GKIMES ET AL.: REPRODUCTIVE BIOLOGY OP' TILEFISH 



Branchiostegus japonicus) all exhibit extended re- 

 productive seasons centered around summer (Haya- 

 shi 1977, 1979; Dooley 1978; Ross and Merriner 

 1983). However, Erickson et al. (1985) reported that 

 L. chamaeleonticeps spawned from March through 

 June off Georgia in the South Atlantic Bight. The 

 reason for the shorter reported spawning season for 

 L. chamaeleonticeps in the South Atlantic Bight is 

 uncertain. 



Fractional spawning by individual females was in- 

 dicated by the polymodal frequency distribution of 

 ovum diameters. This asynchronous follicular devel- 

 opment is typical of fractional spawning (deVlam- 

 ing 1983), and has been reported for L. chamaele- 

 onticeps in the South Atlantic Bight (Erickson et al. 

 1985) and Mid-Atlantic Bight (Morse fn. 9). Frac- 

 tional spawning based upon polymodal ovum-diam- 

 eter distributions has also been reported for the con- 

 familials C. microps and B. wardi (Dooley 1978; 

 Ross and Merriner 1983). 



Our estimates of fecundity in tilefish are consis- 

 tent with those in the literature. We agree with Gale 

 and Deutsch (1985) that the term fecundity is incor- 

 rectly applied to many fishes, especially fractional 

 spawners, because in most cases there is no reason- 

 able means to determine which or how many oocytes 

 or developing ova will be released, or how many 

 ovarian ova will be resorbed after spawning and 

 never released. Therefore, we recognize that our 

 ovarian egg count data provides only a rough esti- 

 mate of actual fecundity. Erickson and Grossman 

 (1986) found the relationship between fecundity and 

 weight to be best described by a log transformation 

 (log, F = 1.497 log, W + 12.59, r- = 0.93); while 

 Morse (fn. 9) found the relationship best fitted to 

 a linear form (F = -966,471 + 887 W, r"~ = 0.61). 

 To compare our estimates further we calculated 

 relative fecundity using extreme point estimates of 

 fecundity reported by other authors (Erickson and 

 Grossman 1986, 414 -g"^ and 950 -g'' for 2.0 and 

 8.0 kg fish; Morse fn. 9, 543 -g-^ and 867 -g-^ for 

 3.5 and 9.0 kg fish) and this study (119 -g"' and 

 769 -g"' for 2.1 and 13.0 kg fish), and using predic- 

 tive equations (Table 10). Comparing the range of 

 point estimates, the three studies are similar, how- 

 ever our findings agree more closely with Erickson 

 and Grossman (1986) when the comparison is based 

 upon predictive equations. The inconsistency in the 

 comparison is because Morse (fn. 9) used a linear 

 equation, thus assuming that egg production per 

 gram of body weight was constant for all body 

 weights, and Erickson and Grossman (1986) and this 

 study chose curvilinear equations. It is possible to 



conclude that small tilefish, like some other fishes 

 (Grimes 1987), produce fewer eggs per unit body 

 weight than larger fish. 



We compared numbers of ova actually found in 

 ripe ovaries collected on 1 July (n = 3) and 28 

 August {n = 6) to ovarian egg numbers predicted 

 for the same size females to estimate what propor- 

 tion of ova had been spawned in the mid- and late- 

 spawning season. The analysis suggested that ap- 

 proximately 25% of ovarian eggs were spawned by 

 July and 50% by the end of August. The ovaries of 

 two postspawning females collected on 10 October 

 contained considerable quantities of ovarian eggs 

 >0.15 mm in diameter. These ova accounted for 

 15-20% of the predicted ovarian egg number. By 

 December 11 (n = 3) the number of ovarian eggs 

 >0.15 mm in diameter had decreased to about 5% 

 of the predicted maximum ovarian egg number. 

 Since atretic ova were observed in histological 

 preparations of ovaries in spent and resting stages, 

 and resorption is well documented in other teleosts 

 (Hoar 1957; Smith 1965; Combs 1969; Foucher and 

 Beamish 1977; LaRoche and Richardson 1980; 

 Waltz et al. 1982), it seems reasonable to conclude 

 that at least 15-20% of the maximum ovarian egg 

 number are never spawned and are resorbed dur- 

 ing the winter. 



Sexuality 



Our results agree with Erickson and Grossman 

 (1986) that tilefish are gonochoristic and that sec- 

 ondary gonochorism is a possibility. Gonad micro- 

 structure and development of adult (>50 cm FL) 

 tilefish were typical of most male and female ovi- 

 parous teleosts, and identical to that described for 



Table 10. — Comparison of ovarian egg number and 

 relative ovarian egg number for small (2.0 kg) and large 

 (9.0 kg) female tilefish calculated following Erickson 

 and Grossman (1986), Morse,' and this study. 



'Morse, W. W. Length, weight, spawning and fecundity of 

 the tilefish, Lopholatilus chamaeleonticeps. from New Jersey 

 waters, Unpubl. manuscr Northeast Fish. Cent. Sandy Hook 

 Lab , Natl Mar. Fish. Serv., NOAA, Highlands, NJ 07732. 



757 



