384 



Fishery Bulletin 93(2), 1995 



by about age 50 h (stage 13), where they reside until 

 about age 150 h (stage 26); then they sink rapidly 

 for about 50 more h before hatching. Since the hatch- 

 ing eggs sink at about 381 m-day -1 and their density 

 is greater than seawater near the bottom in the 

 spawning area on the North Chatham Rise (870 m), 

 they may hatch near the bottom. This vertical dis- 

 tribution pattern of the eggs may partially explain 

 the distributions of early juvenile orange roughy on 

 the North Chatham Rise. The spawning aggregation 

 at 177°W (Fig. 8A) forms each year and has supported 

 the highest measured orange roughy biomasses on 

 the North Chatham Rise since the inception of the 

 trawl surveys on the stock (Zeldis, 1993). There are 

 high concentrations of 0+ to 1+ annual cohort orange 

 roughy about 50 to 175 km east of this aggregation 

 (Fig. 8B; Mace et al., 1990) and this is the only area 

 on the rise where early juveniles have been found in 

 large quantities. Mean, residual currents are directed 

 eastwards along the North Chatham Rise (Heath, 

 1983; Chiswell, 1994). These currents are generally 

 weak (about 6 cm-sec -1 in the mixed layer), and the 

 current velocity decreases with depth (Heath, 1983; 

 Chiswell, 1994). During their 10-day development 

 period, orange roughy eggs spend about 6 days in 

 the mixed layer and a total of about 4 days ascend- 

 ing to and sinking out of the mixed layer. Therefore, 

 on average, eggs spawned in the aggregation at 

 177°W would be expected to drift somewhat less than 

 50 km east before hatching near the bottom. The 

 subsequent distribution of the newly hatched larvae 

 is unknown, but since the larvae are more dense than 

 bottom waters of the North Chatham Rise (Grimes 

 and Zeldis, unpubl. data) they may remain near the 

 bottom and drift into the area of high 0+ and 1+ ju- 

 venile density just downstream. 



In this study, middle-age and older eggs were dis- 

 placed in near-surface waters some 18 km to the west 

 of the concentration of spawning adults (Figs. 7 and 

 8A), contrary to the expected eastward drift. This 

 unexpected location may have resulted from smaller- 

 scale eddy and tidal flow embedded in the large-scale 

 mean eastward flow on the North Chatham Rise 

 (Chiswell, 1994). 



This study has determined the age-at-stage of or- 

 ange roughy egg development as vertical distribu- 

 tion changes during ontogeny. These results were 

 formerly unknown for this species. They are, how- 

 ever, essential for egg production estimation of or- 

 ange roughy biomass where they are used for esti- 

 mating egg mortality rates and rates of egg produc- 

 tion at spawning. A further step is to ask how sensi- 

 tive egg production estimation is to inaccuracies in 

 egg age determination. This is significant in regard 

 to other studies which may have less information 



than this one on vertical distributions of eggs during 

 ontogeny. An example is the egg production study by 

 Lo et al. ( 1992) for Dover sole, Microstomas pari ficus, 

 biomass, in which eggs were aged on the assump- 

 tion that the temperature of development was the 

 mean temperature of the stratified layer for the eggs 

 they found in that layer. This assumption does not 

 use depth-specific development rates to age eggs in 

 the stratified layer and if it were applied to orange 

 roughy, would cause egg age to be underestimated 

 (dashed line in Fig. 5). To examine this effect on or- 

 ange roughy egg production estimates, these under- 

 estimated egg ages were used to estimate the pro- 

 duction rate of orange roughy eggs from the Ritchie 

 Bank egg production survey and were compared with 

 the result when the thermal history model (solid line 

 in Fig. 5) was used to age eggs. Maximum likelihood 

 fits of both sets of egg abundance-at-age data were 

 done by using only stages <10, because older eggs 

 were advected out of the survey area. 2 Egg produc- 

 tion rate was biased upward by 31% for the fit with 

 underestimated ages relative to the fit in which ages 

 from the thermal history model were used. This hap- 

 pened because the same numbers of eggs-at-age were 

 assumed to have occurred but over a shorter time 

 for the fit with underestimated ages. 



Dover sole eggs have a similar diameter and 

 spawning depth range to that of orange roughy eggs 

 (Lo et al., 1993). Therefore, they should ascend the 

 water column at a similar rate, excluding large dif- 

 ferences in water column density and egg density. 

 Therefore, the ages of young Dover sole eggs may be 

 underestimated by a similar extent as shown above 

 for orange roughy, and their production rate may be 

 similarly overestimated when they are aged without 

 using depth-specific development rates. Thus, it 

 seems prudent to reduce ageing bias by using a 

 method that accounts accurately for thermal history 

 of eggs. This is especially important if eggs found 

 below the mixed layer contribute exclusively or sub- 

 stantially to the data set used for estimating egg pro- 

 duction. This was the case for the Ritchie Bank egg 

 production survey and has been suggested by Lo et al. 

 ( 1993) for future Dover sole egg production surveys. 



The consistency between the thermal history model 

 results and field results on orange roughy egg verti- 

 cal distributions suggests that age-at-stage should 

 be predictable in other orange roughy spawning ar- 

 eas (e.g. the Ritchie Bank) where egg production sur- 

 veys are conducted, provided data are available on 

 water column density and temperature structure. It 

 should also be possible to age other pelagic fish eggs 

 by using depth-specific development rates, as long 

 as ascent rates are not affected by chorionic sculp- 

 turing (Robertson, 1981). This would require data 



