overall seasonal increase in size of the testes, and 

 some milt is usually present throughout the year. 



On the other hand, the ovary as an indicator of 

 maturity has been well documented (Clark, 1925, 

 1929, 1934; Hickling and Rutenberg, 1936). As 

 oogenesis proceeds, characteristic changes occur 

 which can be easily detected macroscopically or 

 microscopically. We therefore chose to use ovarian 

 characteristics to represent maturity of billfishes in 

 this study. 



The most precise method of determining the stage 

 of ovarian maturity is to histologically examine the 

 tissues as performed by Merrett (1970) or Moser 

 (1967). This procedure, however, is lengthy and 

 time-consuming. Another reliable technique is to 

 measure a large number of ova from the same ovary, 

 a method used for many species (Clark, 1929, 1934; 

 June, 1953; and Brock, 1954). This method is based 

 on the assumption that as the spawning season pro- 

 gresses, the group or groups of maturing ova will be 

 distinguished as advancing modes in size-frequency 

 distributions. This method is also time-consuming 

 and laborious, but has a definite advantage in charac- 

 terizing the frequency of spawning when a fully ma- 

 ture specimen is examined. When many fish are 

 examined over a time interval, the progression of the 

 modes of developing ova may provide information 

 on the rate of maturation, time of spawning, and size 

 at maturity. Two variations of this process which 

 require the measurement of fewer ova are the use of 

 "'maximum ova diameter" (Otsu and Uchida, 1959) 

 and the position of the 95th centile (Schaefer and 

 Orange, 1956). The latter is particularly useful when 

 the exact position of the developing mode is difficult 

 to distinguish, as in early maturation stages. 



Indirect methods to measure sexual maturity in- 

 volve the relationship between some measure of the 

 fish's size (either length or weight) and gonad 

 weight. The use of fish length assumes that fish 

 weight is nearly proportional to the cube of the 

 length, a true situation with regard to the billfishes in 

 this study as determined by length-weight analyses 

 for eastern Pacific billfish. It is also assumed that 

 fecundity is proportional to size. Kume and Joseph 

 (1969b) have plotted ovary weight versus eye-fork 

 length and also utilized the gonad index (GI) 

 computed as 



GI = (W/L 3 ) ■ 10 4 



where 



W = total weight of gonads in grams, and 

 L = eye-fork length in cm. 



Table 2. — Regression of maximum ova diameter and 95th 

 centile of ova diameter on gonad index (n = sample size, r 

 = coefficient of correlation, b = slope, a = v axis inter- 

 cept.) 





n 



r 



b 



a 



Striped Martin 











95th centile on GI 



3! 



0.936* 



3.02 



1.46 



Max. ova diameter on GI 



269 



0.797* 



3.78 



1.48 



Sailfish 











95th centile on GI 



21 



0.913* 



3.91 



2.47 



Max. ova diameter on GI 



184 



0.859* 



4.78 



3.43 



*Significant at 0.01 level. 



Merrett (1971) used another type of gonad matura- 

 tion index which related the macroscopic appear- 

 ance (color, yolk presence, egg diameter, and gen- 

 eral appearance) of the gonad to recognizable stages 

 in its histology. 



To evaluate these different measures of maturity 

 and to determine the degree of correlation between 

 them, we applied regression analyses to our data 

 (Table 2). As can be seen, the gonad index is highly 

 correlated. In each of the four regressions, the corre- 

 lation coefficients exceeded the 0.01 significance 

 levels when tested against a Student's r-distribution. 

 The lower r values for regression of maximum ova 

 diameter on gonad index can be explained by the fact 

 that maximum egg diameters do not always repre- 

 sent the size of the advanced mode. For example, 

 the presence of a few residual eggs in an ovary which 

 is in the resting or early maturation stages will not 

 reflect the true stage of development of the ovary. 



We have included both direct and indirect 

 methods to analyze the spawning of striped marlin 

 and sailfish. But, based on the above comparison 

 and considering the time and manpower costs and 

 the degree of accuracy desired, we conclude that the 

 gonad index represents the most practical indicator 

 of the stage of sexual maturity for a study of this 

 type. 



Size at First Spawning 



The reported size at which striped marlin attain 

 sexual maturity varies little among previous studies. 

 Merrett (1971) reported first maturity at 140-160 cm 

 eye-fork length. This agrees with the conclusion of 

 Williams (1963). Kume and Joseph (1969b) stated 

 that individuals greater than 160 cm from the eastern 

 Pacific regularly occur in the spawning group (3.0 

 GI), however, they did collect a mature specimen in 

 the 148-cm class. 



93 



