WANKOWSKI: ESTIMATED GROWTH OF SURFACE-SCHOOLING TUNAS 



80 



70 



E 

 3 60 



X 

 H- 

 O 



z 

 liJ 



cr 

 O 50 



40 



J J A 

 1977 



S O N D J 



M A M J J A 

 1978 



S O N D J F 



M A M J J A 

 1979 



S O N D 



Figure 4. — Modal fork lengths of samples of yellowfin tuna from all Papua New Guinea areas as a fimction of time (in months). Modal 



progressions are indicated by lines. See Figure 3 for explanation of symbols. 



suits of tagging studies in Papua New Guinea 

 (Lewis 1980a, b, see footnote 8) indicate that skip- 

 jack tuna recruitment may be intermittent, and 

 possibly dependent on periodic influxes from north 

 and east of the area. 



It has been suggested (Lewis 1980a, b, footnote 

 3; Anonymous^) that there are at least three par- 

 tially mixing components to the skipjack tuna 

 population in the western Pacific: one ranging 

 from Japanese waters to the Equator, one centered 

 on the Bismarck Sea and ranging to lat. 10° N, and 

 the last extending south from the Bismarck Sea; 

 and that these components are composed of sepa- 

 rate spawning units, distinguishable on the basis 



*Lewis, A. D. 1977. Tuna tagging in Papua New Guinea. 

 Harvest, Papua New Guinea 4:13-17. 



^Anonymous. 1976. Ad -hoc meeting of scientists to discuss 

 skipjack fisheries developments and research requirements. 

 Rep. meet., South Pac. Comm., Noumea, New Caledonia, 6-10 

 Dec. 1976, 27 p. 



of spawning periodicity with respect to northern 

 and southern summers. The following analysis of 

 data from the present study provides further evi- 

 dence supporting this view. 



Josse et al.'s (1979) estimates of the von Ber- 

 talanffy parameters {k = 0.94512 on an annual 

 basis and L-^ = 65.47 cm) for tagged skipjack tuna 

 in the Papua New Guinea region were applied to 

 the length-frequency modal data to estimate the 

 dates of birth from all modal lengths available for 

 skipjack tuna samples from the four areas. Josse et 

 al.'s estimate is independent of the modal data, 

 being derived from an earlier study using a differ- 

 ent technique, and for this reason it is preferable 

 (for present purposes) to the estimate derived in 

 this report. 



The frequency distribution of dates of birth is 

 shown in Figure 6. The extremes of the range 

 (pre-1975 and post-1977) are poorly represented; 

 however, a 6-mo periodicity is indicated. Although 



523 



