584 



Fishery Bulletin 97(3), 1999 



adjusted to compensate for sampling losses (F). Abun- 

 dance at each sampling time was adjusted by sub- 

 tracting the number of individuals collected ( removed 

 from population) with the epibenthic sled from the 

 overall abundance at the site (site abundance = area 

 of site/area sampled xnumber of red drum collected). 

 Regressions of log^-adjusted abundance on age were 

 then run and mortality coefficients derived in this 

 manner were basically estimates of M since the ef- 

 fect of F was removed. Consequently, F was deter- 

 mined by subtracting adjusted mortality rate (M) 

 from Z. Because F values were relatively small, 2.3 

 to 6.9% of Z, only estimates of Z are presented. 



The relative recruitment potential of individual 

 cohorts was assessed by examining the ratio of 

 weight-specific growth (G) to mortality (Z). The ra- 

 tio is commonly used as an index of stage-specific 

 survival because it incorporates both growth and 

 mortality (Werner and Gilliam, 1984; Houde, 1996). 

 Because cohorts with G:Z ratios >1 gain biomass, the 

 probability of survival (recruitment potential) for 

 these individuals is assumed to be high. 



Data analysis 



Analysis of covariance (ANCOVA) was used to test 

 for intra- and interannual differences in growth and 

 mortality (covariate: age). Prior to each ANCOVA 

 test, a preliminary model ( interaction regression) was 

 tested to determine if the slopes of the regression 

 lines differed (homogeneity of slopes assumption; 

 Sokal and Rohlf, 1981). The main significance test 

 of the ANCOVA (homogeneity of y-intercepts) was 

 performed when the parallelism of slopes assump- 

 tion was met. Since differences in abundance affect 

 the elevation of regression plots used to estimate mor- 

 tality, spatial and temporal trends in mortality were 

 evaluated by comparing slopes (y-intercept test not per- 

 formed). Analysis of variance (ANOVA) was used to 

 examine variability in density. Estimates of density 

 were log -transformed to minimize heteroscedasticity. 



Results 



Environmental conditions 



Water temperature and salinity were recorded daily 

 from August to December in 1994 and 1995. Tempo- 

 ral variation in both parameters was pronounced and 

 trends were similar between years (Fig. 2). Tempera- 

 ture during the primary settlement period was vari- 

 able (September-October) and ranged from 23.3° to 

 30.2°C in 1994 and from 22.7° to 31.6°C in 1995, re- 

 spectively. Mean temperature during this period was 



similar between years 1995 (27.5°C) and 1994 

 (27.2°C). Salinity ranged from 25.3'?, to 37.07ff and 

 from 25.3%r to 30.8%r in 1994 and 1995, respectively 

 Similar to temperature, mean salinity was relatively 

 similar between years: m.OVcc (1994) and 28.67« 

 (1995). Peak values for both parameters occurred 

 during the initial spawning period and declined for 

 later spawning and settlement dates. Cohorts arriv- 

 ing early in the season experienced high tempera- 

 ture and salinity, whereas late-season cohorts were 

 exposed to lower temperature and salinity (Fig. 2). 

 No conspicuous differences in temperature or salin- 

 ity were observed between sampling sites (ABl, AB2). 



Catch characteristics 



Overall, 1306 and 8144 red drum lai^vae and early 

 juveniles (3-40 mm) were collected from the two sam- 

 pling sites in the Aransas Estuary in 1994 and 1995. 

 In both years, settlers were first detected in early 

 September and peak densities of recently settled red 

 drum (<10 mm) were present in early October (Fig 

 3). Larger postsettlers (>10 mm) were most abun- 

 dant from mid-October to early November. Mean 

 densities of recently settled red drum (<10 mm) 

 ranged from 0.0 to 1.7/m' in 1994 and from 0.0 to 

 4.1/m- in 1995; densities of all postsettlement red 

 drum (<40 mm) ranged from 0.0 to 3.0/m" in 1994 

 and from 0.0 to 4.2/m- in 1995 (Fig. 3). Maximum 

 densities (per sled tow) observed in 1994 and 1995 

 in the Aransas Estuary were 3.4/m" and 11.5/m-, re- 

 spectively. Interannual variation in postsettlement 

 density during the two seasons (October-November) 

 was small: 1.53/m2 (1994), 1.63/m'- (1995). No sig- 

 nificant differences in density were detected between 

 years or sites (ANOVA, P>0. 05). 



Although collection effort varied between years, 

 length-frequency distributions of red drum from the 

 Aransas Estuary were similar (Fig. 4 ). In both years, 

 the smallest individuals collected were 3 mm and 

 large recruits (>30 mm) were collected infrequently. 

 Catch rates showed a steeply ascending left limb 

 which peaked at approximately 8-9 mm, suggesting 

 that recruitment to seagrass meadows was complete 

 for these individuals. A long descending right limb 

 characterized catches for individuals >10 mm. 



Age and growth 



Age-length relationships (age-length keys) were devel- 

 oped for each year class and described by the following 

 equations: Age = -17.05 -i- 42.25 LogSL (1994, «=249, 

 r^=0.90, CV=8.9); Age = -11.43 -i-40.il LogSL (1995, 

 n = UO, 7^=0.90, CV=7.5). Predicted ages of red drum 

 (3-30 mm) ranged ft-om 6 to 45 d in 1994 and from 10 



