Thompson et al.: Age distribution and growth of Senola dumenii 



363 



and-line catches from an ofFshore gas production plat- 

 form (;i=48). In general sampling was not system- 

 atic; we examined all fish available. 



Morphometric measurements and otoliths were 

 collected. Fish were measured by using fork and to- 

 tal lengths (FL and TL> in mm and weighed by using 

 total and gutted weights in g. Both sagittae were 

 removed and stored in 95*^ ethanol for later exami- 

 nation. Lapilli, asterisci, and dorsal spines were re- 

 moved from several fish to evaluate their usefulness 

 for age estimation. Otoliths were cleaned of organic 

 tissue by rinsing in a 50% hypochlorite solution 

 iClorox), air dried, weighed to the nearest 0.1 mg 

 with a Sartorius model 1801 microbalance, and 

 stored dry (Wilson et al.,1991). 



Age estimates were made by using techniques simi- 

 lar to those of Wilson et al. (1991) for billfish. Whole 

 otoliths were sputter coated with a mixture of gold 

 and palladium and viewed with reflected light un- 

 der a dissecting microscope for description of exter- 

 nal morphology. These samples were also viewed with 

 a Cambridge Stereoscan 150 scanning electron mi- 

 croscope (SEM) to obtain detailed photographs at 

 niagnifications ranging from 20 to 650x. 



Sagittae were embedded in epoxy resin (Spurr, 

 Kmbed 812, or Araldite GY 502) and sectioned with 

 a Buehler Isomet low-speed saw to yield a thin (about 

 1-mm) transverse section containing the core 

 I Beckman, 1989). Transverse sections were ground 

 with various grades (300-2000 grit) of wet and dry 

 .'sandpaper until the core was at the surface, then 

 polished with 0.3 |.im alumina polish. Thin sections 

 were mounted on glass slides with a clear thermo- 

 plastic cement (Crystalbond 509) and viewed under 

 transmitted light with an Olympus BH-2 compound 

 microscope at 50-250x magnification. Age estimates 

 from sectioned sagittae were made by combining 

 counts of translucent and opaque zones and other 

 growth features viewed under transmitted light and 

 counts of associated ridges on the ventral and 

 medioventral portions of the rostrum as determined 

 in billfish species (Wilson and Dean, 1983; Prince et 

 al., 1986; Wilson et al., 1991). 



Validation and verification of our age estimation 

 technique were attempted by using several tech- 

 niques. Validation of annulus formation using mar- 

 ginal increment analysis proved futile owing to our 

 inability to determine the condition of the otolith 

 edge. Therefore we pursued additional techniques 

 that would corroborate our age estimates. A mark- 

 recapture study was carried out from a gas produc- 

 tion platform (Mobil USA, West Cameron block 352) 

 located in the Gulf of Mexico 50 miles south of 

 Cameron, Louisiana. Greater amberjackt « =48) were 

 captured by hook and line, measured as described 



above, tagged at the base of the dorsal fin with a Hall- 

 print dart tag, injected with oxytetracycline hydro- 

 chloride (Agrimycin-100) at a rate of 20-40 mg/kg fish 

 weight, and released. Tagged greater amberjack (6) 

 were recaptured with hook and line or by spear fish- 

 ermen at the release site and sampled as described 

 above. Sectioned sagittae were viewed under ultra- 

 violet light (405-435 nm wavelength) at 50-250x 

 magnification for detection of the tetracycline mark. 

 Attempts made to compare age estimates from sag- 

 ittal otoliths with dorsal and anal spines and verte- 

 brae were not successful. To determine reproducibil- 

 ity of age estimates, sagittae were aged indepen- 

 dently by two readers. Reproducibility of age esti- 

 mates was compared by using the coefficient of varia- 

 tion (CV) and index of precision (D) (Chang, 1982). 

 Tag-recapture data for the Gulf of Mexico and 

 South Atlantic (7!=711) were obtained from the Co- 

 operative Gamefish Tagging Program (CGTP). Fish 

 that had been at large for at least 365 days after 

 tagging and that showed positive growth were used. 

 Twenty-five specimens met these criteria. Following 

 the method of Labelle et al. (1993), we used the von 

 Bertalanffy growth curve parameters to predict 

 changes in length from the tag recapture data to 

 verify our age estimates. Like Labelle et al. (1993) 

 we employed Fabens' ( 1965) length increment model 

 to predict length at recapture: 



lr,=l,+{L 



■/,)(l-e-*^'), 



where l^ 

 At 

 Ir 



length at release of individual i; 

 time of liberty of individual /; 

 estimated length-at-recapture of indi- 

 vidual ;; and 

 L^ and k = von Bertalanffy parameters estimated 

 from otolith ages. 



Predicted recapture lengths were then compared with 

 observed recapture lengths. As Labelle et al. (1993) 

 pointed out, this procedure is not statistically rigor- 

 ous but is useful for comparative purposes. 



An analysis of variance (ANOVA) was used to test 

 for sex-related differences in mean fork length, mean 

 gutted weight, mean sagittal weight, and mean age 

 among sources and within age class. Analysis of co- 

 variance (ANCOVA) was used to test for differences 

 in the relations (fork length [L] versus total weight 

 [Wt], fork length versus sagittal weight [SW], age 

 versus sagittal weight) between sexes (Cerrato, 1990; 

 Kimura, 1980). Statistical inferences were made with 

 a significance level of a=0.05. The relationship be- 

 tween fork length and otolith weight was modeled 

 with a negative exponential (von Bertalanffy) be- 

 cause it provided the best fit (highest /•-). 



