Newman and Dunk; Age validation, growth, mortality, and additional population parameters of Phstipomoides multidens 



125 



chemistry of these fish, culminating in the formation of an- 

 nual growth increments. 



Female and male fish older than 20 years of age were 

 uncommon in the landed catch. Fish of both sexes between 

 5 and 12 years of age were common in the landed catch. 

 The maximum age of P. multidens observed in our study 

 was much greater than that recorded previously. Richards^ 

 reported a maximum age of 14 years in Papua New Guinea 

 from counts of daily rings on otoliths, whereas Brouard 

 et al. (1984) recorded a maximum age of only 8 years in 

 Vanuatu with a similar method. Edwards (198.5) analyzed 

 vertebrae and scales of this species in the Timor Sea and re- 

 ported a maximum age of 14 years. In contrast, Mohsin and 

 Ambak (1996) estimated a maximum age of only 5 years 

 from the east coast of peninsular Malaysia with length- 

 frequency analysis. Variation in the longevity estimates of 

 earlier works is related to the aging methods used and their 

 biases. For example, growth increments in vertebrae are 

 often difficult to detect despite the presence of numerous 

 discontinuities in bone growth (Marriott and Cappo, 2000). 

 Alternatively earlier longevity estimates may have been 

 drawn from sample populations biased by gear selectivity 

 or from populations with varying degrees of exploitation. 



Otolith weight was a good predictor of age in P. multidens, 

 accounting for 94% of the variability in age. The strong lin- 

 ear relationship between otolith weight and fish age from 

 a very large sample size implies that otolith weight may be 

 used as a proxy for age. The coefficient of determination of 

 the regression model is affected by the degree of colinearity 

 of the independent variables. The high r^ value observed in 

 our study provides the basis for a first-order age approxi- 

 mation. Thus, the potential exists for an age-otolith-weight 

 key to be derived for P. multidens, as for an age-length key, 

 whereby the age composition of the landed catch in future 

 years may be obtained by weighing large numbers of oto- 

 liths. However, the accuracy and precision of adopting this 

 monitoring strategy remains to be tested. 



The fit of the regression model for the otolith weight- 

 age relationship was much more precise than the fit of the 

 fork-length-age relationship as described by the von Ber- 

 talanffy growth model. Considerable variation in length 

 was observed within most age groups for both sexes. The 

 large variation in length at a given age makes it difficult to 

 accurately determine the age of P multidens from length 

 data alone. For example, fish ranging in length from 450 to 

 550 mm FL may vary in age from 5 to 30 years. This vari- 

 ability may explain the very low estimate of maximum age 

 obtained by Mohsin and Ambak ( 1996), which was derived 

 with length-frequency analysis. 



Growth was most rapid through age 9 for both sexes. 

 From age 9 onwards somatic growth slows with increasing 

 age. The estimation of growth parameters is dependent 

 upon adequate sampling across the length range of any 

 species. The fish sampled in our study ranged in length 

 from 80 to 701 mm FL, covering most of the length range 

 of P multidens. Therefore, it is unlikely that the growth 

 parameters of P. multidens are biased because of inad- 

 equate sampling across the length range. 



Despite methodological differences in age estimation, 

 the estimates of K derived from the studies of Richards' 



1995 

 n = 328 



IS 20 22 24 26 28 30 32 



1996 

 n=984 



Hn-n^ 



4 fi S :() 22 24 2f> 28 .^0 32 



1997 

 n.702 



„nnn^ 



s 1(1 i; 14 



199S 

 11=1126 



OQDii 



14 16 IK 211 



24 26 2S M) ^2 



1999 

 n=573 



24 26 2s Ml M 



Age (years) 

 Figure 9 



Age-frequency distributioii.s ol P. multidens in the 

 northern demersal scalefish fishery from ISO,') to 1999. 



