SPAWNING OF GALAXIAS OLIDUS IN BRUCES CREEK 
121 
likely for G. olidus as relatively few fish were 
collected from the riffle areas and there was no 
evidence of eggs being guarded. We surmise that 
ripe males and females from the pools move into 
the riffles to spawn, then return to their pre¬ 
ferred habitat. The collection of eggs in drift nets 
from both disturbed and undisturbed sites indi¬ 
cates that not all eggs are attached to the 
spawning site. Such “free” eggs may originate in 
three ways. 
1. Eggs not becoming successfully attached to 
the spawning site during spawning. This may 
be expected when fish are attempting to at¬ 
tach and fertilise eggs on the underside of a 
boulder. There appears to be a large discrep¬ 
ancy between female fecundities and the 
number of eggs found attached to boulders, 
suggesting that not all eggs become at¬ 
tached. 
2. Eggs being laid at less sheltered sites in the 
stream where attachment does not occur or 
where the eggs may easily be displaced. There 
is, however, no evidence for such sites. 
3. Eggs being dislodged from the spawning 
boulder. This appears unlikely as the eggs ad¬ 
here strongly to the rock and to each other. 
Some dislodgement may be possible, how¬ 
ever, if subsequent spawning activity occurs 
at the same site. 
The collection of single, non-adhesive, de¬ 
veloping eggs at undisturbed sites suggests that 
some eggs may be carried downstream either by 
drifting in the current or by rolling along the 
streambed. Such a mechanism would aid 
downstream dispersal. G. olidus eggs are initially 
very adhesive, but those at a later stage of devel¬ 
opment that were unattached were non¬ 
adhesive. Some of these eggs probably lodge in 
the interstices of the substrate in riffles or settle 
to the streambed in pools. Such eggs were col¬ 
lected by disturbing the substrate. 
Hence, the eggs of G. olidus may be found in 
three situations: (a) attached, (b) lodged in the 
substrate, or (c) drifting. Each of these situations 
involves a different degree of disturbance to the 
egg itself, possibly leading to variations in hatch¬ 
ing times. Cowden (1988) found that hatching of 
eggs could be induced by gently swirling the 
water in which they were held. The induced 
hatching of eggs of G. truttaceus and G. brevipin- 
nisby disturbance has also been noticed (Koehn 
& O’Connor unpubl. data). Cowden (1988) 
recorded, however, that “induced” fry were 
smaller and less active than normal. 
Exact hatching times were not obtained, but 
fertilized eggs collected in the stream at an early 
embryonic stage took 21 days to hatch at tem¬ 
peratures of 12.9° to 14.8°C. Using Cowden’s 
(1988) results, hatching times were estimated to 
be 47 days at about 9°C and 32 days at about 
13°C. Average temperatures in Bruces Creek 
from August to mid October were about 9°C, 
while temperatures in late October were about 
13°C. Assuming hatching is temperature depen¬ 
dent, these figures suggest that hatchings in 
Bruces Creek probably occurred from early 
October until mid December. 
Newly hatched larvae emerged at 9.4 mm TL 
with relatively small yolk sacs which were fully 
absorbed after about 5 days, with feeding com¬ 
mencing after a further 3 days. Large, well 
developed larvae with small yolk sacs are more 
mobile, able to seek shelter and hence less vul¬ 
nerable than smaller fry with large yolk sacs. 
This is consistent with the spawning strategy we 
have described for G. olidus. 
Increased sediment input into streams has 
been recognised as having adverse effects on 
Victorian native freshwater fish (Cadwallader & 
Backhouse 1983, OCE 1988, Mitchell 1990, 
Koehn & O'Connor 1990a, 1990b). Sedimen¬ 
tation would be likely to fill or cover potential 
spawning sites, and the availability of sites un¬ 
der boulders appears essential for spawning. If 
such sites were not available or were heavily 
silted so that eggs do not adhere, spawning may 
not occur or may have reduced success. Like¬ 
wise, the loss of interstitial spaces in the sub¬ 
strate due to sedimentation may cause increased 
mortalities to unattached eggs which normally 
become lodged there. Eggs of Gadopsis marmo- 
ratus, Galaxias truttaceus and G. brevipinnis 
have shown high mortalities when covered with 
light layers of silt (Koehn & O’Connor unpubl. 
data). 
Sedimentation and the well-documented 
threat of predation by introduced species are 
likely to be major environmental threats to G. 
olidus in otherwise natural streams. 
ACKNOWLEDGEMENTS 
We thank Damien O’Mahony, Matt Westaway 
and Melinda Millar for technical assistance, and 
Tim Doeg, Tarmo Raadik, Sandy Morison, 
Andrew Sanger and Darwin Evans for valuable 
comments on the manuscript. Tim Doeg’s tran¬ 
sition in interest from bugs to fish has been an 
inspiration to all. Artwork was completed by 
Justin O’Connor and wordprocessing by Kae 
Winch. This work was completed as part of the 
