FISHERY BULLETIN: VOL. 81. NO 4 



older tilefish; the mean marginal increment for fish 

 with 1-3 and 4 bands did not decline until late in 

 spring, while the mean marginal increment of older 

 fish usually began to decline in March (Fig. 3). Eleven 

 fish apparently had not formed annuli by 1 June. 

 They had hyaline edges and marginal increments 

 equal to 50-120% of the increase in their otolith size 

 in the previous year. Four of these fish (one female 

 and three of unknown sex) had formed 3 or 4 annuli, 

 six were females with 7-9 annuli, and one was a male 

 with 6 annuli. 



Otolith size increased with fish size and thus pro- 

 vided added evidence for the validity of using the 

 otoliths for aging tilefish. A log-log regression model 

 fit the data best. ANCO VA showed no significant dif- 

 ferences in slope or elevation between males and 

 females(P> 0.05) inthe50-73 cm FL range for which 

 there were nearly equal numbers of each sex. 

 Therefore, one least- squares regression line was 

 derived for all fish (r 2 = 0.90, n — 663) and converted 

 to a functional regression equation [In (FL) = 

 -0.4369 + 1.1112 In (OS)]. 



Seasonal length- frequency analysis of small tilefish 

 (2-30 cm FL) taken over several years (Fig. 4) showed 

 modes near the time of annulus formation that 

 agreed closely with the back-calculated lengths 



5i 



tt 



r~i .r-i 



Ik 



SEPT.- NOV. 



n , n 



1 — nr 



15 



10- 

 5- 



n 



E 



Z 



XL 



n 



DEC- FEB. 



5 i 



MARCH -MAY 



^^ 



XL 



10n 

 5 



n JUNE -AUG. 



 i— i , nn , nip 



10 20 30 



Length (cm) 



FIGURE 4. — Length frequency of small tilefish sampled by various 

 researchers (see text) in 3-mo intervals. 



(Table 1). These histograms were compiled from our 

 data, other published data (Bigelow and Schroeder 

 1953; Fahay and Berrien 1981), and unpublished 

 data (Fahay and Berrien 6 ). In the winter and spring, 

 three distinct modes occurred at 6- 11, 17-21, and27- 

 30 cm FL which approximated the mean back- 

 calculated lengths (see below) at ages 1, 2, and 3 yr, 

 respectively. Tilefish spawn from at least May into 

 October (Freeman and Turner footnote 4; Idelberger 

 et al. 1981; Fahay and Berrien 1981), and the small- 

 est fish observed were in the falL The progressive 

 increase in size of the smallest fish from the fall 

 through the summer, and the progressive increase of 

 other modes, also indicated that these modes rep- 

 resented age groups. 



Progressive increase of modes in the longline length 

 frequencies for 1977-79 (Fig. 5) also suggested that 

 age data were valid. In 1977 a mode existed at 41-45 

 cm FL; in 1978 there was a more pronounced mode at 

 51-55 cm FL; and in 1979 there was one at 56-60 cm 

 FL. The sizes of these fish in 1977-79 compared well 

 with mean empirical lengths at ages 4-6 of fish we 

 aged (see below). 



Longevity and Length at Age 



Length-at-age data suggested that males grow fast- 

 er than females, but females live longer. On average, 

 tilefish (sexes combined) grew about 10 cmFL/yrfor 

 the first 4 yr and thereafter growth slowed, especially 

 for the females (Table 1). After age 3, mean back- 

 calculated lengths of males were larger than those of 

 females. At age 4, males and females averaged 43 and 

 41 cm FL, respectively, and by the ninth year males 

 averaged 74 cm FL while females averaged 64 cm FL. 

 The oldest fish was a 35-yr-old female of 89 cm FL, 

 and the largest female was 95 cm FL at 32 yr. The 

 largest male was 112 cm FL at 20 yr old, and the 

 oldest male was 96 cm FL at 26 yr. 



Back-calculated growth increments for males and 

 females were significantly different in years 3-16 

 (ANOVA: P< 0.05). For years 17-25, they were either 

 not significantly different (P > 0.10) or not compa- 

 rable because of small sample sizes. Statistical 

 analyses were not performed on data involving back- 

 calculated lengths at age 1, because the OS:FL con- 

 version formula was fit to fish >16 cm FL so back- 

 calculations below that size (including nearly all 

 back-calculated lengths at age 1 ) may have been inac- 

 curate. Males achieved significantly more growth 

 than females in allyears (3-16) except the third (Dun- 



6 M. P. Fahay and P. L. Berrien, Northeast Fisheries Center Sandy 

 Hook Laboratory, National Marine Fisheries Service, NOAA, 

 Highlands, NJ 07732, pers. commun., 1982. 



756 





