nificant difference and concluded that plankton nets 

 were reliable tools for sampling larval tunas within 

 the limitations of the method. All of his tows were 

 taken at the surface and at night. 



To determine if catches could be duplicated during 

 other hours and with oblique tows, the data from the 

 39 pairs of plankton tows taken during CHG-35 were 

 examined. The abundance of larval tunas calcu- 

 lated from catches of the first tow did not differ sig- 

 nificantly (Wilcoxon matched-pairs signed-ranks 

 test) from that of the second. We thus concluded 

 that catches by oblique tows taken during day or 

 night were duplicative. 



SPECIES COMPOSITION 



Larvae of the following species of tunas were 

 identified in the Marquesan samples: skipjack, yel- 

 lowfin, bigeye, albacore, little tunny {Euthijnnus 

 affinis), and frigate mackerel {Auxis sp.)^. Adults 

 of all except .4 uxis have been caught either by long- 

 line, trolling, or pole-and-line fishing in the Mar- 

 quesas (King et al., 1957; Austin, 1957; Wilson and 

 Rinkel, 1957; Wilson et al., 1958; Yoshida, 1900). 

 Another species, the dogtooth tuna [Gymnosarda 

 nuda), also has been caught in the Marquesas by 

 trolling near the islands, but its larva has not been 

 identified. 



The species composition of the larval tunas is 

 shown in table 3. Skipjack was the dominant spe- 

 cies throughout the offshore surveys and at all but 

 one of the diel variability stations. At the diel 

 variability station occupied during HMS-43, yellow- 

 fin was dominant (appendix table A-2). Other 

 species were found in sporadic abundance, e.g., Auxis 

 at the second diel variability station of CHG-35 

 (appendix table A-1), bigeye at the diel variability 



T.\BLE .3. — Species composition of larval tunas collected in 

 Marquesan waters, 1957-58 



^ May include both A. thazard and .1. rochei (=-4. thynnoides); see 

 Matsumoto (1959). 



station of H;MS-43 (appendix table A-2) and on the 

 oft'shore surveys of CHG-38 (appendix table A-7) 

 and HMS-45 (appendix table A-8). 



DIEL AND SEASONAL DISTRIBUTION 



Diel variation in catches of larval tunas has been 

 discussed by Wade (1951), Matsumoto (1958), and 

 Strasburg (1960). All reported greater catches at 

 night. The latter two authors attributed this pri- 

 marily to vertical migration by the larvae into the 

 upper surface layers of the ocean at night, although 

 they did not rule out the possibility of net dodging 

 during daylight. 



Since Strasburg (1960) found practically no larval 

 tunas below 140 m., our sampling did not extend 

 below this depth (except at the diel variability sta- 

 tion on HMS-45). By sampling the 0- to 140-m. 

 depth range, we hoped that variations due to diel 

 vertical migration would be kept to a minimum or 

 even possibly eliminated. 



Catches at the diel variability station during 

 CHG-35, HMS-43, and CHG-38 provided evidence 

 that this variation was not eliminated. Prominent 

 diel variations occurred in December and March for 

 skipjack and in January for yellowfin (fig. 2). 

 Either the larvae did occur below 140 m., or they 

 were more successful in escaping the net during the 

 day in these instances. The problem of diel varia- 

 tion was complicated further by the inconsistency of 

 the catches during the other months. For example, 

 the highest catches of larval skipjack were obtained 

 during or near twilight in April and June, while dur- 

 ing October and January day catches were as good 

 as night catches. 



Average larval abundance during the offshore sur- 

 veys of each cruise was compared with those of the 

 other cruises for seasonal variation (fig. 3). The 

 abundance during HMS-43 (January to February) 

 and during CHG-38 (!\Iarch to April) was signifi- 

 cantly greater (Mann-Whitney U test, p<0.05) 

 on both occasions than for either HMS-45 (May) or 

 CHG-35 (October to November). No significant 

 differences were found in other comparisons of off- 

 shore averages. 



Averages for the diel variability station, computed 

 from results of tows taken between 2000 and 0400 

 hours, the same hours during which tows were taken 

 on the offshore surveys, also are showai in figure 3. 

 The magnitudes and variations of these averages 

 differ considerably from those of the offshore surveys. 



LARVAL TUNAS IN MARQUESAN WATERS 



