Livingston et al.: Estimating the annual proportion of nonspawning adult Macruronus novaezelandiae 
101 
proportion of female hoki on the Southern Plateau 
in May that would spawn in the coming season and 
2) the proportion of females that already have begun 
their spawning migration by May. In this paper, we 
detail 1) the histological basis for classifying hoki as 
either nonspawners or prespawners and 2) the analy- 
ses used to estimate the maturity ogive and the to- 
tal proportion of fish that would spawn in July or 
August. 
Materials and methods 
Trawl surveys 
Two sets of surveys were completed: 1) 12 Novem- 
ber-23 December 1991 and 17 April-21 May 1992; 
2) 14 November-22 December 1992 and 1 May-4 
June 1993. The survey area (275,356 km 2 ) incorpo- 
rated depths of 300-800 m south of Puysegur Point, 
excluding rough ground and the Bounty Platform 
(Fig. 1). The surveys used a two-phase random strati- 
fication design (Francis, 1984), and because hoki tend 
to be near the seabed during daylight, coming up off 
the bottom to feed at night (Kerstan and Sarhrage, 
1980), trawling was carried out during daylight hours 
only. Trawling procedure and standardization of gear, 
station, and stratum details are reported for each 
survey individually by Chatterton and Hanchet 
(1994), Schofield and Livingston (1994, a and b), and 
Ingerson et al. (1995). 
Length-frequency samples of about 200 female hoki 
were collected from each tow on a random basis. The 
length-frequency distribution and the total numbers 
of fish were scaled up to the total stratum area by 
using the Trawlsurvey Analysis Program, as de- 
scribed by Vignaux. 4 The scaling was done by assum- 
ing a catchability and vulnerability of 1.0 in all sur- 
veys. Because these assumptions were unlikely to 
be valid, the numbers of fish were used only in a rela- 
tive sense. 
Additional samples of 20 female hoki were collected 
from each tow to measure gonad and total body 
weight, to identify macroscopic gonad stage, and to 
obtain histological samples from ovaries. Otoliths 
were also collected from these fish for ageing. Ma- 
turing hoki in the later stages of vitellogenesis can 
be macroscopically distinguished from resting hoki. 
The ovaries at that point are swollen, and the indi- 
vidual oocytes, visible to the naked eye, are opaque 
and creamy pink. Resting and immature ovaries are 
small and translucent and no oocytes are visible. 
Previous surveys of hoki in April-May (e.g. van den 
Broek et al., 1981) reported few hoki that were suffi- 
ciently developed to be identified macroscopically as 
maturing. We therefore obtained histological samples 
of ovaries as well as recorded their macroscopic ap- 
pearance. The central portion of the ovary from each 
sample was preserved at sea in 8%-10% buffered 
formalin. Samples were later processed to produce 
thin sections that were stained with standard 
haemotoxylin and eosin preparations. 
Histological staging 
In developing a method to distinguish prespawning 
hoki from nonspawning hoki, we classified ovaries 
into the stages given by West (1990). Ovaries were 
classified according to the most advanced oocyte 
present in the ovary. A summary of these stages (as 
given by West) is as follows: 
Chromatin nucleolar 
stage 
Perinucleolar stage 
Yolk vesicle 
(cortical alveoli) stage 
Vitellogenic (yolk) 
stage 
Ripe (mature) stage 
The oocyte is surrounded 
by a few follicle cells and 
contains a large nucleus 
surrounded by a thin layer 
of cytoplasm. The nucleus 
has one large nucleolus 
and several small nucleoli. 
The nucleus has multiple 
nucleoli at its periphery. 
Late perinucleolar oo- 
cytes may have vacuoles 
in the cytoplasm. 
This stage is charact- 
erised by the appearance 
of large numbers of yolk 
vesicles in the cytoplasm. 
They increase in size and 
number to form several pe- 
ripheral rows. The chorion 
is visible at this stage. 
Small yolk granules which 
gradually enlarge until 
they form fluid-filled 
spheres are typical. The 
spheres may eventually 
fuse to form a continuous 
mass of fluid yolk. 
The nucleus may be pe- 
ripheral or may have dis- 
integrated completely. 
4 Vignaux, M. 1994. Documentation of trawlsurvey analysis 
program. NIWA Greta Point Internal Report 225, NIWA Greta 
Point library, Wellington, New Zealand, 44 p. 
Criteria to distinguish prespawners from non- 
spawners were developed from a subsample of ovary 
sections from each survey and from some commer- 
