Shi et al.: Growth and survival of Pteuronectes vetulus 
163 
ha) (Shi et al., 1995). Within each stratum, stations 
were randomly selected from sampling units super- 
imposed on nautical charts, with the constraint that 
no two stations were immediately adjacent to one 
another. The effort (number of stations) allocated to 
each stratum was proportional to the abundance of 
English sole in that stratum (Shi et al., 1995). 
The nearshore area was sampled along fixed 
transects oriented east-west and trawl stations were 
located at discrete depths (Fig. 1). Five transects were 
established, and sampling stations were located at 
depths of 5, 9, 18, 27, 36, 46, and 55 m. The 55-m 
station was not sampled on the northernmost 
transect because of frequent gear damage at this lo- 
cation. Additional effort was allocated to the inter- 
mediate stratum; two trawl samples were taken at 
all 27- and 36-m depths. Sampling stations were 
stratified according to depth to obtain population 
estimates. The outer boundary for the nearshore 
study area was the 59-m (32.5-fm) isobath, and the 
mean low low water (MLLW) mark was the inner 
boundary. The boundary separating the inner and 
middle strata followed the 14-m (7.5-fm) isobath, 
whereas the boundary between the middle and outer 
strata was located at 41 m (22.5 fm). The northern 
and southern limits of the survey area were posi- 
tioned 5 km beyond the northernmost and southern- 
most transects. 
Each of the three areas was visited once a month. 
Sampling in estuaries was planned during low spring 
tides of the month (April or May through Septem- 
ber) so that we could navigate among unmarked 
channels, which otherwise are difficult to see. Sta- 
tions in close proximity to intertidal areas were 
sampled preferentially at low tide to minimize bias 
associated with fish movement onto the tideflats at 
higher stages of tide. More exposed sites were typi- 
cally sampled at high water. Trawling operations 
ceased when tidal currents were judged sufficiently 
strong that the trawl gear would not tend the ocean 
bottom properly. Nearshore sampling trips were usu- 
ally made between the two estuary trips in that 
month. 
Survey samples throughout the study area were 
collected with a 3-m beam trawl specifically devel- 
oped for this study (Gunderson and Ellis, 1986). Ef- 
fective width of the net was 2.3 m, whereas the esti- 
mated vertical opening was 0.6 m. The body of the 
net was composed of 7-9 mm (lumen) knotless ny- 
lon, and the codend was lined with 4-mm stretch 
mesh. A double tickler chain array was attached to a 
9.5-kg wingtip weight at each corner of the net. The 
tickler chain array, together with the turbulent zone 
it creates, dislodges small animals from the sub- 
strate, thus promoting capture by the net. 
Nearshore sampling was conducted from the 17- 
m stern trawler F/V Karelia. Tows in the nearshore 
were taken parallel to isobaths. Scope was routinely 
5:1, except at the 5 and 9 m stations where it was 8:1 
and 9:1, respectively. Time on the ocean bottom was 
estimated by using a trigonometric relationship be- 
tween water depth and wire out, whereas the linear 
distance towed (mean: 750 m) was determined from 
LORAN-C readings. Tow duration was routinely 20 
minutes at a mean towing speed of 2.6 km/hr (1.4 
knots), except at the 5- and 9-m stations, which of- 
ten yielded excessive quantities of sand dollars 
( Dendraster excentricus) and gravel; tows in these 
areas were limited to 5 or 10 minutes. 
A 6.4-m Boston whaler with a 150-hp outboard 
engine was used for estuarine trawling. Buoys were 
deployed at the points where the net first contacted 
the bottom and subsequently left bottom upon re- 
trieval. The distance towed (mean: 260 m) was esti- 
mated with an optical rangefinder. Mean towing 
speed was 2.8 km/hr (1.5 knots), comparable to that 
used in the nearshore area. 
Data analysis 
Length Growth rate estimates were obtained by 
regressing the mean length of a recruitment influx 
(indicated by a mode in the length-frequency distri- 
bution [Shi et al., 1995]) against the time when 
samples were taken. There was a linear relationship 
between modal length and time, as was the case in 
previous growth studies on juvenile English sole 
(Ketchen, 1956; Kendall, 1966; Rosenberg, 1982). 
Because size-dependent migration between near- 
shore and estuarine systems occurs, with smallest 
juveniles migrating into estuaries and larger fish 
moving offshore (Gunderson et al., 1990; Shi et al., 
1995), separate estimates of growth rates for 
nearshore and estuarine fish would be inappropri- 
ate. To minimize the effect of interregional migra- 
tions, the mean lengths at each mode (defined on the 
basis of visual inspection of monthly length frequency 
plots [Shi et al., 1995]) were calculated from the es- 
timated size composition of the overall population. 
The length statistic used was the mean modal length 
(MML), which is defined as the mean length within 
a mode, weighted by the estimated population size 
for each size group: 
MML = 
(1) 
