tion data for 1960 only. My analysis of their data 

 shows a significant difference for June and July 

 but no significant difference for August and Sep- 

 tember. Their findings for June, August, and Sep- 

 tember agree with those in this study. The ex- 

 treme difference in se.x ratios found in the present 

 study for the gulf (May through October) and for 

 the Florida Straits in April has not been reported 

 previously. The above findings suggest that some 

 white marlin segregate into distinct areal groups 

 according to the predominating sex and that sex 

 ratios may change with season. A similar occur- 

 rence has been noted for the blue marlin, Makuira 

 nigricans (Kume and Joseph 1969). 



Length- Weight Relationship 



The average length of females is greater than 

 that of males from both the gulf and the Atlantic 

 (Figure 1). It is also apparent (Figures 2, 3) that 

 the average length of females is greater than that 

 of males from each area for each month studied. 

 This difference may be due to faster growth of 

 females or higher mortality of males and should be 

 considered in future growth studies of the white 

 marlin. 



The length- weight relationship by sex was de- 

 termined for white marlin taken in the gulf from 

 1971 through 1976 (Figure 4) and in the Atlantic 

 from 1972 through 1976 (Figure 5). Analysis of 

 covariance (Table 2) indicated that length-weight 

 regression coefficients were significantly different 

 between gulf females and males (F = 16.0; df = 1, 

 900;P<0.001), gulf males and Atlantic males (F = 

 19.2; df= 1, 514; P<0.001), and gulf females and 

 Atlantic females (F = 10.8; df = 1, 871;P<0.001). 

 The adjusted means were also significantly differ- 

 ent between Atlantic females and males (F = 13.4; 

 df = 1, 486; P<0.001). These findings agree with 

 those of Lenarz and Nakamura (1974), who found 

 a significant difference between sexes in the rela- 

 tionship between weight and eye-fork length for 

 white marlin from the gulf during 1971. 



Analysis of covariance was conducted for the 

 length-weight relationship, on a monthly basis, 

 for which sufficient samples were available: gulf 

 females versus Atlantic females in May, June, 

 August, and September, and gulf males versus 

 Atlantic males for June, August, and September. 

 A significant difference in the regression 

 coefficients was found only for the August males (F 

 = 13.7; df= 1,211;P<0.001). A significant differ- 

 ence in adjusted means was found for females dur- 



215- 



210- 



205- 



200- 



E 195- 



u 



190- 



o 



185- 



180- 



5 175- 



o 

 it. 



I 170- 



< 165- 



I 155J 



150- 



145- 

 140- 

 135- 



■r 



? 



N=277 

 31% 



N=248 

 51% 



0*' 



N = 627 

 69% 



N=241 



49% 



Figure 1. — Comparison of length between female and male 

 white marlin collected in the northern Gulf of Mexico (1971 

 through 1976) and in the Atlantic (1972 through 1976). The 

 number, percent, mean (horizontal line), range (vertical line), 1 

 SD on each side of the mean ( open box), and 2 SE on each side of 

 the mean (shaded box) are shown. 



ing June (F = 9.3; df = 1, 74;P<0.005) and August 

 (F = 12.0 df = 1, 286;P<0.001), and for males 

 during September (P = 7.6; df = 1, 73; P<0.01). 

 Differences between length-weight relation- 

 ships of white marlin from the gulf and the Atlan- 

 tic suggest the possibility of separate groups in- 

 habiting the two areas. Tag returns, however, 

 showed there is at least some migratory move- 

 ment from the Middle Atlantic Bight to the gulf. 

 To date, tag return data have not shown white 

 marlin migrations in the reverse direction, al- 

 though one fish tagged in the gulf was recaptured 

 off Cuba, giving some support to the likelihood 

 that they do migrate in the opposite direction 

 (Chester C. Buchanan, Southeast Fisheries 

 Center, National Marine Fisheries Service, 

 NOAA, 75 Virginia Beach Drive, Miami, FL 

 33149, pers. commun.) 



921 



