reefs in the Central Pacific (Marshal Islands). In addition, the 

 latter authors categorized the algal feeding fish into four categories 

 as follows, depending on the nature of their foraging habits: (1) 

 consumers of unicellular algae; (2) grazers on algae; (3) browsers on 

 algae; (4) incidental algal feeders. Herbivorous fish were found to 

 dominate the reef, as in the Virgin Islands. Many of the fish described 

 from the Marshal Islands are represented by comparable types in the 

 Virgin Islands (Randall, 1968). 



Our studies indicate that reduced grazing pressure is apparent at about 

 30m from a reef and that it causes a corresponding increase in plant 

 material. Most of the reef fish probably do not swim 20 - 30m from the 

 reef, for they are afraid of other predatory fish. Localized patches 

 of coral heads serve as a source of protection for reef fish, which live 

 on the sea grass beds. In addition, migratory fish and possibly juvenile 

 stages of reef fish were observed grazing on algae on the flats. It 

 should be emphasized that the depths on transects 1 and 2 were reasonably 

 uniform, and that linear changes in plant materials (biomass and 

 diversity) were much greater than any recorded at different depths. 



The results of our cage experiments, although of only 19 days duration, 

 confirmed the importance of fish grazing on the plant communities. Thus, 

 we recorded an increased diversity and abundance of plant materials under 

 two of the cages. No change was noted for the 3 and 180m cages. The 

 silty substrate at 180m probably restricted the establishment of new 

 plants. The 3m cage was located in the disturbed barren zone near the 

 perimeter of the reef, and one would expect a longer time for regrowth 

 to occur. 



Randall (1965) has given a detailed account of fish grazing on sea 

 grasses in the West Indies (including observations in Lameshur Bay), 

 After a combination of cage and transplant experiments, he concluded that 

 the bands of bare sand separating coral reefs and beds of sea grass 

 ( Thalassia testudinum and Syringodium filiforme ) resulted from heavy 

 grazing by Parrot fish ( Scarus and Sparisoma ) and Surgeon fish 

 ( Acanthurus ) . The present studies confirm his findings. 



Our feeding experiments showed that many plants (particularly Halimeda 

 incrassata ) were rapidly eaten by herbivorous fish. In addition, the 

 feeding experiments also showed some correlations between the linear 

 distribution of algae on transect 2 and the feeding preferences of the 

 fish. For example, Halimeda incrassata occurred as small plants and in 

 limited quantities near the reef, and it was actively eaten in the feed- 

 ing experiments. Such information suggests that it may be selectively 

 grazed upon. In contrast, Udotea conglutinata was recorded as one of the 

 closest algae to the reef and it was not grazed upon like Halimeda . It 

 should be noted that although selective fish grazing may occur, all of 

 the species present in the transitional areas (2,4 - 30m) on transect 2 

 were exposed to heavy grazing. Thus, each species found in the transi- 

 tional area exhibited either an increase in size or abundance with an 



VI-128 



