340 
Fishery Bulletin 99(2) 
8000 
A 
9,2 
B 
7000 
9 
9 
9 
8,8 
9 
6000 
9 
8,4 
• 
5000 
'' - ^ 9 
9 9 
CO 
g 4000 
© 
& 8,0 
CD 
3000 
9 
9 
* . * • 
7,6 
• * - 
9 
2000 
. * * * • • . * 
9 
• • 
7,2 
m • 
1000 
9 
6,8 
1970 1976 1982 1988 1994 2000 
3,3 
3,4 3,5 3,6 3,7 3,8 
3,9 4,0 
Year 
Ln(f) 
7500 
c 
9,0 
D 
6500 
• 
8,6 
9 
5500 
8,2 
9 
9 
4500 
cr 
* - * . • 
"co 
Cl 
• 
1 7,8 
9 
^ 3500 
9 9 
C 
9 
9 
9 
9 
9 
7,4 
9 * 9 
9 * • 
2500 
• 
. 
9 
9 
1500 
9 • 9 9 9 ® • 
* *9 
• 9 
7,0 
9 
6,6 
1970 1976 1982 1988 1994 2000 
3,3 
3,4 3,5 3,6 3,7 3,8 
3,9 4,0 
Year 
Ln(f) 
Figure 1 
Relation between average effort per stock fraction (epsf) per vessel in the fleet and year (A), epsf for an individual vessel and year 
(C), log- 
transformed epfs and time (in ln( years)) for the average 
vessel in the fleet (B) and for an individual vessel (D 
). Note that 
the first observed time value in B and D is ln( 27 ) because 1946 is 
assumed to be the first year in the learning process. The slope and 
intercept for the fitted lines and the correlation between the van 
ables in B and D 
are given in the text. 
Discussion 
A learning curve ideally requires data points from the 
init ial phase of the process. For many fish st ocks, catch and 
effort data, together with an independent estimate of total 
biomass, exist only for recent years. Given the approximate 
year the fleet entered the fishery, it is necessary to assume 
that epsf follows the learning curve pattern from that year 
onwards. Under this assumption, it is possible to adjust 
effort backwards in time to the level of one year in the 
time series by using the estimated learning curve. Tech- 
nical revolutions, such as the introduction of hydraulic 
wires, may cause a very dramatic change in the fishery, 
and if this happens the learning curve should be esti- 
mated from the time when the new technology appeared. 
Good knowledge about the history of the fishery is thus 
necessary. There are also alternative methods for dealing 
with increases in efficiency with time in a fishery. Gulland 
(1983) suggested constant monitoring of the changes in 
the fishing gears by conducting experiments. This solution 
may, however, be very expensive. 
The question whether the effort should be adjusted with 
a learning curve from the fleet level or from the vessel 
level depends on the resolution of the catch and effort da- 
ta and on other standardization of effort. If effort is first 
adjusted within the fleet because of the individual differ- 
ences in fishing power, adjustment of the effort due to 
learning should be based on a learning curve from the 
standard vessel or the group of standard vessels to avoid 
double standardization. It is important to be aware that 
individual vessels also may show different learning rates 
and these may differ from the average learning rate of the 
fleet. An individual vessel may improve its efficiency by in- 
creasing the cooperation with other vessels, by buying bet- 
ter searching equipment and more efficient fishing gear, 
and by hiring more skilled crew. 
