68 
Fishery Bulletin 106(1 ) 
constituting the Actaeon region samples were within 
relatively low productivity areas that may be consid- 
ered to be atypically low for the Actaeon area (see the 
seasonal analysis). 
Table 4 
Seasonal growth-description parameter estimates. All 
parameter definitions are given with Equation 1 under 
the heading “Growth model” in the Materials and meth- 
ods section. In the Actaeon Island estimates L| 0 and L s 95 
reached the limits placed on the parameters. At Sterile 
Island, the parameter estimates obtained did not change 
when the L' | 0 and L | 5 estimates were replaced with 209 
and 210, respectively. In all cases the logistic reduction 
in the variance was negligible. The -veLL is the nega- 
tive log-likelihood. Productivity, in kilograms, is the rela- 
tive productivity derived from the respective transition 
matrix. 
Parameter 
Actaeon Island 
Sterile Island 
MaxAL 
21.8464 
20.0111 
^50 
129.9702 
121.7953 
T m 
^ 95 
162.1723 
161.4016 
Maxa L 
6.7129 
6.8766 
^50 
209 
186.9362 
210 
192.3000 
C 
0.1123 
0.0907 
P 
0.1263 
0.1561 
C, 
-0.0821 
-0.0489 
Productivity Kg 
484.2 
419.4 
-veLL 
1249.673 
944.564 
Oct- Oct- Oct- Oct- Oct- Oct- Oct- Oct- Oct- Oct- Oct- 
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 
Figure 6 
A comparison of the implied seasonal growth curves at 
the Actaeon and Sterile Islands in southern Tasmania. 
The fine horizontal dashed line at 138 mm is the current 
legal minimum length. The curves remain approximately 
linear until about 100 mm shell length. The slowest growth 
periods line up approximately with October each year. 
Seasonal growth descriptions 
Seasonal growth descriptions were fitted to data from 
two sites in southeast Tasmania: Actaeon Island and 
Sterile Island (Table 1; Fig. 2). In each case, when the 
inverse-logistic curve describing the standard deviation 
of the residual errors was fitted, the estimates of L| 0 
and L s 95 were both much larger than the maximum size 
observed. In effect, this result implied that the vari- 
ability was best described with Equation 7 rather than 
Equation 3. For this reason, the 7-parameter model was 
preferred to the 9-parameter model, even though the 
optimum fit in each case lead to the same results in 
terms of parameter estimates and log-likelihood. The 
two sites are very close together geographically (~2 km 
apart) and, in this case, the parameter estimates were 
similar between the two sites (Table 4). 
By assuming a starting size of 0.25 mm on the 
25 th November (a typical size and date of settle- 
ment for a newly settled blacklip abalone) in an 
arbitrary year, and by calculating the expected 
size increment in a series of 8-day steps forward 
in time, adding those increments to the initial size 
and then repeating the process, it was possible to 
visualize the seasonal growth of animals (Fig. 6). 
The main difference in the parameters is in the L^ 0 , 
which was 8.0 mm larger at Actaeon Island (Table 4). 
This difference led to the abalone at Actaeon Island 
reaching the current minimum legal length of 138 mm 
in approximately eight years, whereas the difference led 
to the current minimum legal length in nine years, on 
average, at Sterile Island (Fig. 6). 
The Actaeon Island site was far more productive than 
Gagens Point and the Middle Ground, both in the Ac- 
taeon region and, in fact, was as productive as both 
the Bruny Island and southwest regions (Tables 2 
and 4). The Sterile Island site had a productivity 
that was only 89% of that of the Actaeon Island site 
(431 kg vs. 485 kg; Table 4), this was also reflected 
in a one year difference in time to legal minimum 
size (Fig. 6). 
The pattern of seasonality was very similar be- 
tween the two sampled sites; linear-like initial 
growth proceeded at least until 100 mm shell length. 
The phase parameter, p, value of 0.118 implies that 
the fastest period of growth occurred on 11 th Febru- 
ary at the Actaeon Islands, whereas at Sterile Island 
the estimate of 0.156 indicates the fastest growth 
occurred two weeks later on 25 th February. 
Discussion 
Comparison of inverse-logistic, von Bertalanffy, 
and Gompertz models 
The major difference between the inverse-logistic 
growth description, the von Bertalanffy growth curve, 
and the Gompertz growth curve is seen in the model 
fit at the extremes of the growth trajectory. Without 
