transplanted onto the reef slope for the herbivory assay (Fig. 

 8), young segments still contained 150% more organic content and 

 210% more nitrogen than old segments. Given the high food value 

 of young versus old segments, it is surprising that grazing was 

 only 39% higher on young segments (Fig. 8). We assume that the 

 increased abundance of the halimedatrials and 



halimedatetraacetate (Figs 6 and 7, Table 1) in younger segments 

 kept herbivory from more closely tracking the differences in 

 segment food value. Halimedatrial clearly deterred grazing by 

 both parrotfishes and surgeonf ishes on the reef at St. Croix 

 (Fig. 9, Table 2) and Targett et al . (1986) recently showed that 

 halimedatetraacetate deterred grazing by the parrotfish Sparisoma 

 radians in laboratory assays. Additionally, Paul and Van Alstyne 

 ( in press ) have shown that halimedatrial and halimedatetraacetate 

 both deter feeding by Pacific reef fishes and that halimedatrial 

 is a stronger deterrent than halimedatetraacetate. 



In contrast to our observations on fish feeding, Targett et 

 al. (1986) noted that Sparisoma radians avoided the more 

 chemically defended tips of Halimeda in favor of the less 

 chemically rich, but more heavily calcified, basal portions of 

 the plant. Additionally, Pacific reef fishes appear to graze new 

 unpigmented tips less than older pigmented ones (Paul and Van 

 Alstyne, 1987), suggesting that new tips are better defended 

 despite their lack of calcification. These observations suggest 

 that different fishes may respond very differently to various 

 combinations of chemical and morphological defenses. The 

 unusually high incidence of chemical defenses in tropical 

 seaweeds that are calcified (Hay, 1984b; Paul and Hay, 1986) 

 suggests that multiple defenses may be necessary for plants that 

 live in habitats with a diverse assemblage of herbivores, several 

 of which may be immune to any single defensive trait. 



If different herbivores are differentially affected by 

 halimedatrial versus halimedatetraacetate, then the production of 

 the halimedatrials primarily at night and of halimedatetraacetate 

 early in the day (Fig. 6) may correspond to the activity patterns 

 of different herbivores. This could be a particularly 

 interesting area for future study since different types of marine 

 herbivores have recently been shown to have very different 

 responses to terpenoids produced by other seaweeds (Hay et al., 

 1987a, b; Paul et al., 1987). The fact that halimedatrial 

 strongly deters fish grazing (Fig. 9; Paul and Van Alstyne, 

 1987) yet is highest in concentration at night when herbivorous 

 fishes are inactive is puzzling. This deserves further study and 

 may suggest that high nocturnal concentrations of halimedatrial 

 serve to defend the soft, new segments from night-active grazers 

 like sea urchins or from more inconspicuous nocturnal herbivores 

 such as amphipods, crabs, and small gastropods that are not 

 normally considered to be capable of grazing mature, calcified 

 segments. Additionally, preliminary investigations (V. Paul, 

 work in progress) suggest that biosynthesis of 



392 



