FISHERY BULLETIN: VOL. 87, NO. 1 



Table 1 .—South Carolina and Georgia tilefish, Lopholatilus cham- 

 aeleonticeps. landings from 1978 to 1986. Georgia landings of 

 tilefisfi and blueline tilefish, Caulolatilus microps, were not reported 

 separately routinely prior to 1985. Landings by gear type were 

 available for 1980-85, however, so rod and reel and electric reel 

 catches of unclassified tilefish (which are almost exclusively blue- 

 line tilefish) were excluded to yield a more accurate estimate of 

 Georgia tilefish landings. 



and Grossman 1986) to estimate maturity at age by 

 using a logistic model: 



lAndy Jennings, Fisheries Statistics Office. South Carolina Wildlife and 

 fvlarine Resources Department. Cfiarleston SC. pers. commun 1987, 



2S Gordon Rogers, Fisheries Statistics Coordinator, Georgia Department 

 of Natural Resources, Brunswicl^, GA, pers commun, August 1987. 



^Georgia Department of Natural Resources, 1983 Georgia landings an- 

 nual summary 1980. Ga Dep Nat. Resour,, Coastal Resour. Div., Bruns- 

 wick, GA. 



from about 1982 to 1984 were landed in Florida; 

 therefore, landings in Table 1 are a conservative 

 estimate of removals from these grounds. 



Low et al. (1983) provided a preliminary estimate 

 of maximum sustained yield (MSY = 162 t) for the 

 developing fishery. Their estimate was derived from 

 Gulland's (1971) model MSY = O.SMBq, where M 

 was the natural mortality rate and Sq was an esti- 

 mate of virgin biomass. Our objective was to pro- 

 vide updated yield estimates for use in managing 

 the tilefish fishery off South Carolina and Georgia. 

 We base these estimates on the 9-yr sequence of 

 catches from the developing fishery as well as re- 

 cently obtained information on the growth, mortal- 

 ity, and reproductive biology of tilefish from this 

 area. 



METHODS 



We used a deterministic sex- and age-structured 

 model to simulate the developing tilefish fishery and 

 to calculate sustainable yields. The model was based 

 on the following assumptions and data sources. We 

 assumed that the natural mortality rate (M) ranged 

 from 0.10 to 0.25 (Harris and Grossman 1985). We 

 used sex-specific estimates of weight-at-age because 

 the von Bertalanffy growth curves and length- 

 weight relationships used to calculate weight-at-age 

 differed significantly by sex (Harris and Grossman 

 1985). We used maturity-at-length data (Erickson 



Ps.a = 1/(1 + exp{-bMa,{L,a ' L50Mn,))) 



(1) 



where p^^ was the proportion of sex-s, age-a fish 

 that were sexually mature, 6ji^„, was a parameter 

 affecting the steepness of the curve, L^^ was the 

 standard length (SL) of sex-s fish at age a, and 

 L50ji^„, was a parameter representing the length at 

 which 50% of the fish were sexually mature. We 

 assumed that the total biomass of sexually mature 

 females (Sf) was an adequate measure of spawning 

 potential. 



We used a logistic model to relate selectivity to 

 length: 



set 



1/(1 + expi- bseliLs,, - LbOsel))) 



(2) 



where sel^^ was the proportion of sex-s, age-a fish 

 that were vulnerable to fishing, 6^^, was a param- 

 eter affecting the steepness of the curve, and 

 LbOggi was a parameter representing the length at 

 which 50% of the fish were vulnerable to fishing. 

 Based on length-frequency data (Harris and Gross- 

 man 1985), we assumed that female and male tile- 

 fish reached 50% vulnerability at about 475-500 and 

 500-525 mm SL, respectively. We used a slope 

 parameter (65^,) of 0.05 so that selectivity-at-age in 

 the simulated fishery ranged from about at age 

 5 to 1.0 at age 11 (Harris and Grossmar '985). 

 Parameter estimates used in the model are sum- 

 marized in Table 2. 



We assumed that the relationship between spawn- 

 ing stock size and subsequent recruitment was weak 

 or nonexistent because 1) tilefish produce pelagic 

 larvae (Fahay and Berrien 1981); 2) there may be 

 substantia! egg or larval transport between the Gulf 

 of Mexico and SAB (Katz et al. 1983); and 3) tile- 

 fish are dependent on the availability of shelter 

 (Valentine et al. 1980; Able et al. 1982; Low and 

 Ulrich 1983; Grossman et al. 1985). To represent the 

 stock-recruitment relationship, we used a Beverton- 

 Holt curve of the following form (Kimura 1988): 



N,.S + 6] = 



0.5Afe[0]g/[<]«/[0] 

 1 - A(l - Sa«]/S,[0]) 



(3) 



where N^f^lt + 6] was the number of sex-s, age-6 

 recruits in year t + 6, N(,[0] was the virgin recruit- 

 ment level for both sexes combined, and S,[t] and 

 S,[0] were the biomass levels for spawning females 

 in year t and prior to fishing, respectively. The pa- 



178 



