Biesiot et al.: Ovarian development in Rachycentron canadum 



689 



where the null hypothesis was rejected (oc<0.05), 

 nonparametric Tukey-type multiple comparisons 

 were performed according to Zar (1984) in order to 

 determine between which of the mean values signifi- 

 cant differences occurred. 



Results 



Histology 



Histological analyses were performed on the gonads 

 of 115 female cobia collected from the northern Gulf 

 of Mexico over the course of two breeding seasons 

 (n=42 in 1991 and n=73 in 1992). Of these fish, 14 

 were caught in Florida waters, 6 in Alabama, 60 in 

 Mississippi, 26 in Louisiana, and 7 in Texas; loca- 

 tion data could not be obtained for two fish, but they 

 were probably caught in either Mississippi or Loui- 

 siana waters. We observed, as had Lotz et al., 2 that 

 cobia oocyte production appeared to be group syn- 

 chronous as defined by Wallace and Selman (1981), 

 such that each ovary examined contained oocytes at 

 different stages of maturation. However, ovaries 

 could be assigned to specific categories based on the 

 dominant oocyte maturity stage. 



In stage-1 previtellogenesis, the oocytes were 

 small, compact, and irregularly shaped (Fig. 2A). The 

 previtellogenic stage comprised three substages: a) 

 early previtellogenesis, characterized by small oo- 

 cytes in which the nucleus had swollen to form a large 

 germinal vesicle; b) middle previtellogenesis, char- 

 acterized by nucleoli developing within the nucleus 

 and causing evaginations to form in the nuclear en- 

 velope; and c) late previtellogenesis, characterized 

 by the presence of cortical alveoli. The latter substage 

 marked the beginning of the transition to stage 2. 



In stage-2 vitellogenesis, the oocytes increased in 

 size as the yolk material increased (Fig. 2B) and 

 formed unevenly dispersed lipid vacuoles. Vitel- 

 logenic oocytes were somewhat more rounded and 

 were not as compacted as previtellogenic oocytes. 



During stage-3 final maturation, the oocytes were 

 larger and the lipid vacuoles and proteinaceous yolk 

 material had become more evenly dispersed (Fig. 2C). 

 The lipid droplets fused and congregated around the 

 periphery of the oocytes, resulting in a clearing of 

 that region of the cell. Note that although most of 

 the oocytes in Fig. 2C were stage-3 oocytes; some 

 stage-1 oocytes and late stage-2 oocytes were also 

 present. Chromosomes condensed during stage 3 for 

 the initiation of meiosis (Fig. 2D). 



During stage-4 postovulation, unspent oocytes and 

 postovulatory follicles (POF) were resorbed (Fig. 3A). 

 The oocytes became distorted and compacted, as did 



the POF. Residual lipid vacuoles were observed in 

 the resorbing oocytes. (A few early previtellogenic 

 oocytes can also be seen in Figure 3A, concurrent 

 with the resorption process. ) 



A sequential round of ovarian development was 

 observed in some cobia ovaries categorized as stage 

 1' (Fig. 3B). The presence of resorbing oocytes and 

 POF in ovaries suggested that a prior spawning epi- 

 sode had recently occurred. Early previtellogenic 

 oocytes, resorbing oocytes, and POF were not seen 

 simultaneously in ovaries categorized as stage 1. A 

 sequential vitellogenic stage, stage 2', characterized 

 by oocytes with numerous small lipid vacuoles, was 

 also observed (Fig. 3C). Previtellogenic and resorb- 

 ing oocytes as well as resorbing POF were also 

 present in the ovary during this stage. 



Timing of ovarian development 



The stages of cobia ovarian development were tabu- 

 lated according to month of capture for 1991 and 1992 

 data combined (Table 2). In both April and May, 14- 

 15% of the ovaries were developing (stages 1 and 2), 

 -60% of the ovaries were ripe and about to be 

 spawned (stage 3), -20% were postspawning (stage 

 4), and 3—5% had already spawned but were prepar- 

 ing for a sequential spawning episode (stages 1' and 

 2'). The similarity in ovarian developmental stages 

 for these two months is not surprising because all 

 the April fish were collected in the last week of the 

 month whereas all the May fish were collected in the 

 first week of that month. During July (first and sec- 

 ond weeks of the month), again -15% of the ovaries 

 were developing (stages 1 and 2) but only -30% were 

 ripe (stage 3); 15% were postspawning (stage 4) and 

 over 40% had already spawned at least once and were 

 in the process of developing for a subsequent spawn- 

 ing (stages 1' and 2'). Fewer numbers offish were 

 collected in August ( last week of the month) and Sep- 

 tember (first week of the month) but the predomi- 

 nant stages of ovarian development were dramati- 

 cally different from fish collected earlier in the sea- 

 son. The majority of ovaries (over 80% ) were previtel- 

 logenic or vitellogenic (stages 1, 2, and 2') whereas 

 fewer than 20% were ripe (stage 3) or postspawn 

 (stage 4); no stage-1' ovaries were seen. 



Gonosomatic index 



Cobia with ovaries in stages 1, 2, and 3 had increas- 

 ing mean GSI's of 1.1 ± 0.6, 5.0 ± 2.2, and 5.4 ± 2.2, 

 respectively (Fig. 4A). GSI declined to 3.5 ± 1.6 in 

 cobia with stage-4 ovaries (postovulation); the lower 

 GSI reflects the loss of oocytes to spawning. Almost 

 all of the pairwise comparisons were significantly 

 different (Tukey-type multiple comparison test, 



