Collette and Aadland Revision of the frigate tunas (Scombridae. Auxis) 



425 



measurements are approximately equal, whereas 

 A. thazard is more tuna-shaped, having narrow en- 

 tering and departing wedges and more unequal body 

 depths. 



Morphometric data were analyzed with regression 

 analysis, analysis of covariance ( ANCOVA), and prin- 

 cipal component analysis (PCA). Regression analy- 

 sis was run on the combined morphometric data for 

 all populations of A. rochei vs. the combined popula- 

 tions of A. thazard to determine whether morpho- 

 metric differences exist between species. Regression 

 analyses were also run on morphometric data for 

 major populations of each species to look for intraspe- 

 cific differences. The following combinations of geo- 

 graphic populations were tested: for A. rochei — west- 

 ern Atlantic (n-36) vs. Mediterranean (a; =34) plus 

 eastern Atlantic (n=64), western Pacific (n=82) vs. 

 eastern Pacific (/i=37); for A. thazard — Gulf of Guinea 

 (n=20) vs. Indian Ocean (n=38), western Pacific 

 (n=75) vs. central Pacific (re=17), and westcentral 

 Pacific (??=92) vs. eastern Pacific (n=69). 



PCA was performed with the character variables 

 FL, IDor, IDend, 2Dor, and Aor to determine whether 

 these characters supported species level separation 

 of A. rochei from A. thazard and separation between 

 the eastern Pacific population and the other popula- 

 tions for both species. Very little separation was ob- 

 served with PCA because 97.9% of the variation is 

 explained by principal component 1. This means that 

 almost 98% of the variation can be explained by dif- 

 ferences in size, whereas only a little more than 2% 

 of the variation can be explained by character differ- 

 ences among the different populations. 



ANCOVA was performed on the variables IDor, 

 IDend, 2Dor, Aor, P1L, and CW with size (FL) as the 

 covariate to determine whether these characters dif- 

 ferentiated A. rochei from A. thazard and to see 

 whether eastern Pacific populations of each species 

 were different from other populations. ANCOVA 

 showed that morphometric data for each species were 

 the same for all populations, except for the eastern 

 Pacific, and could be combined. For each species, the 

 combined populations were run against the eastern 

 Pacific population to test the level of differentiation 

 of the eastern Pacific population. For A. rochei, the 

 combined populations we tested against the eastern 

 Pacific (??=30) were western Atlantic (n=36), Gulf of 

 Guinea (n=64), Mediterranean (n=34), Indian Ocean 

 (n=9), and western Pacific (n=82). For A. thazard, 

 the combined populations we tested against the east- 

 ern Pacific (n=63) were Indian Ocean (n=38), west- 

 ern Pacific (« =71), central Pacific (n = 17), and Aus- 

 tralia (n=4). 



Material examined is listed at the end of each sub- 

 species account and includes 267 specimens of A. 



thazard and 291 specimens of A. rochei. Abbrevia- 

 tions of institutions housing the material examined 

 follow Leviton et al. (1985). CMFRI is the Central 

 Marine Fisheries Research Institute, Cochin, India. 

 ZRC is the current institutional code for the Zoologi- 

 cal Reference Collection at the National University 

 of Singapore. 



Systematics 



Auxis Cuvier, 1 829 



Auxis Cuvier, 1829:199. Type-species: Scomber rochei 

 Risso, 1810, by subsequent designation of Gill, 

 1862 



Diagnosis (Modified from Collette, 1979.) Auxis 

 lacks the prominent paired fronto-parietal fenestra 

 in the top of the skull found in all the other tunas, 

 tribe Thunnini (except AUothunnus). The dorsal and 

 ventral branches of the cutaneous artery originate 

 separately from the dorsal aorta in Auxis (Godsil, 

 1954); a common cutaneous artery divides into dor- 

 sal and ventral branches laterally from the dorsal 

 aorta in the other Thunnini. The ventral branch is 

 very poorly developed, less so than in other Thunnini. 

 The dorsal cutaneous artery lies ventral to the cor- 

 responding vein in Aims; it lies dorsal in Euthynnus 

 (Godsil, 1954). The vertebral column differs from all 

 other scombrids in having a long pedicel or epihaemal 

 process (1.5-2 times the centrum depth) that forces 

 the dorsal aorta to run much further ventrally from 

 the vertebral column than in other tunas. The 

 interpelvic process is single and very long, equal to 

 or longer than the pelvic fins themselves; it is shorter 

 than the pelvic fins in all other Scombridae. 



Description Body robust, elongate, and rounded. 

 Teeth small and conical, in a single series. Two dor- 

 sal fins, the first with 10-12 spines, separated from 

 the second by a long interspace (at least equal to 

 length of first dorsal-fin base), the second fin followed 

 by 8 finlets; pectoral fins short; anal fin followed by 

 7 finlets. Gill rakers 36-49. Body naked except for 

 the corselet, which extends posteriorly along the lat- 

 eral line past the origin of the second dorsal fin. In 

 A. thazard, the extension is only 1-5 scales wide be- 

 neath the second dorsal fin, in A. rochei it is six scales 

 or more. A strong central keel is present on each side 

 of the caudal-fin base between two smaller keels. 

 Vertebrae: 20 precaudal plus 19 caudal, total 39. 

 Color: dorsum bluish, turning to deep purple or al- 

 most black on the head; belly white, usually without 

 stripes or spots, occasionally black spots present on 



