WANKOWSKI: ESTIMATED GROWTH OF SURFACE-SCHOOLING TUNAS 



and Ly, — 180.9 cm for yellowfin tuna {k in both 

 cases is estimated on a monthly basis). The calcu- 

 lated length-at-age curves are shown in Figures 9 

 and 10, which also include modal progression data 



speaking, a true correlation coefficient. The mean 

 square error was found to be 0.74 for the estimate 

 of skipjack tuna growth and 2.01 for that of yel- 

 lowfin tuna growth. 



Figure 9.— Plot of length at age for 

 skipjack tuna derived from all modal 

 progressions for all Papua New Guinea 

 areas combined, fitting the von Ber- 

 talanffy function by Fabens' (1965) pro- 

 cedure. Original modal data are shown 

 fitted to the curve. The time scale is in 

 elapsed months and does not indicate 

 apparent age. Dashed line shows ex- 

 trapolation beyond the range of obser- 

 vations. 



70-1 



E 

 u 



O50- 



z 



Ul 



cc 

 o 



30- 



20 



40 



MONTHS 



60 



fitted to the curve. Tables 2 and 3 list modal pro- 

 gressions used in the estimations. The mean square 

 error was calculated from the formula: 



'/-. 



N 



where A/^ is the number of pairs of observations and 

 r a coefficient which is minimized in the curve fit. 



^ 



(L - Li) 



where L is the length predicted by the particular 

 curve and Li the observed length value. Since age 

 at length data are unknown, r is not, strictly 



80 



Figure lO. — Plot of length at age for yellowfin tuna derived from 

 all modal progressions for all Papua New Guinea areas com- 

 bined, fitting the von Bertalanffy function by Fabens' (1965) 

 procedure. Original modal data are shown fitted to the curve. The 

 time scale is in elapsed months and does not indicate apparent 

 age. Dashed line shows extrapolation beyond the range of obser- 

 vations. 



E 



I 

 t- 

 O 



z 



UJ 



oc 

 O 



40- 



I 

 I 

 I 



I 

 I 



1 1 1— 



10 20 30 



MONTHS 



527 



