WANKOWSKI: ESTIMATED GROWTH OF SURFACE-SCHOOLING TUNAS 



(1968), and others since, have pointed out the dan- 

 gers of extrapolating the results beyond the range 

 of observations on which they are based. However, 

 provided that these limitations are recognized, 

 and the results considered as representative of 

 average growth of the stock, the length-at-age 

 data are reasonable estimates of growth during 

 the period of the study. 



However, several qualifying comments must 

 be made regarding these estimates. Length- 

 frequency samples are subject to errors from two 

 main sources: 1) Suzuki (1971) showed that sam- 

 ples obtained using a size-specific fishing 

 technique are unlikely to be representative of the 

 size class sampled. Fish at either extreme of the 

 size range would be underrepresented: the mean 

 length of modal groups is likely to be overesti- 

 mated at the lower extreme and underestimated 

 at the upper. 2) Josse et al. (1979) demonstrated 

 that the modal progression method is sensitive to 

 inadequate sampling: estimated growth rates 

 vary widely if few landings are sampled. However, 

 since daily landings from virtually every operat- 

 ing vessel were sampled, possible problems arising 

 from the second source of error were minimized, if 

 not eliminated entirely. The modal progression 

 method itself is considered to be subjective (e.g., 

 Joseph and Calkins 1969), both in view of the 

 methods used to determine modal lengths and in 

 connecting modal values to form serial progres- 

 sions. However this technique has proved effective 

 provided that the derived growth functions are 

 considered estimates only. 



Various studies on the growth of skipjack and 

 yellowfin tunas have been conducted throughout 

 the world, but few have been carried out in the 

 western Pacific. Good estimates for the growth of 

 medium-sized skipjack tuna are available from 

 the Papua New Guinea tagging study. As Le Guen 

 and Sakagawa (1973) pointed out, a comparison of 

 the von Bertalanffy parameters k, L-^, and ^o 

 often gives the misleading impression that growth 

 is different in different regions. Two recent studies 

 on yellowfin and skipjack tuna growth (Le Guen 

 and Sakagawa 1973; Marcille and Stequert 1976b) 

 compare these parameters from various regions 

 and studies. In this paper, calculated growth 

 curves from various studies in the Pacific Ocean 

 are compared with those obtained by the present 

 study. Such comparison requires the use of a com- 

 mon time base, in the form of apparent or esti- 

 mated age, which may be achieved by fixing one 

 common age at length for each species. In the 



absence of a reliable method for aging skipjack or 

 yellowfin tuna it is necessary to estimate likely 

 age at length by combining results obtained using 

 several techniques. 



Estimates of fork length of skipjack tuna at age 

 12 mo agree very closely. Batts (1972) and Cayre^'^ 

 used dorsal spine readings to obtain lengths of 

 40.6 and 40.7 cm for the western and eastern At- 

 lantic. Yoshida (1971), using modal progression 

 data from juveniles recovered from the stomachs of 

 billfish, and Uchiyama and Struhsaker^'' from 

 readings of sagittae, estimated lengths of 35.0 and 

 42.6 cm for the central Pacific. Finally Lewis 

 (footnote 3), also reported in Josse et al. 1979) re- 

 ported estimates from Papua New Guinea waters 

 of between 40.0 and 45.0 cm, again from sagittal 

 readings. In the present study an approximate 

 average of these quoted figures was used: 40.0 cm 

 at 12 -mo age. 



In comparing growth curves for skipjack tuna 

 from different regions, only those from the Pacific 

 have been used (Figure 11), since Josse et al. (1979) 

 showed that, due to sample variability, no signifi- 

 cant difference was detectable in growth among 

 regions, nor between growth in the eastern and 

 western Pacific as calculated from tagging studies. 

 Underestimation (Josse et al. 1979) of growth of 

 larger fish and their low Ly, (65.47 cm) may ac- 

 count for the difference in slope of the two curves 

 derived for Papua New Guinea skipjack tuna, and 

 for the difference in the two values oik ( Josse et al. 

 1979: k = 0.9451; present paper: k = 0.5148, both 

 estimated on an annual basis). 



Age at length estimates for yellowfin tuna ob- 

 tained from scale readings and modal progression 

 data show good agreement over the range of ob- 

 served values only (Suzuki 1971; Le Guen and 

 Sakagawa 1973), with the exception of the study 

 by Yabuta et al. (1960) which appears to have un- 

 derestimated growth rate. Estimates for age at 

 length have been obtained for yellowfin tuna from 

 the Atlantic from scale readings by Yang et al. 

 (1969), whose observed fork lengths averaged 66.1 

 cm at 18-mo age, and by calculation from spawn- 

 ing and recruitment data (Le Guen et al. 1969). 

 The latter study estimated fork length at 18 mo to 



'^Cayre, P 1978. Determination de I'age de listao Kat- 

 suwonus pelamis L.. debarques a Dakar. Int. Comm. Conserv. 

 Atl. Tuna, Collect. Sci. Pap., SCR 78/50. 



"'Uchiyama, J. H., and P Struhsaker. 1975. Age and 

 growth of skipjack tuna, Katsuwonus pelamis, yellow'fin tuna, 

 Thunnus albacares, and albacore, Thunnus alalunga. as indi- 

 cated by daily growth increments of sagittae. Int. Comm. Con- 

 serv. Atl. Tuna, SCRS 75/57. 



529 



