ECOLOGY AND TAXONOMY OF Holimeda 283 



in Hillis-Colinvaux (1974). Goreau's (1973) palmata zone has a more 

 restricted interpretation than in the 1959 description of the reefs, where 

 it is appHed to the entire region between reef flat and buttresses, which 

 in turn are followed by a cervicornis zone. The palmata zone, as delimited 

 by Goreau and Goreau (1973) (their region 5), is poorly developed at 

 Glory Be. 



There is no need for concern that the obvious divisions of the Glory 

 Be reef do not match precisely with Goreau's generalized model 

 because every section of reef can be expected to be modified to local 

 circumstances of tide, current and coastal morphology. And yet there 

 is, perhaps, a question concerning the line at which the waves break; is 

 the pattern at Glory Be an aberration because of the sheltered character 

 of the bay? 



1 . Methods of census 



The zonal boundaries in this system are distinct and can easily be 

 identified both in the water and on the panorama (Fig. 89) made from 

 the top of the cliff. The lengths and widths of the principal features of 

 Fig. 89 were measured by identifying salient features such as individual 

 rocks from the top of the cliff and on copies of the panorama, then 

 taping the distances directly with a nylon line. Twenty-four lengths 

 measured in this way were sufficient to compute the surface area 

 occupied by each zone. Rough estimates for the mean width of each 

 zone can also be made from a copy of the panorama on which the 

 measured distances are superimposed, allowing the calculation of the 

 relative areas of each zone along a metre-wide transect from the cliff to 

 the outer channel. These estimates (Table XXXIII) are obviously 

 crude, but the errors are probably not important for extrapolation 

 purposes since reef-to-reef fluctuations in relative zone areas must be 

 very large. 



Estimates of densities of Halimeda and other algae were made by 

 quadrat sampling, sets of nearest-neighbour samples and subjective 

 estimates of percentage cover. A convenient quadrat for dense popula- 

 tions of Halimeda is a rectangle of plexiglass placed on top of the 

 Halimeda population, when it is easy to count the plants underneath it. 

 Nearest-neighbour samples were most convenient in sandy areas where 

 Halimeda populations were sparse and where the swell made it hard 

 for a diver to keep station. The diver chose one plant as it drifted into 

 view, placed one edge of a steel rule against it, and pivoted to find the 

 other plant of the pair. Estimates of percentage cover were found to 

 be the only practicable way of assessing densities of the sprawling 



