168 



Fishery Bulletin 93(1), 1995 



Results 



Seven specific nuclear shading characters were iden- 

 tified (Table 1): 1) the opacity of the primordium; 2) 

 the opacity of the markings inside the nucleus; 3) 

 the opacity of the increments directly outside the 

 nucleus; 4) the opacity of the nuclear edge; 5) the 

 existence of a dark inner band near the nuclear edge; 

 6) the existence of a light inner band near the nuclear 

 edge; and 7) the existence of a dark inner ring encir- 

 cling the primordium. 



In some cases, combinations of the shading pat- 

 terns were sufficient to characterize a species. For 

 example, 84% of the otoliths of S. goodei were found 

 to have a dark primordium, a dark nuclear edge, and 

 a dark inner ring surrounding the primordium, while 

 this combination was never found in the other spe- 

 cies examined (Fig. 1). Likewise, in S.jordani, a light 

 penumbra was regularly found (76% occurrence) 

 adjacent to the inner edge of the nucleus, along with 

 a dark primordium and many inner rings (faint mi- 

 crostructure occurring inside the nucleus). In S. 

 paucispinis and S. flavidus, there was usually a dark 

 inner band next to the edge of the nucleus (87% and 

 73% occurrence, respectively) . No annual variation 

 was observed for the nuclear shading patterns. How- 

 ever, in the remaining species, the shading patterns 

 were too variable to establish consistent identifiable 

 character states that would distinguish species. 



Annual variations in nuclear radius and the width 

 of the first increment were examined (Fig. 2). An- 

 nual variation in nuclear radius among the species 

 was not significant; however, annual variability in 

 the width of the first increment was significant 

 (P<0.05). In addition, there was a significant inter- 

 action (P<0.05) between year and species. The mean 

 width of the first increment of S. paucispinis, for 



example, declined from 1984 to 1989, whereas that 

 of S. flavidus increased (Fig. 2A). 



Sebastes jordani was found to have the largest 

 average nuclear radius (Table 2; Fig. 2B). The rank 

 order for the remaining species, from largest to small- 

 est average nuclear radii, was S. goodei, S. 

 auriculatus, S. paucispinis, S. flavidus, S. entomelas, 

 S. saxicola, andS. mystinus. Individually, the nuclear 

 radii of S. jordani, S. goodei, S. auriculatus, and S. 

 mystinus were significantly different (P<0.05) from 

 all other species and from each other. 



Sebastes jordani was also found to have the larg- 

 est average first increment of all species studied 

 (Table 2; Fig. 2A). The rank order of the other spe- 

 cies, from largest to smallest average widths of the 

 first increment, was S. paucispinis, S. goodei, S. 

 auriculatus, S. saxicola, S. flavidus, S. entomelas, 

 and S. mystinus. Sebastes jordani and S. paucispinis 

 had significantly larger average first increment 

 widths (P<0.05) than all other species studied (0.97 

 //m and 0.91 /im, respectively) (Table 2). 



A discriminant analysis was performed on the clas- 

 sification of species by using nuclear radius and the 

 width of the first increment as predictor variables 

 (Table 3). Sebastes goodei, S. jordani, S. mystinus, 

 and S. paucispinis were correctly identified from 57 

 to 83% of the time; Sebastes auriculatus and S. 

 flavidus were classified correctly 33.9% and 41.8% 

 of the time, respectively. Although these values are 

 less than 50% correct, they represent the largest 

 single classification for each species. Two species, S. 

 entomelas and S. saxicola, were not often classified 

 correctly (9.5% and 0%, respectively). 



In a blind test, the reader correctly classified 70 of 

 the 100 otoliths (70% correct; Table 4), demonstrating 

 that otoliths provide useful information in species iden- 

 tification (P<0.001). Greater than 90% of Sebastes 



