416 
Fishery Bulletin 96(3), 1998 
either fecundity or spawning frequency, and none 
have used the methods developed by Hunter et al. 
(1985) for multiple-spawning fishes. Hunter’s meth- 
ods have been used successfully to estimate batch 
fecundity, spawning frequency, and annual fecundity 
in other Gulf of Mexico fishes similar to gag in size 
and longevity (e.g. Fitzhugh et al., 1988; Taylor and 
Murphy, 1992; Render and Wilson, 1992; Fitzhugh 
et al., 1993; Nieland and Wilson, 1993; Wilson and 
Nieland, 1994; and Collins et al., 1996). Such esti- 
mates obtained with these newer methods are far 
more accurate than the previously used “ovarian egg 
number” methods for multiple spawning fishes (as 
described in Hunter et al., 1985). 
In this study, we estimate sex ratio, spawning sea- 
son duration, length and age at first spawning, depth 
of spawning, batch fecundity, spawning frequency by 
year, spawning frequency by size and age, and an- 
nual fecundity in gag from the Gulf of Mexico. We 
also test the assumption that heavily pigmented gag 
represent only males using histological examination 
of gonads of heavily pigmented fish. 
Methods 
Gag were randomly sampled from commercial and 
recreational landings from Panama City to St. Pe- 
tersburg, Florida, in 1991, 1992, and 1993. In 1994, 
we asked commercial fishermen to leave all heavily 
pigmented gag ungutted so that they could be exam- 
ined for sex. Gag were considered heavily pigmented 
only if they retained the heavy-black pigment on the 
ventral portion of the abdomen after death (e.g. plate 
XVI, Fig. D., page 237, in Bullock and Smith, 1991). 
Total length (TL, in mm), fork length (FL, in mm), 
and total (ungutted) weight (TW, in g) were recorded 
before removing gonads, which were kept on ice prior 
to examination. 
Gonads of reproductively active gag are so large that 
they require subsampling for determination of sex. To 
ensure that a single tissue sample per female would be 
adequate to estimate maximum oocyte diameter and 
hydrated oocyte counts, we first tested for tissue ho- 
mogeneity. To test for homogeneity of oocyte diameter, 
three bilobed ovaries with hydrated oocytes were di- 
vided into six equal sections (three from each lobe) and 
subsampled to determine the frequency distributions 
of oocyte diameters; they were tested for homogeneity 
with a Kolmogorov-Smimov two-sample test (Sokal and 
Rohlf, 1981). To test for homogeneity of density of hy- 
drated oocytes, we used six hydrated ovaries divided 
in the same manner and determined the number of 
hydrated oocytes per gram; samples were tested for ho- 
mogeneity with a two-way AN OVA (SAS, 1988). 
We made a preliminary determination of sex in gag 
using fresh, unstained samples. A small sample was 
removed from each gonad, teased apart, and viewed 
microscopically (250x). Sex was assigned according 
to the following criteria: females had oocytes and no 
tissues that could be mistaken for spermatogenic tis- 
sue; males had no oocytes (or few small oocyte rem- 
nants); and transitional males had high numbers of 
oocytes and possible sperm (Shapiro et al., 1993). 
Female gonads were then microscopically staged. 
Maximum oocyte diameter (MOD) was recorded; the 
most advanced oocyte type present was noted and 
used to assign individuals to one of five ovarian-matu- 
ration stages: 1) mature, resting (MOD<0. 12 mm and 
clear); 2) early developing (MOD<0.20 mm and 
slightly opaque); 3) late developing (MOD=0.20 to 
0.59 mm and opaque); 4) ripe (MOD>0.59 mm and 
mostly transparent); and 5) spent (much loose par- 
ticulate matter in flaccid ovary) (West, 1990). After 
fat and mesentery were removed, gonads were blot- 
ted dry, weighed to the nearest 0.1 g., and placed in 
10% buffered formaldehyde solution. 
We then selected and prepared some of the gonad 
samples (above) to verify sex and stage using stan- 
dard histological techniques. Five-pm-thick sections 
were prepared from tissues embedded in paraffin and 
stained with Harris’s hematoxylin and eosin. Ova- 
rian stages were assigned on the basis of the most 
advanced oocyte or follicle stage present: 1) primary 
growth; 2) cortical alveolar; 3) vitellogenic; 4) hy- 
drated; and 5) spent (presence of postovulatory fol- 
licles [POFs]). Stages 1-4 followed Wallace and 
Selman (1981). Histological stages of possible early 
transitional males and functional males were as fol- 
lows: 6) possible early transitional males had active 
female tissue with possible crypts containing primary 
spermatocytes (Moe, 1969; Johnson, 1995); 7) ma- 
ture inactive males had a few secondary spermato- 
cytes and remnant oocytes (primary growth and 
atretic only); 8) mature active males had many sec- 
ondary spermatocytes and remnant oocytes; 9) rip- 
ening males had spermatids in ducts; and 10) ripe 
males had tailed spermatozoa in ducts. Sex and stage 
were compared by gonad-region for all heavily pig- 
mented gag. Physical condition of the gonad (e.g. in- 
tactness, presence of parasites, possible preservation 
problems) was also noted. 
We estimated duration of the spawning season by 
using two techniques: 1) plotting gonadosomatic in- 
dex (GSI=gonad weight (1Q0)/TW) by month for all 
gag with gonads in good condition (>90% intact) and 
2) determining when hydrated ovaries with non- 
degenerated POFs appeared. Proof of actual spawn- 
ing (and not just the presence of “ripe” ovaries, as 
discussed in Sadovy et al., 1994) was provided by 
