Schaefer Geographic variation in morphology and menstics of Thunnus albacares 



291 



Acronym Morphometry Character 



TL Total length 



SFD Snout to insertion of first dorsal fin 



SSD Snout to insertion of second dorsal fin 



SAF Snout to insertion of anal fin 



HL Head length 



FDSD Insertion of first dorsal fin to insertion of second dorsal fin 



FDAF Insertion of first dorsal fin to insertion of anal fin 



SDAF Insertion of second dorsal fin to insertion of anal fin 



FDPF Insertion of first dorsal fin to insertion of pelvic fin 



SDPF Insertion of second dorsal fin to insertion of pelvic fin 



PFL Pectoral fin length 



SOFL Second dorsal fin length 



AFL Anal fin length 



Figure 2 



Morphometric characters measured and acronyms used for analysis of geographic variation of Pacific 

 yellowfin tuna. 



common within-group regression are used to allomet- 

 rically adjust variates when between-group heterogen- 

 eity exists (Thorpe 1976, Reist 1985 and 1986). The 

 measurements of the morphometric characters were 

 adjusted to those expected for the overall mean total 

 length with a modification of the allometric formula 

 given by Thorpe (1975): 



Y; = log 10 Yi - [/?(logi Xi - log 10 X)] 



where 

 Yj = adjusted logarithmic character measurement of 



the ith specimen, 

 Yj = unadjusted character measurement of the ith 



specimen, 

 ft = common within-group regression coefficient of 



log 10 Y against log 10 X, 

 X; = total length of the ith specimen, and 

 X = overall mean total length. 



Reist (1985) has shown that this allometric adjust- 

 ment effectively removes size variation from the data 

 he examined. This statistical approach used to remove 

 size effects from morphometric data has been shown 

 to be an appropriate procedure for objective analysis 

 of the data when there is size overlap among the groups 

 examined (Claytor and MacCrimmon 1986). 



I did not adjust gill-raker counts because Spearman's 

 rank correlation procedure indicated that there were 

 no significant correlations between gill-raker counts 

 and total lengths. The Kruskal-Wallis test and a non- 

 parametric multiple comparison test (Zar 1974) were 

 utilized to test for differences among gill-raker counts 

 from the five areas. 



I used canonical variate analysis to examine the size- 

 adjusted morphometric data for yellowfin tuna from 

 five locations in the Pacific Ocean. This technique, also 

 known as multiple discriminant function analysis (Pie- 

 lou 1977), is appropriate when separation of more than 



