250 



Fishery Bulletin 92(2), 1994 



actively spawning fish were caught, perhaps because 

 the survey had been taken in June/early July and 

 peak spawning on the Westpac Bank could have 

 occurred slightly later than in the other two areas 

 (Clark, 1991b). Originally it was thought that fish 

 on the Westpac Bank migrated east to the Central 

 Flat and Pinnacles for spawning (Clark and Tracey, 

 1988). However, in subsequent years ripe and run- 

 ning-ripe fish were found. Nevertheless, in 1987 

 there was a large proportion of fish that were not 

 spawning that year. It is possible that the Westpac 

 Bank has only developed as a spawning ground 

 since 1987, but if so, whether heavy fishing on the 

 other grounds was a factor is unknown. The bio- 

 mass index on the Westpac Bank has declined since 

 1987, so it is unlikely that there has been a major 

 shift of fish from the Central Flat or Pinnacles. 



The time of spawning (defined by 20% spent) has 

 consistently been in the second and third weeks of 

 July. Pankhurst (1988) reported that day length was 

 a critical factor in synchronizing the reproductive 

 cycle of orange roughy. The changes in dates of 20% 

 spent between years do not correlate exactly with 

 annual changes in the shortest day, but day length 

 could nevertheless be an important general cue. 

 Gonadal development has remained consistent de- 

 spite major changes in the size of the population and 

 the spawning school structure. There are indications 

 that heavy fishing pressure may disrupt the stabil- 

 ity of schools of orange roughy (Clark and Tracey, 

 1991). In 1989 when fishing effort was at its peak, 

 a comparatively high proportion of the biomass was 

 in 'background' areas, outside the three main re- 

 gions of spawning activity. Catch rates in the spawn- 

 ing areas were lower than in other years. It is pos- 

 sible that fishing pressure was affecting the forma- 

 tion of aggregations, but nevertheless reproductive 

 development still occurred normally. However, the 

 success of spawning could have been reduced be- 

 cause the fish were more dispersed. 



The diet of orange roughy, and the relative fre- 

 quency of occurrence of prey groups, were similar 

 over the period examined. Natant decapod crusta- 

 ceans and fish remains have dominated the diet. 

 This diet composition concurs with other accounts 

 of orange roughy feeding habits in New Zealand 

 waters (e.g. Liwoch and Linkowski 1986; Rosecchi 

 et al., 1988) and is similar to diet composition of 

 orange roughy in the North Atlantic Ocean (e.g. 

 Mauchline and Gordon, 1984; Gordon and Duncan, 

 1987), Indian Ocean (Kotlyar and Lipskaya, 1981), 

 and off southeastern Australia (Bulman and Koslow, 

 1992). The trophic effects of the decline in orange 

 roughy biomass are unknown. There is little infor- 

 mation on predator-prey relationships within com- 



munities containing orange roughy. Published feed- 

 ing studies and observations from research cruises 

 at different times during the year (authors' unpubl. 

 data) indicate that orange roughy do not prey on 

 eggs or larvae of other fish species. The only pub- 

 lished data on predation of orange roughy record 

 them in stomachs of seal shark on the Challenger 

 Plateau (Clark and Tracey, 1988). Sperm whales 

 (Physeter catodon) are often observed in orange 

 roughy spawning areas, and although they can dive 

 to depths of over 1000 m. It is uncertain whether 

 they feed on orange roughy. 



Orange roughy are slow-growing and long-lived 

 with low productivity, making them highly suscep- 

 tible to the effects of overfishing. Long-term sustain- 

 able yield for the Challenger Plateau stock is esti- 

 mated at 1.6% of virgin biomass (Clark, 1992). In 

 the early years of a developing fishery catch levels 

 are likely to be high. The schooling behavior of or- 

 ange roughy for spawning or feeding means that 

 large catches can be taken in a short time, and high 

 catch rates may be maintained despite decreasing 

 biomass. CPUE declined on the Challenger Plateau 

 but not as consistently on the Chatham Rise where 

 the population was also reduced by heavy fishing 

 (Francis et al., 1992), although there was a progres- 

 sive shortening of the period over which high catch 

 rates occurred (Coburn and Doonan, in press). 



In 1986, the TAC on the Challenger Plateau was 

 increased from 6,000 to 10,000 t in order to assess 

 the impact of heavier fishing and to learn more 

 about the productivity of orange roughy. At that 

 stage there was little understanding of stock size, 

 or age and growth characteristics of the species. 

 Hence the intention was to increase catch suffi- 

 ciently to provide a contrast in abundance indices 

 and give information on the resilience of the stock. 

 However, there were several problems with this 

 'adaptive management' strategy as applied to Chal- 

 lenger Plateau orange roughy. The first was that it 

 began without good data on the abundance of the 

 stock against which to measure any change. It would 

 have been preferable to have had at least two years 

 of abundance data before increasing fishing pres- 

 sure. At the time, CPUE had not been examined, 

 and trawl survey results were inadequate to esti- 

 mate biomass. A new time series of trawl surveys 

 began in 1987, and although the 1988 survey 

 showed a large decrease, with only two survey re- 

 sults we could not be confident about interpreting 

 the differences as a strong decline in stock size. A 

 further difficulty with such management is that 

 with a slow-growing species like orange roughy, 

 potential effects of any changes in spawning stock 

 size on recruitment will not be evident for 20-25 



