Brown-Peterson et al.: Meta-analysis of reproductive parameters of Lut/anus campechanus in the Gulf of Mexico 
43 
Figure 3 
Photomicrograph of ovarian tissue from a female red snapper ( Lutja- 
nus campechanus ) in the actively spawning reproductive subphase with 
24-h postovulatory follicles (POFs) that show evidence of daily spawning. 
OM=oocytes undergoing oocyte maturation. 
POF method. In contrast, spawning interval differed 
with the OM method when regions were considered 
separately. Although spawning interval remained sta¬ 
ble in the northeastern GOM with a median estimated 
value of 2.30 days in 1991 to 2.37 days in 2017, the 
median estimate for the northwestern GOM increased 
from 2.3 days in 1992 to 4.7 days in 2015 (Fig. 4, B 
and C). Additionally, estimations of spawning interval 
for the northwestern region showed more uncertainty 
because the upper 95% credible intervals were larger 
than those for the northeastern region, particularly 
from 2009 through 2015 (excluding missing data from 
2003 through 2008). In contrast, spawning interval re¬ 
mained relatively stable across time in both regions 
with the POF method (Fig. 4, E and F), but the north¬ 
east region had slightly higher estimated values for 
spawning interval (median estimated spawning inter¬ 
val: northeastern=2.2-2.4 days, 1991-2017; northwest¬ 
ern^.0-1.9 days, 1992-2015). Both regions showed 
higher uncertainty from 1999 through 2010 (excluding 
missing data from 2003 through 2008), but the upper 
95% credible intervals were larger in the northeastern 
region. However, it should be noted that mean spawn¬ 
ing interval estimates (Fig. 4, horizontal lines) are less 
precise when individual study-level spawning interval 
values are large. Large spawning interval values are 
inherently more uncertain because they correspond 
with a very low proportion of individuals spawning at 
any given time, which can occur in studies with small 
sample sizes or with sample collections focused either 
early or late in the spawning season. Thus, even a 
small amount of variation is greatly magnified when 
spawning interval is calculated (e.g., 2009 POF data 
and 2010 OM data). 
Batch fecundity 
Individual RBF values varied greatly within all studies. 
Values ranged from 0.179 eggs/g ovary-free body weight 
in 1994 to 394.0 eggs/g ovary-free body weight in 2013 
in the northeastern GOM and 0.048 eggs/g ovary-free 
body weight in 1998 to 557.39 eggs/g ovary-free body 
weight in 2001 in the northwestern GOM. Some of this 
variation was likely due to including RBF values col¬ 
lected from all months of the spawning season, includ¬ 
ing the beginning (April) and ending (September) of the 
spawning season. Despite the wide variation in individ¬ 
ual RBF values the model is moderately robust to outli¬ 
ers, which can be seen in the relatively narrow credible 
intervals (Fig. 5). There were no RBF data available 
from the northeastern GOM during 2005-2006 (2003- 
2008 for the northwestern GOM), and therefore those 
years are not included in the model. 
Relative batch fecundity showed a trend of decreas¬ 
ing values over time, particularly after 2012 (Fig. 5). 
Estimated mean (50% confidence interval [lower-upper 
as 25% and 75% Cl]) RBF values for regions combined 
were initially 58.0 (45.8-71.4) eggs/g ovary-free body 
weight in 1991, increased to 83.8 (74.4-93.5) eggs/g 
ovary-free body weight in 2001, then decreased to 
