Kwok: Age and growth of Tiichiurus spp. 



751 



?=-0.097, df=917, P>0.90; T. nanhaiensis: f =-0.762, df=518, 

 P>0.44) and length (T. leptiirus: t=-0.47l, df=916, P>0.45; 

 T. nanhaiensis: <=-0.689, df=523,P>0.49). Therefore, aver- 

 age otolith length and weight were used in the analyses. 

 For all the linear regressions mentioned above, analysis 

 of covariance ( ANCOVA) was used to compare regressions 

 between sexes and species. 



We assigned 1 May and 1 June as the birth dates for 

 T. Ieptu7-us and T. nanhaiensis, respectively (Kwok and 

 Ni, 1999). Relative ages derived from aging were then con- 

 verted to absolute ages. Von Bertalanffy growth curves 

 were fitted by nonlinear regression on age and preanal 

 length data (SPSS vers. 7.5). The von Bertalanffy growth 

 equation for length is 



where PL^ 

 k 



PL, = PL Jl- el-*"- 'o'l), 



the asymptotic length; 

 growth coefficient; and 



Iq = the hypothetical age at zero length. 



Plots of residuals from regression models were used to check 

 the assumption of normality. ANCOVA was used to com- 

 pare log-transformed age-at-length regressions between 

 sexes and species. 



Results 



The preanal length (mm) and gutted weight (g) regres- 

 sion models were significantly different between sexes 

 (ANCOVA: T. leptiirus: F.-,g.y,=A.m.P<Om\ T. nanhaiensis: 

 P., 530=3.34, P<0.05) and'species (ANCOVA: F.,.^^^^=83.76, 

 P<0.001). The regression models were 



T. leptiirus 

 males: 



W= 1.513 X 10-*PL2'"i (H=212, r2=0.9777,P<0.001); 

 females: 



W = 1.748 X 10-» PL- 5-»9 (w=724, ?-2=0.9761, P<0.001 ); 

 sexes combined: 



W= 1.624 X lO-iPL-sei (,,=936, ,-2=0.9771, P<0.001); 



T. nanhaiensis 

 males: 



W = 3.363 X 10-"^ PL2«-i6 („=282, r2=0.8506, P<0.001); 

 females: 



W = 6.553 X 10-'^PL-^29 (,,=252, r2=0.9299.P<0.001); 

 sexes combined: 



W = 5.672 X 10-5 p/^2 755 (,,=534^ ,.2 =0.8968, P<0.001 ). 



Sectioned sagittae of both species had an opaque nucleus 

 located above the sulcal groove toward the dorsum. The 

 nucleus was surrounded by a pattern of alternating wide, 

 translucent zones and thin, opaque zones; the latter were 

 considered annuli (Fig. 3). Annuli were distinct on the 

 dorsum of the sections but were usually indecipherable 

 on the ventral side. A total of 757 and 534 otoliths were 

 embedded in resin and sectioned for T. lepturus and T. 

 nanhaiensis. respectively. Of these, 33 (4.3%) and 9 ( 1.7%) 



were unreadable, and the percentage agreement between 

 the two readings for each species was 95.4% and 92.7%, 

 respectively. 



The least marginal increment values (Fig. 4) occurred in 

 February for both species, suggesting that one growth ring 

 (annulus) formed each year. Only specimens age 1-4 were 

 included in the analyses because older fish were rare in 

 our collections. The mean otolith annular radius (MOAR) 

 of the first annulus (AN OVA: F^ ^^.,=3.046, P<0.05) and 

 third annulus (ANOVA: P3go=4.024, P<0.05) of T. lepturus 

 were significantly different among different age groups 

 (Fig. 5); MOARs increased slightly with older age groups. 

 However, no particular trend was found with regard to the 

 MOARs of T. nanhaiensis (Fig. 5). Lee's phenomenon was 

 not evident for either species, although reverse Lee's phe- 

 nomenon was possible for T. lepturus. 



Otolith weight accounted for 68.7% and 68.9% (Table 2) 

 of the variability in age for T. lepturus and T. nanhaien- 

 sis, respectively. A negligible amount of the remaining 

 variability was explained by considering otolith length 

 in addition to otolith weight. The otolith weight-age 

 regression was improved by fitting the untreated vari- 

 ables (otolith weight and age) with simple linear regres- 

 sion models: 



T. lepturus: 



OW = 6 3533 -I- 5.2913Age {n=718, ,-2=0.7168, P<0.001); 

 T. nanhaiensis: 



OW = 6.3921 -I- 3.6850Age (;!=515, ,2=0.7561, P<0.001). 



These regi-ession results suggest a linear relationship 

 between otolith weight and age (Fig. 6). The regression 

 models were significantly different between the two spe- 

 cies (ANCOVA: P2.i229=224.17, P<0.001). Normal prob- 

 ability and residual plots showed that the regressions 

 complied with the assumptions of normality and homoge- 

 neous variance. 



Von Bertalanffy growth equations for both species were 



T. lepturus 

 males: 

 PL = 755.2 (1 -e l-o.ii6(/ + 2.73711) (,, = 146, ,-2=0.684, 



P<0.001); 

 females: 

 PL = 601.4 (1-e l-0-i58(( + 2.85011) (,z=578, ,-2=0.765, 



P<0.001); 

 sexes combined: 

 PL = 589.1 II - e 1-0 168 r/ * 2.682.I) (,,=724, ,-2=0.749, 



P<0.001); 

 T. nanhaiensis 

 males: 

 PL = 501.7 II -e 1-0306"* 167311) (,,=281, r2=0.682, 



P<0.001); 

 females: 

 PL = 612.6 II - e 1-0 220 1 ( * 1.79211) (,,=244, ,-2=0.726, 



P<0.001); 

 sexes combined: 

 PL = 602.1 II -e I-0 207W + 2.0441I) (,1=525, ,-2=0.699, 



P<0.001). 



