Schiel and Breen. Population, ageing, and fishing mortality of Hahotis ins 



685 



and egg production. For each region, the von Berta- 

 lanffy growth curves from mark-recapture data were 

 used. 



We held the age-at-first-capture constant at the first 

 age-class with a mean length equal to or greater than 

 125 mm. Then we varied the instantaneous rate of 

 fishing mortality, F, to estimate F 0l and F 2 s%. F .i is 

 the fishing mortality rate at the point where yield-per- 

 recruit increases with F one-tenth as fast as it does at 

 the origin. Although arbitrary, F ()A is used by several 

 major international stock assessment agencies as a 

 guide to appropriate target fishing mortalities (Mace 

 1988). F 2 5% is the fishing mortality rate at which equi- 

 librium egg production is 25% that of an equilibrium 

 virgin population. 



Results 



Field sampling 



Haliotis iris were abundant at all sites, despite more 

 than 30 years of commercial fishing activity in all 

 regions. It was not our intention to estimate densities, 

 but an indication of abundance is given by the length 

 of time it took to obtain our samples. The mean was 

 155 paua per diver-hour (SD 33.6) or about 32kg whole 

 wt per diver-hour. 



The structure of paua populations varied con- 

 siderably among the seven regions (Fig. 3). The three 

 regions of Stewart I. show clear differences in the 

 modes (Fig. 3A-C). The west coast region had the 

 major mode at 140-145 mm shell length, well beyond 

 the minimum legal size of 125 mm, and paua were found 

 up to 187 mm. The major mode for the northern region 

 was at 125-130 mm. The east coast region had a higher 

 proportion of smaller paua and the major mode was at 

 120mm. 



In contrast to Stewart I., length-frequencies for the 

 three regions of the Marlborough Sounds were similar 



to one another (Fig. 3D-F). The major modes were 

 around 125 mm (MLS), and smaller paua were found 

 in each region. There has probably been roughly similar 

 fishing pressure in these three regions in recent times. 

 Although the Karori Rock site is closed to commer- 

 cial activity, it is subjected to considerable illegal fish- 

 ing (Schiel 1989). The major mode in length-frequency 

 is at 130 mm (Fig. 3G). Because illegal fishing removes 

 paua of all sizes, it reduces paua densities but does not 

 necessarily change the population size structure. 



Ageing and growth 



A typical ring-count frequency is shown for the Marl- 

 borough Sounds D'Urville I. region (Fig. 4). Paua ap- 

 pear to be fully recruited to the sampling procedure 

 by 'age' 10. The greatest ring count we observed was 

 34 at the west coast of Stewart Island. Growth rates 

 were obtained from the ring count data for all sites 

 (Table 2). Growth parameters from ring-count data and 

 mark-recapture experimental data for two regions 

 (Table 3) and the mean increments predicted from the 

 two methods (Fig. 5) were compared. A large differ- 

 ence between the two methods was evident: the mark- 

 recapture data resulted in higher K estimates and 

 larger L^ estimates; estimated growth from mark- 

 recapture data was much greater than from ring-count 

 data. This pattern is the converse of what would be ex- 

 pected if the tagging procedure had retarded growth; 

 therefore, we consider the mark-recapture data to be 

 more reliable. The mark-recapture data are inconsis- 

 tent with the assumption that one, or even two, rings 

 are laid down annually. 



Mortality estimation 



Total mortality rates from the catch-curve analyses 

 ranged from Z = 0.216 -0.425 (Table 4). Mortality 

 estimates based on the length-frequency analysis, and 

 the parameters used to obtain them, are shown in Table 

 5. The estimates varied when N, NFULL, K, and the 

 standard deviation constraints were varied; the pattern 

 of change was not consistent between data sets. Only 

 variation in N caused large variation in estimated Z. 

 The Methods describe an objective procedure for choos- 

 ing N. The 'best' estimates (using the criteria described 

 above) were Z = 0.495 and Z = 0.507 for the northern 

 Marlborough Sounds and D'Urville regions, respective- 

 ly. The extreme range observed over all parameter 

 variations in the two data sets was Z = 0.238 -0.995. 



Modeling 



Growth rates used in modeling were those estimated 

 from mark-recapture data (Table 3) for the northern 



