CRIMES ET AL.: REPRODUCTIVE BIOLOGY OF TILEFISH 



when they were functionally mature at 50 cm FL 

 as revealed by histological methods. 



A nonrandom mating system of pairing, involv- 

 ing mate selection by females, is consistent with the 

 both sexual dimorphisms observed in tilefish. 

 Female mate selection can convey a reproductive 

 advantage to large males (Ghiselen 1969; Howard 

 1979), but in randomly mating species, females are 

 usually larger than males (Ghiselen 1969). Female 

 ability to discriminate males is required to support 

 female mate selection (Howard 1979), which often 

 leads to development of specialized structures and 

 colorations by males for display (Krebs 1972; 

 Warner and Robertson 1978). The enlarged adipose 

 flap in tilefish certainly represents a conspicuous, 

 highly visible feature in some adult males that could 

 serve as a visual cue to signal male breeding status. 



The evolution of a female mate selection system 

 requires that a male have the ability to control 

 resources important to the female (Howard 1979; 

 Krebs and Davies 1984), and several lines of evi- 

 dence suggest that male tilefish may be territorial. 

 Burrowing is apparently the rule in the family 

 (Able et al. 1987). Direct observation from sub- 

 mersibles and mark-recapture data indicated that 

 tilefish orient to particular burrows, and may be 

 long-term residents of their habitats (Grimes et al. 

 1983, 1986). Furthermore, time-lapse photography 

 showed the same male-female pair of tilefish utiliz- 

 ing the same burrow over a 26-h period (Grimes et 

 al. 1986), and pair formation has been observed in 

 the branchiostegids Malacanthus plumieri (Clark 

 and Ben-Tuvia 1973), Hoplolatilus sp. (Thresher 

 1984), and H. starcki and H. cuniculus (Randall and 

 Dooley 1974). 



We believe that our data indicate that tilefish have 

 a mating system consisting of two classes of sex- 

 ually mature males, a category actively engaged in 

 spawning and a category of satellite males that do 

 not spawn. Similar breeding systems have been 

 described for several species of hermaphroditic reef 

 fish (Popper and Fishelson 1973; Fishelson 1975; 

 Warner and Robertson 1978; Robertson and Warner 

 1978; Warner and Hoffman 1980; Shapiro 1984). 

 For example, there are scarid and labrid populations 

 with large territorial terminal phase males that have 

 preferred mating status, and nonterritorial initial 

 phase males that have nonpreferred breeding status 

 (Warner and Robertson 1978; Robertson and 

 Warner 1978). Hermaphroditic populations of the 

 serranid Anthias squamipinnis in the Gulf of Eilat 

 have two behaviorally distinct types of males, a 

 dominant territorial male that actively courts, inter- 



acts, and spawns with females, and smaller males 

 that do not interact or spawn with females in the 

 social group. The latter male category have filamen- 

 tous degenerative gonads (Popper and Fishelson 

 1973; Fishelson 1975). These mating systems are 

 characterized by strong sexual selection and main- 

 tenance of reproductive territories by males, and by 

 being reef systems in which fish are habitat limited. 

 Tilefish are also severely habitat limited, i.e., to bur- 

 rowable clay substrate generally (Able et al. 1982; 

 Grossman et al. 1985; Grimes et al. 1986). 



During the period we studied reproduction, the 

 fishery for tilefish was rapidly expanding, and one 

 effect of fishing seems to have been to alter the 

 structure of the mating system. Based upon both 

 age-structured and non-age-structured population 

 modeling, tilefish population density was reduced 

 by about one-half to two-thirds from 1978 to 1982, 

 apparently due to the rapid expansion of the com- 

 mercial longline fishery (Turner 1986). Female size, 

 and particularly age, at maturity do not seem to 

 have been altered in any consistent fashion by the 

 population reduction from fishing. Males, on the 

 other hand, appear to have experienced profound 

 changes in visually assessed maturity; they clearly 

 were mature at smaller size (10 cm) and younger 

 age (2-2.5 years) in 1982 than in 1978. 



We interpret the decrease in size/age of maturity 

 in males to be the effect of fishing. Fishing lowered 

 population density, and in so doing may have made 

 mating territories available to smaller and younger 

 males. This interpretation is supported by the find- 

 ings of Warner and Robertson (1978) and Robert- 

 son and Warner (1978) that the ratio of the two 

 categories of sexually mature males (initial and ter- 

 minal phase) in western Caribbean scarid and labrid 

 populations was density dependent; i.e., relatively 

 more initial phase males were found in dense popu- 

 lations. 



ACKNOWLEDGMENTS 



Our tilefish research was initiated at the sugges- 

 tion of the late Lionel A. Walford. We tender 

 grateful appreciation to the following individuals 

 and institutions for their assistance in this research. 

 Joseph Desfosse, Susan Shepherd, Gary Shepherd, 

 and Stuart Katz helped extract data from samples. 

 Richard Trout and Bruce Babiarz provided statis- 

 tical and histological advice, respectively. Fran 

 Puskus, Louis Puskus, John Larsen, and the fisher- 

 men of Barnegat Light and Sea Isle City, NJ coop- 

 erated in obtaining tilefish samples. Daniel Erickson 



759 



