72 
Fishery Bulletin 107(1 ) 
Achoerdus gouldii undergoes the type of color change 
that is broadly associated with sex change in most pro- 
togynous labrids (e.g., Roede, 1972; Gillanders, 1995a; 
McBride and Johnson, 2007) and which, in A. gouldii, 
involves a shift from green to blue rather than from red- 
dish brown to blue as in A. viridis (Gillanders, 1999). 
However, as with the latter congeneric species, some fe- 
males (3%) were not of the initial color and some of the 
males (11%) did not have the terminal color. Further- 
more, the continuous variable, length, was found to be 
a better predictor of sex than the dichotomous variable, 
color; however, a combination of both of these variables 
further improved one’s ability to predict the sex of indi- 
viduals of A. gouldii. The fitting of logistic curves to the 
length data for fish with testes and with blue coloration 
yielded TL 50 s of 821 and 779 mm, respectively. Although 
these two TL 50 s differed by 42 mm, their 95% confidence 
limits overlapped and therefore color can be used to de- 
rive an approximate value for the TL 50 at sex change 
when it is not possible to record the sex of individuals 
because, for example, their viscera had been removed 
or fish were being viewed live during visual surveys. 
Our approaches to using color as a proxy for sex for 
estimating the TL 50 at sex change and for enhancing 
the description of the growth of females and males are 
likely to be applicable to many of the numerous species 
that exhibit a similar form of sexual dichromatism. 
Recruitment variability, mortality, 
and spawning potential ratio 
The pattern of flow of the Leeuwin Current, the pre- 
dominant current on the southwest coast of Australia 
in winter and therefore during the spawning period of 
A. gouldii, leads to the larvae of certain teleosts and the 
western rock lobster (. Panulirus cygnus) being dispersed 
offshore (Pearce and Phillips, 1988; Gaughan, 2007). 
Thus, because A. gouldii is recruited into nearshore 
waters, it appears relevant that this species exhibited 
particularly strong recruitment in 1972, 1980, 1983, 
and 1990, when the Leeuwin Current was weak (Pearce 
and Phillips, 1988; Caputi et al., 1996), and very poor 
recruitment in 1991 and 1992 when the Leeuwin Cur- 
rent was especially strong. 
Because the current level of fishing mortality for 
A. gouldii in southwestern Australian waters is esti- 
mated to be 74% of natural mortality, this species is 
apparently close to or at full exploitation in these wa- 
ters (after applying a reference point F Um = 2 /3 M based 
on Patterson [1992]). The conclusion that A. gouldii is 
close to or at full exploitation is consistent with the 
estimate for the spawning potential ratio (SPR) for the 
males of this protogynous hermaphroditic labrid. This 
value has therefore declined to 0.52 and, given the 
steepness of the curve relating SPR to fishing mortality, 
is rapidly approaching 0.30, which is often regarded as 
the level at which a stock is considered to be overfished 
(Goodyear, 1993; Mace and Sissenwine, 1993). Indeed, 
the lower 95% CL of 0.27 for the SPR lies below this 
reference point. 
Implications for management 
Most recreational line and spear fishing for A. gouldii on 
the south coast of Western Australia occurs in shallow 
and relatively accessible waters, where the individuals 
of this species are typically smaller than those in deeper 
waters and where most individuals are less than the 
TL 50 of 653 mm at which females attain maturity. Fur- 
thermore, in deeper waters, 52% of the A. gouldii taken 
by the commercial fishery were less than the length at 
maturity and 88% were below the TL 50 of 821 mm at 
which females change sex to males. Moreover, A. goul- 
dii often suffers barotrauma when brought to the sur- 
face from particularly deep waters, as is the case with 
many other demersal species in Western Australia and 
elsewhere (e.g., St John and Syers, 2005; Parker et al., 
2006), and with other labrids (e.g., Nardi et al., 2006). 
Thus, if the current minimum legal length of 500 mm 
is maintained, or slot (minimum and maximum length) 
limits are introduced, fishing crews should be encour- 
aged to adopt fishing practices that minimize the loss of 
released fish. In this context, it would be of great value 
to undertake research into the effects of barotrauma on 
A. gouldii and the ways in which the loss of released fish 
may be minimized. 
Although substantial fishing of protogynous species 
can lead to a marked decline in the relative abundance 
of their males (e.g., Coleman et al., 2000), the size at 
sex change of some protogynous species declines in re- 
sponse to the selective removal of large males by fishing 
(e.g., Platten et al., 2002; Hawkins and Roberts, 2003). 
Because it is not known whether the size at sex change 
of A. gouldii is labile, it would be prudent for managers 
to take the conservative view that this is not necessar- 
ily the case with this labrid. 
Because fisheries for protogynous species, such as 
A. gouldii, have a greater effect on the spawning bio- 
mass of males than females, managers need to take 
into account the potential for fishing to lead to sperm 
limitation, reduced fertilization success, and social or 
behavioral changes, and to consider whether appropri- 
ate biological reference points need to be established 
to ensure that the sex and size structure of the fish 
stock is maintained (Alonzo and Mangel, 2004, 2005; 
Brookes et al., 2008). Because it would be useful for 
managers to have a rapid and inexpensive means for 
determining the sex of individuals of A. gouldii, our 
data are relevant in that they show that body color 
can be used as a broad surrogate for sex and it would 
enable rapid monitoring of the population to detect 
whether changes are occurring in the ratio of the bio- 
masses of mature females and males and in the length 
at sex change, both of which could be used as the 
basis for an appropriate fishery control rule. In our 
estimates of the effect of increased fishing mortality 
on male spawning biomass per recruit, we did not 
assume that compensatory phenotypic or behavioral 
responses of the types explored by Alonzo and Mangel 
(2005) occurred in the pattern of sex change of A. 
gouldii. Consequently, our results provide a precaution- 
