584 



Fishery Bulletin 105(4) 



Present study n=650 



 Females 



CH Males 



^B Transitional-stage fish 



0O0OO00OO0C3OOC3O0OC3OOO0O00O 



1— ■r-'-T-CMCgc\jcsjc\jco<r)(T>':ri<n^ tj-tj- Tj-TrinininmintDco 



B 



T-FO- 



Video survey n=1441 



Hkte 



l~rTTTl m i i 



^ 



Headboat survey 0=1834 



NTT-h-^ -r ^ 



Red porgy size (mm TL) 



Figure 1 



Percent length frequencies (mm total length) of red porgy (Pagrus pagrus) (A) 

 collected by hook-and-line gear in the present study, (B) derived from laser mea- 

 surements from the National Marine Fisheries Service Gulf of Mexico stationary 

 video reef fish surveys, 1995-2002, and (C) in catches of recreational headboats 

 from northwest Florida and Alabama during 1999-2001 sampled by the National 

 Marine Fisheries Service headboat survey. 



fish behave more like males or females, this test was 

 run two ways: one in which transitional-stage fish were 

 considered males and another in which they were con- 

 sidered females. Transitional gag and scamp may be as 

 susceptible to behavior-related sex selectivity as fully 

 transformed males if they adopt male behaviors before 

 completing sex change (Coleman et al., 1996). The lyre- 

 tail anthias (Pseudanthias squamipinnis) exhibits such 

 a shift to male behavior before any visible morphologi- 

 cal changes (Shapiro, 1979). 



Results 



A total of 27 collections of red porgy were made, both 

 during and outside the spawning season, which runs 

 November-March but peaks December-February (DeVr- 

 ies, 2006). Sample sizes were >20 in 21 of 27 instances 

 (range: 10-31). Mean sizes and 95% confidence intervals 

 (CI) of males, females, and transitional-stage fish were 

 296 ±4 mm, 270 ±3 mm, and 291 ±7 mm TL, respec- 

 tively. The overall size distribution is shown in Figure 

 1. Mean proportion and 95% CI of males in 19 collec- 

 tions from five of the seven sites was 0.39 ±0.07, assum- 

 ing transitional-stage fish were males, and 0.24 ±0.05, 



assuming transitional-stage fish were females. The two 

 remaining sites were persistently and significantly more 

 male-biased than the other sites (DeVries, 2006). Mean 

 proportions of males at the eight collections from the 

 two remaining sites were 0.66 ±0.10 (transitional fish as 

 males) and 0.53 ±0.08 (transitional fish as females). For 

 further details on the temporal and spatial distribution 

 of the collections, see DeVries (2006). 



The order in which red porgy were caught in each of 

 27 collections was random with respect to both size and 

 sex, i.e., there was no evidence that larger fish or males 

 were more likely to be caught before smaller individuals 

 or females (Fig. 2). In both the "above and below the 

 median" and "up and down" runs tests, the hypothesis 

 that the sequence of lengths was random could not be 

 rejected at o=0.05 for any of the collections (Table 1.1 

 in DeVries, 2006). Whether transitional-stage fish were 

 considered males or females, the hypothesis that the 

 order in which sexes were caught was random could not 

 be rejected at o=0.05 for all but four of the collections 

 (Table 1.2 in DeVries, 2006). In each of those four col- 

 lections only one male was caught. The statistical table 

 used for determining critical values (Table 28; Sokal 

 and Rohlf, 1981b) required a minimum sample size of 

 two for both of the dichotomous variables used (number 



