LESTKKKTAL.: I'AKASI TKSOK SKIIMACK Tl'NA 



- H, I, J, K), possibly because of the three didymo- 

 zoids suspected of being relatively short-lived (Nos. 

 16, 19, and 20). The east and west New Zealand sam- 

 ples (J, K) were identical on the first two axes, and 

 separated only slightly on the third axis (not shown). 

 There was no obvious link between New Zealand and 

 any particular tropical area. 



Similarly, the larval nematodes and T. cory- 

 phaenae (Nos. 8, 9, 10, and 12) did not suggest a link 

 between New Zealand fish and those from any par- 

 ticular tropical area (Fig. 6). However, west New 

 Zealand (J) now appeared distinct from east New 

 Zealand (K) and New South Wales (I). The separa- 

 tion was due to areas having either high Anisakis I 

 and II and low Terranova and T. coT^phaenae or low 

 Anisakis I and II and high Terranova and T. cory- 

 phaenae. West New Zealand (J) was at one extreme 

 (high Anisakis) and the three most northwestern 

 areas- Ponape (B), Fiji (G), and Marquesas (L)-at 

 the other. Tentacularia coryphaenae and probably 

 Terranova were picked up in the tropics. It seems 

 likely that one or both of the Anisakis larvae were 

 picked up predominantly in temperate waters, par- 

 ticularly in west New Zealand. This may explain the 

 separation of west New Zealand from the other areas 

 in Figure 4. 



In summaiy, the New Zealand fish were not close- 

 ly aligned with any particular tropical sample, and 

 the eastern and western New Zealand fish were 

 probably carrying similar parasite faunas when they 

 arrived in New Zealand. 



Tentacularia coryphaenae 



Data on this parasite are presented in detail 

 because we had more than for any other parasite and 

 because potentially it was our most valuable marker. 

 It also was the subject of many queries from skipjack 

 tuna processors. The parasite was common through- 

 out the south, central, and west Pacific (Table 3, 

 parasite No. 12). The means of samples of over 22 

 fish within the length range 44 to 53.9 cm suggested 

 an east-west cline across the Pacific, with twice as 

 many parasites being found in fish from around the 

 Marquesas (L) as around Papua New Guinea (C and 

 D) (Fig. 7). A regression analysis of number of para- 

 sites against longitude using tropical data on the 

 number of parasites in 972 fish, transformed and ad- 

 justed for differences in host length (data collected 

 independently by the SPC), showed that the relation- 

 ship was statistically significant, though it only ac- 

 counted for about 7% of the fish-to-fish variation. 



Considering fish of all sizes, the number of T. cory- 

 phaenae in the tropics increased with the size of the 



Figure 6. - Results of multivariate analysis using the four 

 "permanent" parasites (anisakids and T. coryphaenae, Nos. 

 8, 9, 10, and 12). 95% confidence rings given for samples of 

 more than 24 fish. 



fish (Fig. 2, solid circles). The increase around 47 cm 

 is due to many of the Marquesas fish being this size 

 and Marquesas fish tended to have more T. cory- 

 phaenae. In New Zealand, smaller fish had about the 

 same average number as fish from the tropics. How- 

 ever, this number did not increase with size (Fig. 2, 

 open circles). Thus, the 58 -t- New Zealand fish had 

 fewer parasites than their peers in the tropics, and 

 about the same number as the 45 to 50 cm fish. 



School-to-School Variation 



An analysis of variance, and a median test, were 

 carried out on 30 schools from the Marquesas and 19 

 schools from eastern New Zealand (areas L and K, 

 respectively. Table 5). The results of the two 

 methods on each data set show close agreement. 



In the Marquesas, five parasites showed strong 

 evidence of association with particular schools, i.e., 

 the probability that schools differed was at least 0.95 

 with both methods. The parasites were Caligtis spp. 

 (No. 1), D. euthynni(bio. 15), H. ventricosa{No. 17), 

 D. intestinomuscularis (No. 19), and Lagenocystisl 

 Univitellannulocystis spp. (No. 20). For these para- 

 sites to show significant differences, they must have 

 heavily infected some schools and not others, and 

 their life span in the fish must have been equal to or 

 shorter than the life of the school. The literature 

 review suggested that the first three species could 

 possibly be readily lost from fish, and this is vin- 

 dicated by their strong school association. The evi- 

 dent impermanence of the last two, however, was 

 unexpected. It was as a consequence of this finding 

 that they were not included in the analysis for Figure 

 3. 



351 



