far in excess of its availability. Pige on plum (Coccoloba diversifolia) and gumbo-limbo (Bursera simaruba) were co nsistently used less than expected from their respective availabilities. Ba sed on Akaike's Information Criterion (AIC), the most parsimonious loglinea r model, after accounting for availability, was one that included only tree species effects. In contrast to the model selected using AIC, likelihood r atio tests indicated that there may also be differences among hammocks, but not in relation to a linear gradient of hurricane damage. We observed diff erences in length among 3-year-old snails on different host trees, and thes e differences were consistent with the hypothesis that host-tree selection reflects food availability. Snails found on host-tree species that were use d in greater than expected frequency had the greatest length, those found o n trees used less than expected frequency were the smallest, and those foun d in proportion to their availability were intermediate in length.