RALSTON and WILLIAMS: AGEING OF TROPICAL FISHES 



mentally altered somatic growth rates of 0. keta 

 juveniles with different experimental feeding 

 regimes and showed a direct linear effect on the 

 mean width of daily increments. SimOarly, Marshall 

 and Parker (1982) presented data showing an in- 

 crease in the relative size of otoliths of starved 0. 

 keta compared with that of fed controls, even though 

 starvation had no effect on the number of incre- 

 ments. In contrast, Neilson and Geen (1985) found 

 no effect due to ration alone on the thickness of in- 

 crements in fry of 0. tshaunjtscha, although an inter- 

 active effect due to ration level and water temper- 

 ature was shovra. These authors also found that 

 increased feeding frequency significantly reduced 

 mean increment width. Lastly, Campana (1984) 

 found that increments of larval (<10 days old) 

 Porichthys notatus were more irregularly spaced 

 than in juveniles, as were the increments of fish ex- 

 posed to a constant photoperiod environment. From 

 these results, it is apparent that the effect of food 

 ration on the width of daily increments is complex 

 and is at present not well understood. 



There is still some question concerning how close 

 the coupling is between somatic and otolith growth 

 rates (Brothers 1981; Bradford and Geen 1987). 

 Over the entire lifespan, otolith length and FL 

 typically are highly correlated (Templeman and 

 Squires 1956; Blacker 1974). This situation could not 

 arise were not the growth rates correlated over a 

 similar scale. Still, in the most rigorous examina- 

 tion of the extent of rate coupling to date, Bradford 

 and Geen (1987) found no correlation between the 

 observed growth rates of individual 0. tshawytscha 

 fry and otolith increment widths over relatively 

 short-term (7-15 d) intervals, although a good corre- 

 lation over a 51 d interval was observed. These 

 authors point to the relatively conservative char- 

 acter of otolith growth (Casselman 1983; Gutierrez 

 and Morales-Nin 1986) as the reason for short-term 

 uncouplings between somatic and otolith growth 

 rates. 



In our study, otolith microstructure typical of daily 

 increments was observed in the sagittae of gindai 

 (Fig. 2). Daily growth increments were previously 

 described and illustrated for congeneric species by 

 Ralston and Miyamoto (1981, 1983), Brouard et al. 

 (1984), and Radtke (1987). Likewise, we observed 

 annual hyaline and opaque zonations, which have 

 been reported in the hard parts (otoliths and verte- 

 brae) of other lutjanids (Loubens 1978; Chen et al. 

 1984; Edwards 1985; Manooch 1987; Samuel et al. 

 1987). Still, of the 11 deep slope species {Pristi- 

 pomoides zonatus, P. auricilla, P. filamentosus, P. 

 sieboldii, P.flavipinnis, Aphareus rutilans, Etelis 



comscans, E. carbunculus, Lutjanus kasmira, 

 Caranx lugubris, and Selar crumenophthalmus) 

 caught during the Marianas survey and whose oto- 

 liths were examined in some detail (Ralston and 

 Williams 1988), only gindai displayed hyaline and 

 opaque zonations, even though all species exhibited 

 microstructure typical of daily growth increments. 

 The absence of annuli in the otoliths of these other 

 species is difficult to explain because many are con- 

 geners, most are confamilials, and all but one (S. 

 crumenophthalmus) occupy the same general deep- 

 water habitat where gindai are found. As a group, 

 these fishes are exposed to virtually identical 

 environmental conditions. Neither is the diet of gin- 

 dai in the Marianas particularly distinctive (Parrish 

 1987). 



In contrast to the situation in the Marianas, 

 studies by Loubens (1978) in New Caledonia and 

 Samuel et al. (1987) in the Persian Gulf document 

 distinctive hyaline and opaque annuli in a wide varie- 

 ty of the taxa indigenous to these areas. Although 

 the occurence of annuli in the otoliths of a variety 

 of tropical and subtropical species is now well docu- 

 mented (Manooch 1987), our understanding of when 

 and how they form is quite limited (see below). 



Von Bertalanffy growth curves were developed 

 for gindai by using both increment microstructure 

 (Fig. 4) and annual marks. Likewise, the L„ param- 

 eter of the von Bertalanffy growth equation was 

 estimated by using the regression method of 

 Wetherall et al. (1987). Moreover, the analysis based 

 on annual markings was tentatively validated with 

 an abbreviated form of marginal increment analysis, 

 wherein the seasonal presence or absence of opaque 

 margins was established for the various pooled ring 

 groups. A preferred approach is to measure the 

 marginal increment for each ring group separately 

 (e.g., Chen et al. 1984; Matheson et al. 1986). Al- 

 though the importance of this type of validation has 

 been overlooked (e.g.. Beamish and McFarlane 

 1983), it is a very useful technique, especially in 

 situations where capture is fatal. 



A comparison of von Bertalanffy parameter esti- 

 mates obtained by the three wholly independent ap- 

 proaches (increment microstructure, annuli, and 

 length-frequency analysis) shows reasonable corre- 

 spondence. The two estimates of growth coefficient 

 (K) differed somewhat (0.234 versus 0.156 yr"'), 

 although estimates of L„ were substantially closer 

 (442, 537, and 466 mm FL, respectively). Given that 

 the annual marks were only weakly expressed, these 

 findings support the conclusion that the microstruc- 

 ture observed in gindai otoliths (Fig. 2) results from 

 the daily accretion of increments and, to the extent 



11 



