Prince et al.: Otolith analysis of Makaira nigricans age and growth 



449 



\ ISL-Jb 



Figure 4 



Photomicrographs of the video image of whole otoliths for (A) sagittae from a 5.74 mm NL Atlantic blue marlin Makaira nigri- 

 cans larvae; (B) sagittae from an 8.5 mm NL swordfish Xiphias gladius larvae; (C) sagitta (left) and lapillus (right) from an 

 8.61mm SL yellowfin tuna Thunnus albacares larvae; (D) sagittae from a larval frigate mackerel Auxis thazard. Black bars = 

 lO^m. 



otoliths, and have been noted in acetate replicas and 

 SEM preparations of many species, including other 

 oceanic pelagics (Brothers et al. 1983). 



Back-calculated spawning dates The primary back- 

 calculated spawning dates (i.e., two or more individuals 

 in a monthly interval) were from May through Novem- 

 ber (Fig. 5A). Except for the 18 larvae which were all 

 sampled in the month of August, the distribution of 

 back-calculated spawning dates did not appear to be 

 associated with the dates of capture (Fig. 5B). The 

 large variation in age of the juveniles and young 

 adults/adults (about 3-16 months) suggests that calcu- 

 lated spawning dates are not a simple reflection of the 

 distribution of capture dates of size selected fish. 



To examine the effect of various increment deposi- 

 tion rates on the distribution of back-calculated spawn- 

 ing dates, we used a contingency table analysis of 

 numbers of spawning dates by season for periodicities 

 (P) of 0.5, 1, and 2 increments per day (Fig. 6A). The 

 distribution of back-calculated spawning dates depends 

 strongly upon the periodicity assumption (x 2 76.88, 

 df 6, P< 0.0001). A deposition rate of one increment 

 per day was the only periodicity of those examined that 

 resulted in most spawning occurring in the summer and 

 fall (Fig. 6A). To examine if substantial undercounting 

 of increments occurred, periodicities ranging from 0.1 

 through 1.0 increments per day were tested (Fig. 6B). 

 Chi-square test for independence (x 2 85.778, df 24, 

 P< 0.0001) was highly significant. Only at periodicities 



