Brown-Peterson et al.: Reproductive biology of Rachycentron cancidum 
17 
hydration and embedment in paraffin, following standard 
histological techniques. The paraffin blocks were sectioned 
at 4 pm by using a rotary microtome. Duplicate slides 
were prepared for each tissue, resulting in a total of four 
slides for each cobia specimen (two from each gonad). The 
slides were stained with Gill’s I hematoxylin and eosin 
phloxine (Polyscientffic Corporation) following standard 
histological procedures. 
Three separate views from each side of the gonad of 
each fish were examined to determine maturity stages. 
Ovarian maturity classes were based on those previously 
described for cobia by Lotz et al. (1996). The entire ovar- 
ian section was examined for the presence of postovulato- 
ry follicles (POF) and oocytes undergoing final oocyte mat- 
uration (FOM). POF stages were classified following the 
methods of Hunter et al. (1986), although age estimates 
for POF stages in cobia are unverified. FOM stages were 
classified following Brown-Peterson et al. (1988). Follow- 
ing inspection of the entire ovarian section, three areas 
were arbitrarily selected from each slide for quantification 
of oocytes. Oocytes in all stages of development (including 
atretic oocytes) and POFs were counted at lOOx and the 
percentage of each oocyte stage in the field of view was es- 
timated. Oocyte atresia stages were classified by following 
the methods of Hunter and Macewicz (1985a). 
The entire testicular section from each cobia was exam- 
ined to determine the maturity classification for male fish. 
Three arbitrarily selected portions of each section were ex- 
amined at lOOx and 400x to classify all stages of spermato- 
genesis observed. Testicular maturity stages were based 
on those described for cobia by Lotz et al. (1996). Partic- 
ular attention was given to the presence and amount of 
spermatogenesis in the testis. 
Estimates of batch fecundity 
Batch fecundity was estimated from the counts of oocytes 
in samples of ovarian tissue. Oocyte counts were obtained 
after teasing oocytes from tissues fixed in either GF or 
NBF for three to four months or from histological evalua- 
tion of tissue sections. The volumetric method was used to 
estimate fecundity for tissues fixed in GF or NBF (Bage- 
nal and Brauin, 1971). All oocytes freed from each GF 
or NBF sample were placed in 50 mL of water, stirred 
to homogeneity, and ten 1-mL samples were removed, 
combined, and the total volume brought to 50-70 mL 
with water. The diluted sample was stirred to homoge- 
neity and 1-mL subsamples were removed, counted and 
replaced three to six times for each ovarian sample. All 
oocytes >700 pm from each subsample were counted and 
measured by using a stereo dissecting microscope and a 
computerized image analysis system. Oocytes of this size 
were used because Lotz et al. (1996) previously showed 
that cobia have a distinct mode of large oocytes prior to 
spawning. Typically, 25-100 oocytes were measured and 
counted in each subsample. 
Estimates of batch fecundity based on histological eval- 
uation were obtained by counting the number of oocytes 
undergoing FOM in six fields under a compound micro- 
scope at lOOx magnification. The area of a single field of 
view was determined to be 0.0249 cm 2 by using a stage 
micrometer. The number of oocytes in final maturation ob- 
served in a field of view was converted to the number per 
mL by the formula 
N x 0.0249 3/2 = Nx 0.003939. 
FOMs were counted as 1 if >50% of the oocyte was in the 
field of view and were uncounted if <50% of the oocyte was 
in the field of view. The total number of FOMs in a fish was 
then determined by multiplying the estimated number 
per mL by the total volume of the ovaries. In all cases, 
fecundity was expressed as both batch fecundity (mean 
number of eggs/batch) and relative fecundity (number of 
eggs/gram of ovary-free body weight). 
Ovarian volume was determined by volumetric displace- 
ment. The observed relationship between ovarian weight 
and ovarian volume was determined and that relationship 
was used to estimate ovarian volumes of fish for which di- 
rect volume measurements were unavailable. The analy- 
sis was restricted to fish with ovarian weights >500g. 
Estimates of spawning frequency 
Two methods based on histological observations were used 
to estimate spawning frequency of cobia: 1) the percentage 
of females in the late developing ovarian class with 0- to 
24-h POF in the ovary and 2) the percentage of females in 
the late developing ovarian class undergoing FOM. Only 
fish in the late developing ovarian class were included 
in these analyses, because this is the only class in which 
cobia have the potential to spawn. For both methods, esti- 
mates of spawning frequency were determined according 
to the procedure of Hunter and Macewicz ( 1985b). The per- 
centage of fish in the late developing maturity stage with 
ovaries containing either FOMs or POFs was calculated 
for each month in each region. This value represents the 
percentage of the fish in the population that are about to 
spawn (FOMs) or have just spawned (POFs). Spawning 
frequency (the number of days between spawnings) was 
determined by dividing 100 (representing the total popu- 
lation of fish) by the percentage of fish with FOMs or POFs 
in the ovaries. 
Statistical analysis 
Student’s t-test was used to test for differences in GSI 
values between years. Batch fecundity data were tested 
for normality and homogeneity of variance. Simple linear 
regression was used to test the relationship between batch 
fecundity as the dependent variable and FL or ovary-free 
body weight as the independent variable. One way analy- 
sis of variance was determined for relative batch fecundity. 
A Mann-Whitney U test was used to compare fecundity 
estimates for the various methods used to determine 
fecundity. A chi-square test was used to test for differences 
in spawning frequency among areas. All statistics were 
computed by using SPSS-PC version 7.5 (SPSS, Inc., 1997) 
or Systat 8.0 (SPSS, Inc., 1998). Results were considered 
significant if P < 0.05. 
