Ralston et al Accuracy of age estimates for Sebastes jordani 



95 



35 



30 

 25 

 20 



15 

 10 



5 





 30 

 25 

 ? 20- 



03 



o) 15 



< 



10 



5 





 30 

 25 

 20 

 15 

 10 



5 







Reader 2 vs. reader 1 (n=52) 

 f- 1.11, P- 0.27 



Reader 3 vs. reader 1 (n-20) 

 ( = 0.00, P = 1.00 



Reader 3 vs. reader 2 (n-1 5) 

 t - -1.31, P- 0.21 



5 10 15 20 25 30 35 

 Age (d) 



Figure 7 



Comparison of age readings of larval Sebastes 

 jordani among three different age readers. Di- 

 agonal lines represent lines of equality (some 

 points superimposed l. 



et al. (1991), who first described an extrusion check 

 in the otoliths of S. jordani. A distinct "birth mark" 

 has also been observed to form in the otoliths of the 

 viviparous surfperch Micrometrus minimus at the 

 time of parturition (Schultz, 1990). 



In certain situations otolith size at hatch is known 

 to depend on temperature (e.g. Campana, 1989) and 

 in our study area temperature varies substantially 

 among years (Ralston, 1995; Ralston and Howard, 

 1995). Because the planktonic larvae we used were 

 collected in 1991, whereas the gestating larvae were 

 collected in 1993, it is important to consider the ef- 

 fect this might have on otolith size at extrusion and 

 the conclusions we have drawn. Results presented 

 in Laidig and Ralston ( 1995) bear specifically on this 



issue; i.e. they found no interannual differences in 

 otolith radius at the extrusion check for six species 

 of Sebastes (including S. jordani) that were sampled 

 over a six-year period (1984-89). However, inter- 

 annual differences in the width of the first postex- 

 trusion increment were observed. 



There are, moreover, interspecific differences in 

 extrusion check radius, and these can aid in identi- 

 fying rockfish species (Laidig and Ralston, 1995). In 

 this regard, S. jordani has a particularly large 

 preextrusion radius and can be distinguished 68% 

 of the time on this basis alone. We speculate that the 

 Sebastes extrusion check may owe its origin to an 

 osmotic or ionic shock that occurs to the larvae when 

 they first contact seawater and that temporarily al- 

 ters calcium metabolism. 



It is noteworthy that many otoliths exhibited in- 

 cremental growth during gestation (see also Schultz, 

 1990). We observed these preextrusion increments 

 in both gestating and planktonic larvae. Although 

 the expression of these structures was quite variable, 

 their existence highlights the importance of accu- 

 rately locating the extrusion check. We speculate that 

 the physiological basis of preextrusion increments 

 in these larvae is linked to the transmittal of the 

 diel maternal calcium cycle into the ovarian embry- 

 onic environment (Mugiya et al., 1981). 



The existence of a clear, well-defined mark that is 

 formed at parturition satisfies an important require- 

 ment for accurate larval age estimates. An unam- 

 biguous starting point for counting increments is a 

 necessary, but not sufficient, condition for obtaining 

 reliable data. Increment counts must also be precise. 

 For S. jordani this requirement was met, given the 

 high percent agreement among age readers (93% 

 within ±2 d) and the lack of systematic reader differ- 

 ences in ageing. High precision was also reflected in 

 the relatively high confidence that readers assigned 

 to their work, stemming from the consistent clarity 

 of increment microstructure in the otoliths. Fish 1- 

 3 d old were the most difficult to interpret. This was 

 probably due 1 ) to optical interference from light dif- 

 fraction around the otolith margin, 2) to the fact 

 that daily increments present at those ages were the 

 narrowest in width, and 3) to the fact that fewer in- 

 crements hampered visual recognition of a growth 

 pattern. 



A serious potential source of bias in ageing larval 

 fish is the limited resolution of optical microscopes 

 ( Neilson, 1992 ). Studies have shown that optical sys- 

 tems are sometimes incapable of resolving fine in- 

 crement microstructure (Campana et al., 1987; Jortes 

 and Brothers, 1987; Morales-Nin and Ralston, 1990; 

 Secor and Dean, 1992). However, our results for S. 

 jordani show that measurements of increment width 



