but we were not able to confirm this. We also observed that there were 

 numerous large juveniles at the reef and that some species ( Ha emu ion 

 f lavolineatum , H. aurolineatum and Chromis cyaneus ) were represented 

 by very distinct size classes - juveniles and adults. At the time of 

 our study a large mixed school of Haemulon juveniles was consistently 

 present over the reef during daylight hours. One suspects that this 

 school might have consisted of different species at another time of 

 year. It is our hypothesis that the juveniles of the species of 

 Haemulon all have similar niches which they share at least partly by 

 using the space at different times of the year. At present we do not 

 have sufficient data to confirm or deny this hypothesis. 



CONCLUSIONS AND RECOMMENDATIONS 



The observations described here confirm the general hypothesis that the 

 fishes of small patch reefs are true communities of interacting 

 species and not merely random assemblages of forms that happen to 

 find suitable conditions in the reef. The existence of resident 

 individuals, particularly of those species that disperse daily and 

 return to the same shelters, in itself is evidence of at I'east short 

 term stability. The large numbers of species constituting this 

 community implies that there is a fine partitioning of the space 

 available and this is confirmed by our observations that each species 

 has a characteristic regimen of space utilization that can be 

 described in terms of activity cycling, feeding areas, shelter sites, 

 home range size, numerical abundance, and social patterns. High 

 diversity communities are generally not subject to drastic population 

 cycles partly because predators have alternate prey species to which 

 they can turn when the most readily available prey species becomes 

 scarce. We suggest space sharing mechanisms also impose a maximum 

 population limit on each species and that this prevents overpopulating 

 with resultant food depletion and abrupt population crashes. This 

 can be operative in shelter sites, feeding area, general home range 

 or territories. Territoriality is most crucial during the breeding 

 season but some species seem to be territorial the year around. 

 Recruitment takes place during a definite limited time of the year 

 and therefore newcomers must compete with larger, established members 

 of the same species not only for food but also for hiding space and 

 feeding area and eventually nesting sites. It is unlikely that a 

 small recruit will be able to displace an established larger individual 

 of the same species and therefore its success is dependent upon the 

 availability of vacant sites. In coral reef environments it appears 

 that shelter competition may be more important than food competition. 

 Since the number of shelter sites is fixed, the maximum population 

 levels are also fixed. Space sharing adaptations permit more species 

 to use the habitat but impose stricter limits on the numbers of any 

 one species. 



VI-268 



