of all obvious herbivores, none of these portions produced new 

 segments during the day and 9 of the 10 produced new segments 

 that night. Control portions that had been transplanted back 

 into the microcosm showed the same pattern; none produced 

 segments during the day and 10 of 10 produced new segments that 

 night. 



Growth of new segments 



The sequence of development shown in Figure 3 is based on 

 the average number of branch tips having each type of terminal 

 segment at each sampling period; it does not clearly show how 

 rapidly an average segment expands or how long the expansion of 

 newly formed segments continues. We examined the time course of 

 segment expansion by individually marking H^ incrassata branches 

 at the earliest visible stage of segment initiation and measuring 

 the width of the newly forming segments and taking notes on their 

 general appearance every 4 hours for the next 48-51 hrs. 

 Production of very small (<1 mm wide), unpigmented buds began 

 between 1300-1600 hrs (Fig. 4). Width of these newly forming 

 segments increased approximately linearly for 24-28 hr after 

 initiation and then leveled off. Between 0400 and 0600 hr the 

 new tips began to show pigmentation and by 0800 hr all tips were 

 pigmented and resembled mature segments except for their lack of 

 obvious calcification and their smaller size (2.5 mm wide versus 

 4 mm wide in older segments). 



Effects of light on the timing of segment production 



To test the effect of changing light regimes on segment 

 production, we performed the following experiments. After 8 hrs 

 of light, we cut ten 50 cm^ diameter cores from mats of Halimeda 

 opuntia growing in the Smithsonian microcosm and transplanted 

 these from the large microcosm into two separate refuge tanks (5 

 per tank) that were connected by water flow with the large 

 microcosm (see Fig. 2). In one of these tanks the plants were 

 subjected to continuous light; in the other tank they were 

 subjected to continuous darkness (all plants had previously been 

 acclimated to 16 hr light: 8 hr dark periods). During the next 24 

 hr, 4 of the 5 plants kept in constant light produced normal 

 segments at the same time that Halimeda populations in the large 

 microcosm (now dark) were producing segments; no plants placed in 

 constant darkness produced any new segments. Water flow to both 

 refuge tanks had been adjusted so that turnover time was about 8 

 minutes. Thus, water chemistry in both tanks should have been 

 similar since all water for both came directly from the large 

 microcosm. Thirty-six hours after the initial transplanting, 

 light cycles in both refuges were set on a reverse 16 hr light: 8 

 hr dark cycle so that the dark cycle in the refuges coincided 

 with the light cycle in the large microcosm. Thus, during the 

 dark cycle in the refuges, plants received water with daytime 

 chemical characteristics from the large microcosm. We allowed 



377 



