FISHERY BULLETIN: VOL. 87, NO. 4, 1989 



Because studies on fish growth are generally 

 conducted at a single time, time series on the 

 order of several decades are either rare or ab- 

 sent. Some available time series are often pro- 

 duced from different studies, which may not 

 have used the same techniques for sampling, 

 ageing, or interpretation (Boehlert and Kappen- 

 man 1980). Using fish scales and interopercular 

 bones from two flatfish species collected over 

 approximately 15 years, Kreuz et al. (1982) 

 established time series of gi-owth and found that 

 growth was negatively correlated with tempera- 

 ture. Width and length of otoliths can be used to 

 estimate length of fish so that past growth pat- 

 terns can be estimated with back-calculation 

 techniques on otoliths from older fish. Extreme 

 longevity has been documented in the scorpaenid 

 genus Sebastes (Bennett et al. 1982; Leaman and 

 Nagtegaal 1987), and ages in excess of 80 years 

 have been reported for several species (Archi- 

 bald et al. 1981; Boehlert and Yoklavich 1984; 

 Leaman and Beamish 1984). Thus, otoliths of 

 these species can potentially be used to estimate 

 growth from several decades ago. 



In this paper, we describe an otolith-based 

 technique for obtaining a historical time series of 

 gi'owth. We apply this technique to two species, 

 the canary rockfish, Sebastes pinniger, and the 

 splitnose rockfish, S. diploproa, and describe the 

 resultant time series of growth in light of 

 physical and biological factors. 



MATERIALS AND METHODS 



Otoliths from S. pinniger and S. diploproa 

 were collected during rockfish surveys con- 

 ducted by the Northwest and Alaska Fisheries 

 Center, National Marine Fisheries Service, 

 NOAA, in 1977 (Boehlert 1980), 1980 (Boehlert 

 and Yoklavich 1984), and 1983 (Wilson 1985). 

 Collection techniques followed Gunderson and 

 Sample (1980). Our objective was to represent as 

 many years of growth as possible; therefore, 

 otolith selection was based upon age alone, and 

 old fish gi'eatly outweighed their relative abun- 

 dance in a random sample. Additional otohths 

 from S. pinniger, collected off the central 

 Oregon coast in 1984, were obtained from the 

 Oregon Department of Fisheries and Wildlife. 



Technical aspects of our methods, described in 

 Boehlert and Yoklavich (1987), are repeated 

 here for completeness because the original infor- 

 mation is not widely available. For otolith sec- 

 tioning and age determination, we followed 

 Boehlert and Yoklavich (1984). Briefly, otohths 



were affixed to cardboard tags, embedded in 

 polyester casting resin, mounted on a diamond 

 lapidary saw, and fed onto a pair of thin diamond 

 blades separated by acetate spacers. Dorsal- 

 ventral sections through the focus and perpen- 

 dicular to the sulcus, approximately 0.4 mm 

 thick, were removed from the center of the 

 otolith, attached to microscope slides, and 

 ground to eliminate artifacts. Total ages were 

 determined from these sections by identifying 

 the first translucent annulus (winter growth 

 zone) and counting sequential gi-owth zones from 

 the center toward the dorsal edge; dorsal annuli 

 at ages near 8-10 years were followed from the 

 dorsal edge to the interior dorsal quadrant, and 

 subsequent annuh were counted to the internal 

 surface. 



Annulus measurements, used as a proxy of 

 annual fish growth, were limited to the first 6 

 years of growth. This limit was imposed because 

 otolith increments became smaller with increas- 

 ing age; eventually, linear gi-owth stopped, and 

 the otolith began to thicken (Bennett et al. 1982; 

 Boehlert 1985). Two different techniques for 

 otolith measurement were used. In the faster 

 growing S. pinniger, whose otoliths are clearer, 

 measurements were from the focus to the dorsal 

 distal edge of each increment (Fig. 1). In S. dip- 

 loproa, whose otoliths are typically more opaque 

 and whose annual increments are smaller, the 

 focus was difficult to identify; for this species, 

 measurements were from the dorsal to ventral 

 distal edges of increments 1-6. 



Data from each fish included sex, fork length, 

 date of collection, location, total age, and widths 

 of the increments measured from the otoliths for 

 ages 1-6. The first increment width was a true 

 reflection of first year growth; subsequent 

 measurements, however, integrated past gi'owth. 

 That is, a large growth increment in year 1 could 

 bias the measurement in years 2-6. To remove 

 this correlation and to provide a measure of 

 gi'owth in each year, our analysis used growth 

 increments (GI(/), where i = 1-6) determined by 

 differencing successive measurements. Thus, 

 gi'owth in a given year was not cumulative and 

 generally did not reflect past growth. 



Age was subtracted from the year of collection 

 to determine the year of birth. For each fish, 

 each growth increment (1-6) was associated with 

 a specific "year of gi'owth" in the following man- 

 ner: Increment 1 represented the year of birth or 

 first year growth, increment 2 represented the 

 following year, and so forth. Data on 6 years of 

 gi'owth were available for each fish with the ex- 



792 



