228 



Fishery Bulletin 97(2), 1999 



geneios, hereafter called hapuku. However, our re- 

 sults are also relevant to other populations of P. 

 oxygeneios and, because both species of Polypriou 

 have similar otolith structure, our results should also 

 assist in the interpretation of otoliths of P. ameri- 

 canits. We used thin otolith sections to age hapuku 

 and attempted to validate ages using oxytetracycline 

 (as a time marker) and otolith marginal state analy- 

 sis. We also obtained independent growth rate esti- 

 mates from length-frequency and tag-recapture data 

 for comparison with growth rate estimates derived 

 from length-at-age data. 



Materials and methods 



Otolith ageing 



Sagittal otoliths were obtained from 1400 hapuku 

 collected from waters off New Zealand, from 1979 to 

 1995. Many otoliths were collected during studies by 

 Johnston ( 1983 ) and Roberts ( 1986 ) and made avail- 

 able to us. Since then, additional otoliths have been 

 collected by us during research trawl surveys and 

 sampling trips aboard commercial line vessels, as 

 well as by Ministry of Fisheries scientific observers 

 aboard commercial line and trawl vessels. 



Most otoliths were accompanied by information on 

 fish total length (TL, to the centimeter below actual 

 length) and sex. Three subsamples of these otoliths 

 were selected for analysis. The first comprised 44 

 otoliths collected from five sites between Three Kings 

 Islands and Great Barrier Island (Fig. 2) (hereafter 

 called the "northern" sample). The second consisted 



of 178 otoliths collected from Cook Strait and 

 Kaikoura ( Fig. 2; "central" sample) and the third con- 

 sisted of 28 otoliths from large hapuku collected from 

 other sites throughout New Zealand ("miscellaneous" 

 sample), selected to boost the number of old fish in 

 the data set. 



One otolith from each pair was viewed under trans- 

 mitted polarized light, and the straightest dorso- 

 nuclear ridge was marked with a pen to guide the 

 saw for sectioning. Each otolith was then embedded 

 in a block of clear epoxy resin and sectioned trans- 

 versely through the core at a thickness of 650 \xm by 

 using a dual-blade high-speed diamond saw. One face 

 of the section was polished with caz'borundum paper 

 and glued on to a microscope slide with thermoplas- 

 tic cement. The opposite face was then ground and 

 polished to a final thickness of 200-350 |.im, depend- 

 ing on the clarity of the otolith bands when viewed 

 under both reflected and transmitted light. 



Opaque otolith bands were counted under trans- 

 mitted light in the dorsal part of each section by two 

 readers. Both readers carried out an initial training 

 exercise by making counts on a subsample of the sec- 

 tions, while knowing the collection details of each 

 fish (TL, sex, and site), and then discussed their re- 

 sults. Subsequent counts were carried out "blind"; 

 i.e. the readers did not know the size, sex, or collec- 

 tion site of the hapuku. Reader 1 counted all sec- 

 tions once (Rj) and reader 2 counted all sections twice, 

 the two counts being separated by two months (R., , 

 and R., .,). The readability of sections was scored on a 

 scale from 1 (unreadable) to 5 (exceptionally clear). 



Otolith band counts were assessed for ageing bias 

 and precision (between readers, and between read- 



