Francis et al,: Age and growth estimates for Polypnon oxygeneios 



235 



suggested that females grew slightly faster than 

 males, although there was strong overlap in the 

 ranges (males: mean residual=-1.8 cm, SD:=8.6, 

 n=90; females: mean=1.6, SD=7.8, n=82). All growth 

 curves had strongly negative values for ^q (Table 2). 



Length-frequency distributions 



Hapuku from the four southern trawl surveys gen- 

 erally were less than 85 cm long (Fig. 9), and be- 

 cause 50% maturity occurs at about 85 cm and 88 

 cm for males and females, respectively (Johnston, 

 1983), these would have been immature. The small- 

 est hapuku was 41 cm long, but few fish were less 

 than 50 cm. This reflects the pattern seen in the en- 

 tire New Zealand research trawl database which con- 

 tains length measurements for 5841 hapuku: the 

 smallest hapuku was 40 cm, and only 55 were less 

 than 50 cm long. 



The best-fit MULTIFAN model consisted of con- 

 stant length standard deviation with 11 age classes 

 (see Table 2 for growth parameters). The estimated 

 time elapsed between the theoretical birthday and 

 the first appearance of juvenile hapuku in the length- 

 frequency samples (t^) (based on an estimated mini- 

 mum age at first capture from otolith ageing of 3 

 years, plus 0.5 years elapsed between the theoreti- 

 cal birthday [1 September] and the sampling date mid- 

 point [1 March] ) was 3.5 years. The age in years since 

 zero length of the youngest age class at the time it first 

 appeared in the length-fi-equency samples (Oj) was es- 

 timated by MULTIFAN to be 11.52 years, producing a 

 t^ estimate of -8.02 years. The standard error estimates 

 provided by MULTIFAN for the growth parameters are 

 not presented in Table 2 because they tended to be 

 unrealistically small (Francis and Francis, 1992). 



50 1 



40 - 



c 

 o 



.« 30 



(0 



> 



c 



(D 

 O 

 it 

 CD 

 O 

 O 



20 - 



10 





5 10 



Age (Rj j) (years) 



15 



Figure 7 



Precision of hapuku age estimates of reader 2 (R., ,) and 

 reader 1 (Rjl plotted against the ages determined for the 

 same fish by reader 2 (R,, .,). The data points are the coeffi- 

 cients of variation of the data shown in Figure 4. 



The growth curve resulting from the best MULTI- 

 FAN fit to the length-frequency data was nearly lin- 

 ear over the range of the data (Fig. 10). The curve 

 was truncated at 14.5 years, which is the maximum 

 age covered by the data (calculated as the sum of the 

 number of age classes detected by MULTIFAN [11] 

 and the estimated age at first recruitment to the 

 samples [3.5 years]). The MULTIFAN curve should 

 not be extrapolated beyond this point because the 



