Zeldis et al.: An estimate of biomass of Hoplostethus atlanticus 
581 
used with the methods of Zeldis et al. (1995 ) and the 
Ritchie Bank CTD temperature profiles described 
below. Ritchie Bank temperatures were within the 
range of those observed in Zeldis et al. (1995). Cur- 
rent velocities as a function of depth were calculated 
from geostrophic velocity profiles (Pond and Pickard, 
1978) by using Guildline CTD profiles taken over 
Ritchie Bank. A reference depth of no motion at 1,000 
m was assumed, and geostrophic velocities at 100 m 
were corrected to match the 100-m velocities mea- 
sured by a buoy drogued to that depth. This buoy 
was deployed over Ritchie Hill on 28 June 1993, re- 
located 20 hours later, and subsequently lost. 
Correction factors were to convert egg counts to 
densities were also calculated (assuming the trajec- 
tory of the net was straight) for comparison with those 
where a curved trajectory was assumed. For this, the 
volume filtered v (m 3 ) was estimated by using 
v = 2 Jp 2 + z 2 , 
where p = the distance the ship drifted from deploy- 
ment to recovery of the net, determined 
with GPS; 
= the maximum depth of the net; and 
factor 2 = the area (m 2 ) of the net mouth. 
It was assumed that, at the start of hauling, the net 
was at the position of the vessel when deployment 
commenced, i.e. the net dropped vertically through 
the water during deployment. 
Estimates of egg abundance by age group in the 
survey area (N a ) were calculated by multiplying the 
mean egg density at age in each stratum by stratum 
area and by summing across strata. In the case of 
the time strata in the center of the survey area, the 
egg abundances were averaged before summing with 
the other strata. The CV of N was calculated by us- 
ing the standard deviation of egg density at age and 
stratum, weighted by stratum area. The average of 
the standard deviations was used in the case of the 
central time strata. 
The maximum age of eggs that could be used in 
estimating daily egg production was the maximum 
age for which there appeared to be no significant 
advection out of the survey area owing to water move- 
ments (a loss of eggs by advection would cause a nega- 
tive bias in N ). Advection was examined by plotting 
centroids of each age group (Appendix 2) and by us- 
ing the buoy and CTD data described above. 
The daily egg production, N 0 (eggs/day), and in- 
stantaneous mortality for eggs, Z (per day), were 
estimated by maximum likelihood with the mortal- 
ity model 
N a =N 0 e^ t 
where t = the mean age (days) of age group a (Ap- 
pendix 3). 
The precision of these estimates was estimated by a 
bootstrap procedure (Appendix 3). 
Daily fecundity per female weight 
Survey design Female fecundity and ovarian stage 
samples were taken from trawls from the RV 
Tangaroa on Ritchie Bank from 7 June-6 July and 
from commercial vessels on 22 June, 11 July, and 13 
July (Table 2). Trawling was done during the week 
before spawning started (8-11 June), to sample to- 
tal annual fecundity for the AEPM, and from the start 
of spawning until spent fish were common (20 June- 
13 July), to sample fecundity reduction for the DFRM. 
Twenty five of the trawls were from Ritchie Hill 
(within the area of the central stratum in Fig. 2B) 
and three were from North Hill, a spur off the north 
end of Ritchie Bank, about 12 km north of Ritchie 
Hill (Fig. 2B) where fish were spawning. 
Oocyte sampling, ovarian staging, and daily fecun- 
dity estimation The oocytes of 569 mature fish 
(about 35 fish/trawl) in macroscopic ovarian stages 
3 (late vitellogenic, prespawning), 4 (hydrated), 5 
(ovulated), or 8 (late vitellogenic, partially spent) 
were counted. Of these, 218 fish were prespawning 
and used for total annual fecundity analysis. The 
remainder were used for DFRM analysis. Oocytes 
were counted by using the automated system de- 
scribed in Appendix 4. 
The proportions of females in ovarian stages 3, 4, 
5, and 8, and stage 6 (spent) in the trawl samples 
were estimated by using macroscopic ovarian stag- 
ing of about 100 randomly chosen females per trawl. 
The ovarian stages were further grouped because 
it was observed that stages 3 and 4 (group 1) had 
higher fecundities/ kg than stages 5 and 8 (group 2); 
this difference was due to fish in group 2 having 
started spawning. Therefore, the estimate of fecun- 
dity per female weight, R p was stratified by these 
groupings to minimize error. Stage-6 fish (spent) were 
placed in group 3. Thus, R t was estimated as the 
mean number of eggs/kg of all females that would 
spawn, were spawning, or were spent, for trawl i: 
tv. 
j = i 
