448 
Fishery Bulletin 96(3), 1998 
used to survey Victorian abalone stocks, about 20- 
28 samples were completed daily; the radial transect 
method provides 27 samples during a typical day and 
requires only one-third the number of ascents per 
diver, reducing the risk of decompression illness 
(Marks and Fallowfield, 1994; Oxer, 1994). As with 
timed searches, the radial transect method has been 
successfully applied under rough sea conditions typi- 
cal of the exposed coastal reefs colonized by abalone. 
McShane and Smith ( 1990) identified the similar- 
ity between the timed search method and the tech- 
niques employed by commercial abalone divers to 
target aggregations of abalone. From this similarity 
one can reasonably infer that in some instances the 
abalone collection rate of research divers may exhibit 
(albeit over a smaller spatial scale) the hyperstability 
that characterizes the catch per unit of effort of com- 
mercial abalone divers. Indeed, in a study involving 
intensive experimental fishing McShane and Smith 
(1989) could not detect a significant difference be- 
tween pre- and postfishing abundance estimates of 
postrecruited abalone from timed searches despite a 
50% decrease in CPUE. In his discussion of a patch- 
frequency estimation method, McShane (1995) noted 
the bias in timed searches when a large proportion 
of the search time is spent collecting abalone from 
dense aggregations. Radial transects target a spe- 
cific area (30 m 2 ); consequently handling time does 
not affect the numbers of abalone counted or collected 
and divers are unable to target aggregations to the 
exclusion of more sparsely distributed abalone. 
Therefore it is reasonable to expect that surveys 
based on radial transects more accurately reflect true 
abalone abundance than do those based on timed 
searches. 
Although there is no doubt that dense patches of 
abalone are important with respect to fishing and 
the maintenance of profitable catch rates (McShane, 
1995), recent work on H. rubra in Victoria (Officer 1 ) 
has demonstrated the importance of postfishing 
movement and reaggregation in maintaining these 
patches. The Victorian studies also showed that dense 
patches contributed most of the variation in abun- 
dance estimates and that the standard deviations of 
samples of sparsely distributed abalone were rela- 
tively small. It is from these sparsely distributed 
abalone, and those occupying cryptic habitat, that 
reaggregation appears to have occurred. Further re- 
search is required to determine the relative impor- 
tance of sparsely distributed abalone in assessing the 
impact of fishing. Abundance estimates from transect 
1 Officer, R. 1997. Marine and Freshwater Resources Institute, 
PO Box 114, Queenscliff, Victoria 3225, Australia. Unpubl. 
manuscript. 
sampling may prove to be more effective indicators 
of the size of stocks if dense aggregations are excluded 
from surveys. 
Because the application of radial transects avoids 
targeting some emergent abalone to the exclusion of 
others, there is less potential for divers to bias their 
sample towards larger abalone as may occur with 
timed searches. By measuring the length of each 
abalone collected, separate estimates of abundance 
for prerecruit (shorter than the legal minimum 
length) and postrecruit (recruited to the stock) aba- 
lone can be made. Timed searches do not necessarily 
permit this separation of prerecruits from post- 
recruits because of the potential for divers to collect 
larger, more accessible abalone at the expense of 
smaller abalone. Timed search estimates of abun- 
dance reported in McShane and Smith (1989) show 
an almost threefold increase in prerecruits after fish- 
ing, suggesting that abundances of prerecruits may 
have been underestimated prior to fishing because 
of the prevalence of postrecruits. 
Our trial stock surveys provided evidence that the 
transect method has the power to detect differences 
in abalone abundance (among locations and sites) and 
is robust in respect to differences in divers’ abilities 
to perform the method. Moreover, annual stock sur- 
veys based on radial transects were able to detect a 
significant change in abundance between consecu- 
tive years and had a high power of detecting smaller 
effect sizes over 2-3 years, thus showing their use- 
fulness as an effective stock monitoring tool. The 
radial transect method as it is currently applied for 
monitoring Victorian abalone stocks should have suf- 
ficient power to detect changes in abundance of about 
10% after 3 years sampling. Monte Carlo simulations 
showed that surveys based on timed searches gener- 
ally would require several years of sampling and con- 
sequently would have lower power than radial 
transect surveys in detecting the same rate of change. 
The simulation results for the eastern zone, where it 
was predicted that only one year would be required 
for a 2.5% change, proved to be the only exception. 
However, unlike the timed search sampling of the 
central and western zones, all sites in the eastern 
zone were sampled twice annually during 1989-91. 
This additional sampling would be expected to in- 
crease substantially the statistical power to detect 
changes in abundance. 
Surveys based on radial transects have been used 
to monitor H. rubra stocks in the Victorian abalone 
fishery since 1992 with the objective of establishing 
a temporal series of abundance data to assist man- 
agers in determining sustainable levels of fishing. 
In monitoring Victoria’s abalone stocks, the aim is to 
detect an overall change in abundance across a man- 
