120 
W. G. O’CONNOR AND J. D. KOEHN 
sides of up to 3.7 fish/m 2 in tributaries of Lake 
Eucumbene. Such densities are likely to be at¬ 
tributable to the presence of high-quality habitat 
and the absence of brown trout, Salmo trutta. 
The effects of S. trutta on the abundance and 
distribution of G. olidus have been comprehen¬ 
sively documented, with mutually exclusive 
populations often occurring and a fragmen¬ 
tation of the range of G. olidus resulting in iso¬ 
lated populations often being reported (Tilzey 
1976, Cadwallader 1979, Fletcher 1979, Jackson 
& Davies 1983, Cowden 1988, Jones et al. 1990, 
Lintermans & Rutzou 1990). 
After using otoliths to age G. olidus , Cowden 
(1988) concluded that most individuals did not 
reach maturity until their third year (age 24-). 
Drayson (1989), however, examined otoliths 
from G. olidus collected in the same catchment 
as Cowden and considered their annular pat¬ 
terns to be uneven and therefore unsuitable as a 
means of accurately aging this species. From the 
length frequency distribution (Fig. 2), we suggest 
estimated age classes 04- (1st year), 14- (2nd 
year), 2 -f (3rd year) and 3+ (4th year) based on 
a similar population structure reported by 
Fletcher (1979). These three age classes make up 
almost the entire population (97%) of G. olidus 
in Bruces Creek. Most G. olidus did not mature 
in their first year but all were mature by their 
second year. These results are consistent with 
those of Fletcher (1979), namely that maturity is 
mostly reached at age (1+) and that the maxi¬ 
mum age is probably 4 years. 
Although the sex ratio for sexually mature fish 
is 1:1, the ratio is not consistent throughout the 
age classes. Males mature earlier than females 
but appear to have a higher mortality; 
consequently females predominate in the 24- 
and 34- age classes. Cowden (1988) also re¬ 
ported a 1:1 sex ratio. 
The fecundity of G. olidus is one of the lowest 
among galaxias species in Australia. A mean fec¬ 
undity of 198 found in this study is similar to the 
value of 243 recorded by Cowden (1988). Of 
Victorian galaxiid species, only Galaxiella 
pusilla , a wholly freshwater species, has a similar 
fecundity (generally 100-200; Humphries 
1986). Fecundities of diadromous galaxiid 
species are generally much higher: common 
galaxias, Galaxias maculatus , up to 13,500 
(McDowall 1968); spotted galaxias, G. trutta- 
ceus , up to 16,000 (Humphries 1989); and 
broad-finned galaxias, G. brevipinnis , up to 
23,000 (Koehn and O’Connor unpubl. data). 
The breeding season of G. olidus has been 
variously reported as winter through to summer 
(Cadwallader & Backhouse 1983). G. olidus 
spawns in October in Victorian alpine streams 
(Fletcher 1979); from early August to early Sep¬ 
tember in the Australian Capital Territory; and 
during late August to early September in 
southern Queensland (Marshall 1989). The vari¬ 
ation in onset of these spawning seasons may be 
due to the corresponding differences in water 
temperatures. Cowden (1988) recorded tem¬ 
peratures of 6.5-8.2°C, whereas Marshall (1989) 
estimated a temperature range of about 15- 
20°C. In Bruces Creek the main breeding season 
extended for over 2 months from early August to 
about mid October, when the temperature range 
was 8.0-10.2°C. 
G. olidus employed a spawning strategy in¬ 
dicative of a fish species that has a low fecundity 
and needs to maximise the survival of eggs and 
larvae. This strategy involves laying a small 
number (average fecundity of 198) of relatively 
large (2.3 mm) adhesive eggs in a protected site, 
usually a boulder more than 180 mm in diameter 
and with a narrow gap between the underside of 
the boulder and the streambed. Similar 
spawning sites for G. olidus have been reported 
by Cowden (1988). In Bruces Creek the eggs 
were attached to the underside of the boulders 
where they may be protected from possible pred¬ 
ators (especially fish), disturbance, strong water 
currents and possible smothering by sediment. 
All the boulders were in riffles where the sur¬ 
rounding water was relatively fast-flowing (0.2- 
0.5 m/sec) and well oxygenated. Eggs were not 
found attached to any other instream objects 
such as wood debris. 
Red bruising on the abdomen and sides of 
many spawning fish of both sexes suggests that 
the fish rubbed against hard objects during 
spawning. Bruising is highly likely in enclosed 
locations where the fish would probably have to 
press hard against the underside of boulders in a 
lateral or upside down position to deposit eggs 
and milt. 
A least three different stages of embryonic 
development were recorded for eggs at one 
spawning site. Because females appear to lay all 
their eggs at once, it appears that more than one 
spawning may take place at a suitable site. The 
strategy of laying a small number of relatively 
large, adhesive eggs at an enclosed site has been 
described for another Victorian freshwater fish, 
the freshwater blackfish, Gadopsis marmoratus 
(Jackson 1978). G. marmoratus attaches its eggs 
to the inside of hollow logs which are then 
guarded by an aggressive male fish (Koehn 
unpubl. data). Parental care of eggs appears un- 
