seaweeds allowed us to separately test the effects of 

 halimedatrial on parrotfishes versus surgeonf ishes. Halimeda- 

 trial was coated on the plants as 1% of their blotted wet mass 

 (our initial HPLC analysis of fresh, young tips of Halimeda had 

 indicated that halimedatrial comprised 2-6% of their blotted wet 

 mass). Since this compound is hydrophobic, it adheres to the 

 plant after the ether evaporates. The plant can then be placed 

 in seawater with minimal loss of the halimedatrial. McConnell et 

 al . (1982) developed this method for similar compounds and 

 demonstrated that 88-100% of these types of hydrophobic compounds 

 remain on coated plants after several hours in seawater. 

 Following our assays, we extracted the uneaten portions of our 

 treatment plants and confirmed by thin layer chromatography (TLC) 

 that halimedatrial was still present and had not degraded to some 

 other compound. We attempted similar tests with 

 halimedatetraacetate but at the end of our assays we could not 

 confirm its presence by TLC. 



Paired groupings of halimedatrial-treated and control plants 

 were placed on either shallow (2-3 m) back reef (assay for 

 surgeonf ishes ) or the deeper (15-18 m) reef slope (assay for 

 parrotfishes). For assays with either Thalassia or Acanthophora , 

 4 halimedatrial-coated plants or 4 control plants ( 6 cm long) 

 were placed in a single 3-strand rope. Ropes with treatment 

 plants and ropes with control plants were paired by placing them 

 within 0.5-1 m of each other on the reef. For Thalassia assays 

 we used 14 rope pairs; for Acanthophora assays we used 10 pairs. 

 During the first hour of each test, each of 4 or 5 divers watched 

 two pairs of ropes and recorded the number of separate fish 

 grazing from the ropes, the number of bites taken from each, and 

 whether the grazers were parrotfishes or surgeonf ishes. After 

 the first hour, all ropes were checked approximately hourly and a 

 pair was removed whenever half or more of the plant material on 

 either rope had been consumed. Some pairs were removed after 

 only 1 hr, some remained for 7 hrs. Consumption was measured as 

 the proportion (in cm) of each blade that was consumed (see Hay, 

 1984a). During our observations, Thalassia was fed on 

 exclusively by the red band parrotfish Sparisoma aurof renatum ; 

 Acanthophora was eaten exclusively by surgeonf ishes, primarily 

 Acanthurus bahianus and to a lesser extent by A^ coeruleus . 



The proximate effects of light cycles and diel changes in 

 water chemistry on the timing of segment production were 

 evaluated by altering the timing of day-night periods within 

 different portions of the Smithsonian's coral-reef microcosm (see 

 Fig. 2). By placing the refuge tanks on light-dark cycles that 

 were opposite those of the large microcosm and rapidly cycling 

 water from the microcosm through the refuge tanks, we were able 

 to place Halimeda plants in a nighttime environment characterized 

 by a daytime chemical regime. Water from the lighted microcosm 

 (7000 1) was supplied at a rate that filled the refuge tanks 

 (150 1) every 8 minutes; we assumed that the relatively small 



374 



