(2) Fork length: tip of bill to tips of mid-caudal 

 rays. 



(3) Body length; tip of lower jaw (with jaws 

 closed) to tips of mid-caudal rays. 



(4) Eye-fork length: posterior margin of eye to 

 tips of mid-caudal rays. 



(5) Caudal spread: dorsal tip to ventral tip of 

 lobes of caudal fm. 



(6) Girth: twice the curved distance along one 

 side of the body from the pelvic groove to the 

 dorsal edge of the dorsal groove. 



METHODS OF ANALYSIS 



Three equations were used in the study. The rela- 

 tion between logio (weight) and logio (eye-fork 

 length) is given by 



Y = A + BiXi (1) 



where 



Y = logio (weight), 



A = intercept, 

 Bi = coefficient, 

 Xi = logio (eye-fork length). 



The equation can be transformed to the familiar form 



weight =A' (eye-fork length) ' 



where 



A' = 10^ 



by taking antilogs of both sides of (1). The relation 

 between logio (weight), logio (eye-fork length), and 

 logio (girth) is given by 



Y =A + BiXi + B2X2 



(2) 



where 



Y = logio (weight), 



A = intercept, 



BiandB2 = coefficients, 



Xi = logio (eye-fork length), 



X2 = logio (girth). 



The equation can be transformed to 



weight = A' (eye-fork length)^' (girth) ^^ 



by taking the antilogs of both sides. The relations 

 between eye-fork length and other measures of 

 length are given by 



Y = A +BiXi (3) 



where 



Y = eye-fork length, 



A = intercept, 



Bi = coefficient, 



Xi = other measure of length. 



Equation ( 1 ) was not used for the relation between 

 the various measures of length because estimates of 

 B were very close to 1, indicating that linear rela- 

 tions among the variables were appropriate. Equa- 

 tion (3) was used instead. 



The parameters of (1). (2), and (3) were estimated 

 by use of linear regressions. Analysis of covariance 

 was used to examine sexual differences. Mul- 

 tivariate analysis was used to determine if white 

 marlin could be sexed or allocated to either Florida 

 or Louisiana given measures of length and weight. 



RESULTS AND DISCUSSION 



Estimates of the parameters of (1), (2), and (3) are 

 shown in Table 1 . All estimates of the parameters are 

 significantly different from at the 0.01 level of 

 significance. 



Analyses of covariance revealed no significant 

 differences between sexes in the relations between 

 weight and eye-fork length, between eye-fork 

 length and the three other measures of length, and 

 between eye-fork length and caudal spread for blue 

 marlin. However, sexual differences were found in 

 the relations between weight and eye-fork length 

 and between eye-fork length and caudal spread for 

 white marlin (Fig. 1 and 2). Female white marlin 

 tend to weigh more at a given length than male 

 white marlin, but this difference tends to disappear 

 at larger sizes. Further examination of the data in- 

 dicates that the difference is partially the result of 

 females tending to have deeper bodies than males. 

 Male white marlin tend to have a wider caudal 

 spread than females and the difference tends to in- 

 crease with size. A sexual difference in caudal 

 spread was also found for sailfish (Fig. 3), but the 

 difference decreases with increased size. Sexual 

 differences were not found in the length-weight re- 

 lation for sailfish. 



Deviations from the length- weight relation of the 



122 



