FISHERY BULLETIN: VOL. 83, NO. 3 



variation in area-to-area differences in parasite 

 numbers. A plot of the first two, accounting for 58% 

 of the variation, showed the Atlantic fish (0) to be 

 distinct from all the Pacific ones, even though only 

 three fish from the Atlantic were dissected (Fig. 3). 

 However, fish from California (M) and Ecuador (N) 

 fell close to the western Pacific samples. They were 

 separated out on the third axis (not shown), but 

 nevertheless it is evident that they had a somewhat 

 similar parasite fauna. The fish from western New 

 Zealand (J) appeared distinct, and so too, to a less ex- 

 tent, were the Papua New Guinea samples (C and D). 

 There is no suggestion that fish from Ponape (A), 

 Palau (B), Solomon Islands (E), Fiji (G), and the Mar- 

 quesas (L) had distinct faunas of these long-lived 

 parasites. 



H- 



Figure 4. -Results of multivariate analyses using 7 "temporary'" 

 parasites (Nos. 1, 6, 15, 17, 18, 21, and 22). 95% confidence rings 

 given for samples of more than 24 fish. 



In this analysis, Anisakis II had the most powerful 

 discriminating properties, though at least 7 of the 10 

 parasites used were capable of substantial discrimi- 

 nation in their own right. 



An analysis based on the 7 "temporary" parasites 

 (Nos. 1, 6, 15, 17, 18, 21, and 22) produced a much 

 greater separation of areas (Fig. 4). They are 

 grouped into two broad classes: one containing New 

 South Wales (I), New Zealand (J, K), and the eastern 

 Pacific (M, N); and the other the western tropical 

 areas. Each area in the latter group had a temporary 

 parasite fauna that was distinct from most other 

 areas. Over 83% of the variation was accounted for 

 by the first two axes, and 90% by the first three. It is 

 interesting to note that New South Wales (I) is more 

 similar to east New Zealand (K) than to west New 

 Zealand (J) (this was much more marked on the third 

 axis, not shown, where I and K were pulled to one 

 side), and that west New Zealand is similar to 

 California (M) and Ecuador (N). 



Taken together. Figures 3 and 4 indicate that 

 several distinct skipjack tuna parasite faunas existed 

 within the tropical Pacific, and the longer lived 

 parasites were more evenly distributed than the 

 shorter lived ones. 



To check these results and to look for links be- 

 tween the New Zealand fish and the tropical areas, 

 the west Pacific data were reanalyzed using first the 

 10 "semipermanent" parasites (the didymozoids) and 

 second the 4 "permanent" parasites (anisakids and T. 

 coryphaenae). 



The average numbers of didymozoids in the New 

 Zealand fish were almost identical to the overall 

 average for the central and western tropics (Table 

 4). In the multivariate analyses, the temperate water 

 samples fell to one side of the tropical samples (Fig. 5 



Figure 5. -Results of multivariate analysis using 10 

 didymozoids only (Nos. 2, 3, 4, 5, 11, 13, 14, IK, 19, and 20). 

 95% confidence rings given for samples of more than 24 fish. 



350 



