Zeldis et a I.: An estimate of biomass of Hoplostethus atlanticus 
585 
Data from the completed subsurveys 3 and 4 
showed that egg catches were highest in the central 
and middle strata (Fig. 3, C and D). These eggs were 
predominately in the young, grouped-age category 
(stage 7 or less, <21.8 h old; Table 1), but many 
middle-aged and some older eggs were also caught 
in these strata. A few orange roughy yolk-sac larvae 
were also caught in the central and middle strata. 
These high catch rates for eggs were spatially corre- 
lated with high research trawl catch rates for adults 
(Fig. 2, A and B; Table 2 ). Catches in the outer strata 
were usually <10 eggs/tow or zero eggs/tow. However, 
there was one large catch of young eggs near North 
Hill (Fig. 3D), on which relatively large catches of 
spawning adults were made with research trawling 
(Table 2) and commercial trawling. All of the undam- 
aged eggs in this sample (19% of all eggs) were at 
the 1-cell stage, indicating that these eggs arose from 
localized spawning on North Hill. Four additional 
random tows were then made within an extra stra- 
tum created in this area (Fig. 3D) at the end of 
subsurvey 4. Only 5 additional eggs in total were 
caught in these tows, indicating that egg production 
on North Hill was low compared with that on Ritchie 
Hill. 
A few moderate egg catches were also made in the 
southwestern corner of the survey area (Fig. 3, C and 
D). Most of these eggs (84%) were > stage 7. Four 
additional tows were then made within an extra stra- 
tum created in this area (Fig. 3D) at the end of 
subsurvey 4, and in three of these tows, all eggs were 
> stage 7. The fourth sample had many damaged eggs 
(< stage 7), but it was likely that these were at the 
older end of the age range of “young” eggs, judging 
by the stages of undamaged young eggs in that 
sample (all were > stage 5). The bottom in the area 
where these samples were taken (Fig. 2B) is deeper 
(>1,000 m) than depths at which orange roughy nor- 
mally spawn (850-900 m), and trawl catch rates in 
the area were very low in this survey and in previ- 
ous surveys (Fig. 2, A and B). This finding indicated 
that these eggs had not been produced locally, but 
rather had been advected from the main spawning 
center on Ritchie Hill. 
The positions of the centroids (centers of gravity) 
of successive egg age groups suggested that advec- 
tion was initially to the southwest out of the survey 
area (Fig. 4 A) but that older eggs (those of the very- 
broad-age-group stage 16+, >54.3 h old; Table 1) re- 
entered the area, possibly from the east. In inter- 
preting Figure 4, it is important to realize that cen- 
troids close to the boundary of the survey area were 
unlikely because only eggs from within the survey 
area were used in the calculation of centroid posi- 
tion. An estimate of the average rate of advection, at 
the depths of these egg stages, was calculated by di- 
viding the distance between the centroids of stages 
<7 and 11 by the difference between the mean ages 
in these two age groups (Appendix 3) and was found 
to be 7.4 cm/sec. 
The distribution of eggs by age within the survey 
area (Fig. 4, B and C) confirmed the southwesterly 
drift pattern. Eggs > stage 11 (> 28.4 h old) became 
increasingly centered in the southwestern corner of 
the survey area (region 1, Fig. 4, B and C). However, 
the old eggs in the stage 16+ group were most abun- 
dant in the eastern and central regions (Fig. 4, B 
and C). 
The advection inferred from egg distributions can 
be compared with hydrographic results. The drogued 
buoy was relocated 11.2 km south-southwest 
(bearing=207°) of the release site after 20 h at lib- 
erty (Fig. 5A), indicating that advection (at 100 m 
depth) was to the south-southwest, at a rate of 16 
cm/sec. Geostrophic analysis of the CTD data (Fig. 
5, A-D) indicated that in the northern part of the 
station grid (in the vicinity of the Ritchie Hill spawn- 
ing site), current directions turned from south-south- 
east through south to southwest as depth increased 
from 100 through 400 to 800 m. The southwestern 
component of current velocity between 800 and 400 
m (the approximate depth range of eggs < stage 11; 
Table 1) in the vicinity of Ritchie Hill averaged about 
7 cm/sec (Fig. 5D). Geostrophic current speeds aver- 
aged about 12-13 cm/sec in the upper 100 m of the 
water column, in the vicinity of Ritchie Hill. These 
speeds were probably underestimates because the 
velocity profiles showed little evidence of reaching 
asymptotically low values as the postulated level of 
no motion (1,000 m) was approached (Fig. 5D), sug- 
gesting that some residual velocity existed at that 
level. This may explain the greater buoy speed than 
geostrophic speed at 100 m. Current speeds toward 
the southwest (Fig. 5D) were higher on the northern 
side of the grid than on the southern side through 
the upper water column; a significant easterly com- 
ponent was observed in the upper 200 m. 
Thus, the advection pattern indicated by egg cen- 
troids and stage distributions (Fig. 4, A and C) was 
consistent with results from the drogued buoy ( Fig. 
5A) and the geostrophic analysis (Fig. 5, A-D), which 
indicated that the direction of drift of young eggs in 
the lower half of the water column was toward the 
southwest. The geostrophic velocities in the lower 
half of the water column in the vicinity of Ritchie 
Hill (at least 7 cm/sec) were consistent with the ve- 
locity of egg advection (7.4 cm/sec) from our calcula- 
tions. The older eggs, which would have spent most 
of their time in the mixed layer (Zeldis et al., 1995), 
might have been conveyed into the survey area from 
