A central hypothesis of biogeography is that consumer-prey interactions are more intense at lower latitudes, leading to increased defenses of prey. Because plants vary in many traits that might affect palatability to herbivores, however, studies of latitudinal variation in single plant traits such as secondary chemistry provide only circumstantial evidence to test this hypothesis. We directly compared the palatability of 10 salt marsh plants from seven northern (Rhode Island and Maine) and eight southern (Georgia and Florida) coastal salt marshes by flying fresh plant material back and forth and allowing 13 species of herbivores direct choices between northern and southern conspecific plants in laboratory assays. In 127 of 149 assays (85%), herbivores showed a significant or marginally significant preference for northern plants. In only one assay did herbivores prefer southern plants. These results occurred regardless of the geographic location of the assay, herbivore species, year, or season of plant collection, although there were hints that latitudinal differences became less pronounced for two plant species late in the growing season. Our results provide the most comprehensive evidence to date for a latitudinal gradient in plant palatability in any community. The proximate plant traits and the ultimate evolutionary factors responsible for this pattern remain to be determined.

Biogeographic theory predicts that intense consumer–prey interactions at low latitudes should select for increased defenses of prey relative to high latitudes. In salt marshes on the Atlantic coast of the United States, a community‐wide pattern exists in which 10 species of low‐latitude plants are less palatable to a diverse suite of herbivores than are high‐latitude conspecifics. Examination of proximate plant traits (toughness, palatability of polar and nonpolar extracts, nitrogen content) of high‐ and low‐latitude conspecifics of nine plant species suggested that all these proximate traits had the potential to contribute to latitudinal differences in palatability of some plant species. Southern plants were tougher than northern plants (five species), had less palatable polar extracts (four species), and had lower N content (six species). Experimental evidence linking traits to latitudinal differences in palatability was strongest for polar extracts and lacking for N content. For one plant species, none of the traits we studied correlated with latitudinal variation in palatability. Because palatability differences may change when moving from fresh plants to freeze‐dried plants to plant traits, studies of latitudinal variation in freeze‐dried plants or plant traits are likely to under‐ or overestimate latitudinal variation in palatability of fresh plants. This study has begun to identify the proximate plant traits responsible for latitudinal variation in plant palatability in Atlantic coast salt marshes, but the ultimate evolutionary factors responsible for variation in these traits remain to be determined.

Few studies have evaluated the relative importance of multiple plant traits to herbivore diet choice, especially with an experimental approach. Moreover, although circumstantial evidence points to plant toughness and silica content as important determinents of diet choice, few studies have experimentally demonstrated that these factors actually deter feeding by herbivores. We examined feeding preferences of a generalist salt‐marsh crab, Armases cinereum, for all the common angiosperms in its habitat. We took an experimental approach to evaluating the importance of toughness, secondary chemistry, silica, salt, and protein in determining feeding preferences. Consumption of plants by Armases in two experiments was correlated with decreasing toughness. Consumption was more equitable when plants were ground up and reconstituted in agar discs, with reduced differences in toughness. In four pairwise choice tests, Armases always preferred the softer plant of the pair, but in three of four cases exhibited no preference when the same plants were presented in reconstituted discs. Several plant extracts significantly stimulated or deterred feeding by Armases, but these effects were not consistent with and could not be used to predict overall preferences for fresh plants. Salt in artificial diets stimulated feeding by Armases. Silica had no effect on feeding in two experiments and stimulated feeding in a third. Armases was mildly stimulated to feed by high levels of protein in artificial diets, but these levels were above those found in plants, and the protein content of plants did not correlate with feeding preferences. Our results point towards plant toughness as most important in determining feeding choices of Armases. Since Armases is omnivorous, it may escape dietary constraints that affect feeding choices of pure herbivores. The dogma that silica defends plants against herbivory needs to be re‐examined with additional experimental studies.