574 



Fishery Bulletin 93(3), 1995 



a fish, I did not encounter that fish again during the 

 drift. To confirm the accuracy of my underwater FL 

 estimates, I compared markings on my slate (1-cm 

 increments) with structural relief that L. analis 

 rested on or passed by at reef 1. 1 repeated this prac- 

 tice until I could discern between the three size 

 classes ±5 cm from distances up to 5 m away. 



On each sampling date, I performed three repli- 

 cate drifts from reef 2 to reef 1. Size frequencies were 

 derived from the largest number of fish counted in 

 each of three size classes, irrespective of replicate, 

 because the number offish in each size class was at 

 least equal to the greatest number of that size class 

 seen during a single drift (Sale and Douglas, 1981). 



Results and discussion 



Four individual L. analis were recognized from cen- 

 sus to census for periods up to one year. Fish were 

 identified by body size, scars, fin anomalies, or shape 

 of the black upper-body spot. However, owing to the 

 wary nature of L. analis, it was difficult to remain 

 within close proximity of subjects to identify ad- 

 equately all fish by this method. Although individu- 

 als were considered resident, it is possible that oth- 

 ers were transient or wandering fish. Still, my ob- 

 servations are consistent with a previous study 

 (Beaumariage, 1969) that indicated little movement 

 in adult L. analis. 



On each sampling date, up to 56 L. analis were 

 counted and measured at the study site. Fish ranged 

 in size from 15 cm FL (half the length of my under- 

 water slate) to 65 cm FL (>twice the length of my 

 slate). Small (20-cm-FL), medium (30-cm-FL), and 

 large (>40-cm-FL) L. analis were aged 1+ yr, 2+ yr, 

 and 3+ yr, respectively (sensu Claro, 1981; Mason 

 and Manooch, 1985). The number of small and me- 

 dium-size fish observed varied from 6 to 24 and 12 

 to 25, respectively. The number of large fish never 

 exceeded 13 (Table 1). From April to July, fish sizes 

 were normally distributed; however, during August, 

 fish sizes became negatively skewed. On 23 Febru- 

 ary 1992, water clarity was extremely poor (<5 m), 

 which may account for the low number of small fish 

 recorded on this date (Table 1). 



The causes for seasonal variation in the size struc- 

 ture of L. analis are difficult to isolate from my ob- 

 servations; however, some inferences can be drawn 

 from previous studies. For example, the increase of 

 small fish in late August (Table 1) corresponds well 

 with the peak recruitment to seagrass beds reported 

 for juvenile L. analis (<7 cm FL) during August 

 (Springer and McErlean, 1962) and September 

 (Garcia-Arteaga et al., 1990). Jones (1990) described 



a similar pattern in ambon damselfish, Pomacentrus 

 amboinensis, Bleeker, 1868, and showed that adult 

 densities increased in proportion to juvenile recruit- 

 ment success after a two-year period (=maturation 

 time). 



The simultaneous decrease in medium-size fish 

 during August (Table 1) is not easily explained. The 

 population structure of unexploited reef fish is de- 

 termined by a complex sequence of events that in- 

 cludes recruitment, demographic changes (i.e. mor- 

 tality, migration, and growth), and ecological pro- 

 cesses such as habitat structure, resource availabil- 

 ity, and intraspecific competition (Jones, 1991). For 

 instance, variable patterns in growth can have a 

 major and direct influence on the size structure of a 

 population (Jones, 1991). Individuals that grow faster 

 may subordinate conspecifics and establish a domi- 

 nance hierarchy (Forrester, 1990). Once established, 

 hierarchies are often extremely stable (Morse, 1980). 

 For example, where individually recognized recruits 

 were followed over time, juvenile humbug, Dascyllus 

 aruanus (Linnaeus, 1758), never outgrew other group 

 members in size during an eight-month period 

 (Forrester, 1990). 



Lutjanus analis exhibits much variation in length 

 for a specific age (Mason and Manooch, 1985) but 

 matures sexually at fork lengths above 38 cm (Claro, 

 1981). In the central Bahamas, L. analis form groups 

 of variable-size fish that occupy seagrass meadows 



