380 



Fishery Bulletin 93(2), 1995 



the mixed layer, at a constant, mean temperature of 

 11.25°C, until about cumulative age 150 h was 

 reached (shown only to stage 15 or cumulative age 

 57.4 h in Fig. 5). 



To examine the effect of variability of the labora- 

 tory-determined ascent rates on the predictions of 

 age-at-stage, the model was rerun by using the as- 

 cent rates in Table 1 ±their 95% confidence inter- 

 vals (taken to be 60 m-day" 1 from Fig. 3), to give 95% 

 confidence intervals about the mean age-at-stage. 

 Depth reached at each stage was also determined (Fig. 

 6) from the ascent rates and time intervals used to run 

 the model, along with the depths reached when the 

 ascent rates ± their 95% confidence intervals were used. 



Modeling of age- and depth-at-stage from stage 26 

 onward was not attempted because of uncertainty 

 in modeling age-at-stage at temperature for these 

 older eggs (see embryology and development rates 

 above). Also, the descent rate data (Fig. 3) were in- 

 sufficient to describe the rate at which eggs in these 

 stages increase in sinking rate and the position of 

 the eggs in the mixed layer when sinking starts was 

 not known. It is likely that mixing would redistribute 

 neutrally buoyant eggs in the mixed layer before the 

 eggs begin to sink. 



Field vertical distribution 



<stage 7 (32 cell); all other samples with eggs older 

 than this had no damaged eggs. The damage was 

 observed prior to preservation and evidently occurred 

 during the hauling of the plankton net. These eggs 

 were staged by proration, from the frequencies of 

 undamaged eggs <stage 7 in these nets. In tow 546 

 (Fig. 7), the nets fishing to 783 and 672 m contained 

 34% and 32% of eggs damaged, respectively, and the 

 bottom net in tow 545 contained 61% damaged, as 

 percentages of total eggs <stage 7. 



Field vertical distributions of egg stages (Fig. 7) 

 were compared with their predicted depths from the 

 thermal history model (Fig. 6). To facilitate this com- 

 parison, the depth intervals of stages 1 to 14 were 

 determined (Table 2) by using the 95% confidence 

 intervals around the depths reached at the begin- 

 ning and end of each stage (Fig. 6). These depth in- 

 tervals were compared with the egg densities-at-stage 

 summed over all MOCNESS tows within 200-m strata 

 (summary column in Fig. 7; Table 3) and with catches 

 from individual nets within these 200-m strata. 



Results 



Two features were evident in the egg catches from 

 the MOCNESS series. First, the abundance of dif- 



To study vertical distribution of eggs in the field, a 

 multiple opening-closing net and environmental 

 sensing system (MOCNESS) was used on the North 

 Chatham Rise to take plankton, temperature, and 

 salinity samples from known depth strata. The 

 MOCNESS tows were made at various times of day 

 in areas where catch rates of adults had been rela- 

 tively high or where exploratory plankton tows con- 

 ducted in previous days had produced relatively high 

 catch rates of orange roughy eggs. On three of the 

 MOCNESS tows, the first of the nine nets was used 

 for an integrated haul from the surface to within 

 about 90 m of the bottom at 870 m; then the ascent 

 was stratified into approximately 100-m intervals by 

 the remaining nets (see Fig. 7 for exact stratum in- 

 tervals). On two of the tows, both descent and ascent 

 were stratified into approximately 200 m intervals. 

 Eggs caught in the tows were either staged on board 

 and then grown in the culture facility or were pre- 

 served in 4% buffered formaldehyde in seawater and 

 staged in the land-based laboratory. Preservation had 

 little effect on staging accuracy, because stage fre- 

 quencies of eggs staged in the laboratory were very 

 similar to those staged in the fresh state on board. 

 However, in three of the nets some of the young eggs 

 had embryos which were damaged and could not be 

 staged with certainty. The damaged eggs were all 



