168 
Fishery Bulletin 95 ( 1 ), 1997 
Table 2 
A summary of effects of year and settlement time 
on the growth of 0 + 
English sole. 
Variable 
Parameter 
Coefficient 
Partial-f 
P (2-tail) 
Intercept 
Constant 
a 
8.44 
3.19 
< 0.01 
Settlement Time (c) 
«i 
43.48 
17.03 
< 0.01 
Year (y p y 2 , y 3 ) 
— 
— 
— 
<0.01 ; 
yi 
a 21 
6.76 
2.61 
< 0.05 
y 2 
a 22 
7.79 
3.58 
< 0.01 
y 3 
a 2 3 
-13.06 
-4.19 
< 0.01 
Slope (Growth) 
PWMD (t,) 
P 
0.43 
18.45 
< 0.01 
Settlement time (c) 
Pi 
-0.11 
-4.04 
< 0.01 
Year (y p y 2 , y 3 ) 
— 
— 
< 0.05 2 
yi 
P 2 I 
0.06 
2.13 
< 0.05 
y 3 
P 23 
0.12 
3.66 
< 0.01 
1 Based on the result of an F-test, F 3 I5 = 4.35. 
2 Based on the result of an F-test, F 3 I5 = 13.25. 
Model 1 : q 2 = 0, P t = P 0l e 
= \ 
42.82 
1985 
21.02 
1986 
66.63 
1987 
40.40 
1988 
-0.0066 f -0 00561, _o 0 056(, 
Model 2 :P t = e 00056 ] P e oi \P ol = 
43.80 
18.62 
73.33 
40.46 
1985 
1986 
1987 
1988 
Model 3: P, 
n -0.0175 f, 
P o, e 
P,,e 
-0. 0 175x62^, -0.0075U, -62). 
t, = 0 to 62 (July and August) 
t; > 62 (September) 
(49.60 
P.. = 
23.23 
74.33 
47.28 
1985 
1986 
1987 
1988 
The estimated instantaneous mortality rates of 0 + 
English sole in July and August were equal, there- 
fore model 3 was reduced from a three-step to a two- 
step model. The data fitted model 3 best (Fig. 5) with 
instantaneous mortality rates of 0.0175 per day in 
July and August and 0.0075 per day in September. 
The value of the Bayesian information criterion was 
3.68 for model 1, 4.04 for model 2, and 2.65 for model 
3. As a result, we concluded that model 3 was the 
best for estimating mortality. 
Table 3 
Back-calculated settlement dates with their ranges, assum- 
ing average length at settling, l settling = 20 mm TL. 
Settlement 
Date of 
95% 
Year 
cohort 
settlement 
prediction interval 
1985 
1 
24 Jan 
27 Nov-22 Mar 
1986 
1 
16 Jan 
17 Dec -16 Feb 
2 
10 May 
21 Apr -29 May 
1987 
2 
8 May 
30 Apr -16 May 
1988 
1 
18 Mar 
17 Feb -16 Apr 
Discussion 
Gear efficiency 
Our estimates of growth and mortality might be bi- 
ased if gear selectivity varied with size. Edwards and 
Steele (1968) suggested that beam trawl efficiency 
depends on a number of factors, such as towing speed, 
bottom type, and fish size. At a speed of 35 m/min, 
the efficiency of their 2-meter beam trawl was 25- 
35%, depending on fish size. They point out that their 
results apply only to their particular gear and the 
special conditions in Loch Ewe. Kuipers (1975) found 
that the efficiency of a 2-m beam trawl in the Dutch 
Wadden Sea declined from 100% at lengths below 70 
mm to 15-30% for plaice larger than 150 mm. Our 
gear was a 3-m beam trawl with effective fishing 
width of 2.3 m, wider than the gear used by either 
Kuipers or Edwards and Steele, and was towed faster 
(41-47 m/min vs. 30-35 m/min). Also the ratio of fish- 
