368 



Fishery Bulletin 88(2). 1990 



first appeared at age I (63.4 cm FL) as they did in 

 Trinidad, but Beaumariage believed that males 

 spawned initially at age III (77.0 cm FL) due to the 

 greater development of spermatogenesis in the testes 

 of older fish. However, in Trinidad initial spawning 

 takes place at age I, as shown by the presence of ripe- 

 running males in this age group. 



The spawning pattern of S. cavalla in Trinidad is ap- 

 parently similar to that off the coast of Ceara, in north- 

 eastern Brazil. Gesteria and Mesquita (1976) observed 

 year-round spawning with maximum intensity from 

 October through March, as in Trinidad. Spawning 

 throughout the year was also recorded by Ivo (1972), 

 but his period of maximum activity was from January 

 through June. Another study (Menezes 1969) indicated 

 that they spawn during the period October through 

 March. Further north, the spawning season is reversed, 

 where in the northeastern Caribbean it lasts from April 

 through November (Erdman 1976) and in the south- 

 eastern United States, from April through October 

 (Beaumariage 1973, Finucane et al. 1986). Spawning 

 migrations in North America are determined by tem- 

 perature (Moe 1972, Beaumariage 1973). In Trinidad, 

 however, spawning and migration may be influenced 

 by salinity changes because there is little variation in 

 the water temperature. Peak spawning starts after the 

 rains have set in, and may be triggered by a drop in 

 salinity; minimum salinities have been recorded in 

 August and September (van Andel and Postma 1954, 

 Edwards 1983). The presence of ripe-running males in 

 the samples indicates spawning in local waters, but the 

 spawning grounds remain to be discovered. The scarc- 

 ity of the fish during November through February 

 remains unexplained. Spawning on the outer Continen- 

 tal Shelf, 50-60 km offshore, as occurs in the north- 

 western Gulf of Mexico (McEachran et al. 1980) would 

 place part of the population beyond the reach of the 

 artisanal fishing fleet which is limited to some 40 km 

 offshore. Another possible factor could be migration 

 along the Venezuelan coast to the northwest and/or the 

 southeast, although part of the population may be resi- 

 dent throughout the year. Decreased vulnerability to 

 hook-and-line due to decreased feeding activity during 

 spawning, and/or gillnets due to spawning in waters 

 deeper than that in which the gillnet is effective, are 

 other possible reasons to be considered. 



Beach landing data indicate that a southerly migra- 

 tion takes place in Trinidad during the months when 

 the fish is seasonally abundant (Sturm et al. 1984). In 

 the present study, gonad analysis did not suggest any 

 migratot7 trends since ripe fish were taken around the 

 coast throughout the year. Moreover, observations on 

 stomach contents showed no clear feeding patterns as 

 evidence of migi-ation (unpubl. data). The spawning and 

 abundance patterns of S. cavalla approximately cor- 



respond to those of S. maculatus ( = S. brasiliensis), 

 which spawns throughout the year with more intense 

 spawning from October through April and is season- 

 ally abundant from May through September (Sturm 

 1978). S. brasiliensis also moves in a southerly direc- 

 tion during peak abundance, part of likely clockwise 

 movements around the island (Sturm 1978, Sturm et al. 

 1984). In Florida, Williams and Sutherland (1979) and 

 Sutherland and Fable (1980) have shown that S. cavalla 

 undertakes long-range migrations compared with the 

 shorter range migrations of S. m.aculatus, which is 

 closely related to S. brasiliensis (Collette et al. 1978). 

 In Trinidad, a similar situation would explain the dif- 

 ficulty in recognizing migratory trends of S. cavalla, 

 compared with the more local movements described for 

 S. brasiliensis (Sturm 1978). 



Female dominance in samples may be the result of 

 more females than males being hatched or mortality 

 being higher in males than females. Higher mortality 

 probably is associated with slower growth in males. 

 Alternatively, female dominance may be a function of 

 gear selection in gillnets or of behavioral differences 

 between the sexes. Increased voracity in females was 

 recorded in Brazil (Menezes 1969) and may explain why 

 the sex ratio favored females in hook-and-line samples. 

 Also, males may inhabit greater depths than females 

 and are less vulnerable to gillnets. If this was the case, 

 a meeting of the sexes for spawning by an upward 

 migration of males and/or a downward migration of 

 females could explain the increase in the male:female 

 ratio observed for gillnets during peak spawning. 



Acknowledgments 



Thanks are due to Michele Julien-Flus for assistance 

 in field work, to Sherry Manickchand-Heileman for 

 useful comments and discussion, to Bruce Lauckner of 

 the Caribbean Agricultural Research and Development 

 Institute for advice on statistical matters, to Rodney 

 Ramkissoon for assistance in figure preparation, and 

 to Avril Siung-Chang and Roland Bailey (Kings Col- 

 lege, London University) for reading the manuscript. 



Citations 



Alves, M.I.M, and G.S. Tome 



1967 Alguns aspectos do desenvolvimento maturativo das 

 gonadas da cavala, Sromheromoru!; ravnlln (Cuvier, 1829). 

 Arq. Estac. Biol. Mar. Univ. Fed. Ceara 7(1): 1-9 [in Portugese. 

 Engl. summ.]. 

 Bagenal, T.B.. and F.W. Tesch 



1978 Age and growth. In Bagenal, T.B. (ed.). Methods for 

 assessment of fish production in fresh waters, 3rd ed., p. 

 101-1.36. Blackwell Sci. Publ., Oxford. 



