NOTE Gobert et al.: Biology of Etelis oculatus in the Caribbean 



419 



Fishing gears used 



In all islands but Guadeloupe, gill nets had mesh sizes 

 of 65 mm (knot to knot) and a stretched height of 6.4 m. 

 In Guadeloupe, mesh was 60 mm and height was 4 m; 

 in addition, the nets were given more slack than in 

 Martinique to increase their efficiency, and thus caught 

 a wider size range offish. 



Trammel nets had mesh sizes of 40 mm (knot to knot) 

 on the central panel and 200 mm on the outer panels, 

 and were 2 m high. All nets (trammel nets and gill 

 nets) were set overnight (15 to 20 hours of fishing time) 

 in units of 200 or 300 m. 



Three types of longlines were used in the fishing ex- 

 periments. Vertical longlines were derived from those 

 used by fishermen in the Lesser Antilles and had about 

 20 hooks on 40 cm-long secondary lines. Pole longlines 

 were adapted from a technique used in Florida and 

 Puerto-Rico: poles about 2 m long are fastened to the 

 main line lying on the bottom, each having 12 to 25 

 hooks on very short secondary lines (20 to 30 cm). Re- 

 inforced longlines are horizontal longlines whose main 

 line is heavier, in order to fish on very rough grounds. 

 All longlines were hauled after 30 to 45 minutes fish- 

 ing. For the analysis, no distinction was made between 

 samples of these three types of longlines. Various kinds 

 of longlines are used in the small-scale queen snap- 

 per fishery in Saint-Lucia. Handlines used in the ar- 

 tisanal and semi-industrial Hondurian fisheries are 

 either mono- or multifilament, and bear one or several 

 hooks. No detailed observations were made on the size 

 of hooks or on the bait used in the commercial queen 

 snapper fisheries. 



1987) could be done reliably in the field for adults and 

 juveniles, but had to be confirmed under the microscope 

 for the smallest individuals (less than 10 cm). Fish 

 length was the only information recordable from profes- 

 sional landings (St-Lucia, Honduras); fishing experi- 

 ments yielded more detailed data, by order of decreasing 

 frequency: length (fork length FL, total length TL, or 

 both; unless specified, all lengths mentioned in the 

 text are fork lengths), weight, and sexual stage, and 

 occasionally a few additional observations (such as 

 unusual number or length of fin rays). Sexual stages 

 were identified by using the macroscopic scale defined 

 by Barnabe (1973) and were coded as follows: 1 (imma- 

 ture, without identifiable sex), 2 (immature, of identifi- 

 able sex), 3 (mature), 4 (prespawning), 5 (spawning), 

 6 (postspawning), and 7 (resting). Depth was recorded 

 only in the fishing experiments; for gillnet and tram- 

 mel-net stations, it was measured at each end of the net, 

 and the depth used in the analysis was the average of 

 these two values. 



Length-frequency analysis 



In most cases, length-frequency analysis was strongly 

 hindered by gear selectivity and sample sizes. We 

 attempted to estimate L, and ZIK with the method 

 of Wetherall et al. (1987) applied to the sample of the 

 semi-industrial Hondurian fishery. All other samples 

 were unsuitable for length-frequency analysis because 

 of severe violations of one or several assumptions, prin- 

 cipally regarding constant catchability above the full 

 selection length, which was obviously not the case for 

 the three gears used in the fishing experiments. 



Data collected 



None of these studies was specifically designed for 

 the study of E. oculatus, and therefore the nature and 

 amount of available information (sampling coverage 

 though time, space, and depth) for this species were 

 variable. Species identification (Allen, 1985; Anderson, 



Results 



Depth distribution 



During the fishing experiments, E. oculatus of market- 

 able size (i.e., larger than about 20 cm) were caught 



