SPAWNING OF GALAX1AS OLIDUS IN BRUCES CREEK 
119 
Eggs were found adhering strongly to the 
underside of boulders larger than 180 mm in 
diameter situated in riffles immediately up¬ 
stream and downstream of pools. Eggs were not 
found attached to wood debris or other instream 
objects. 
The underside of each boulder was relatively 
flat but slightly raised (up to 60 mm) from the 
streambed so that the eggs were sheltered from 
the main water current. Water velocities in the 
rifflesaround the boulders ranged from approxi¬ 
mately 0.2 to 0.5 m/sec. On two boulders which 
had not been disturbed 179 eggs and 26 eggs 
were counted. Eggs were attached either singly 
or in groups of 6 to 62; the eggs were usually one 
and not more than two layers thick and were 
coated with sand and gravel particles. The 
boulder with 179 eggs contained eggs at three 
different stages of embryological development. 
In four drift nets set for 3 hours below undis¬ 
turbed sites only one egg was collected. A repeat 
set of that net 2 days later at the same site col¬ 
lected one more egg. The net had been pos¬ 
itioned in the main gently flowing current (0.1 
m/sec) about 1.5 m downstream of a riffle. No 
eggs were collected in two nets set overnight for 
22 hours. 
When substrate or instream objects such as 
wood debris were disturbed by gentle lifting two 
eggs were caught in one drift net and nine in 
another held immediately downstream of two 
riffles. The substrate of these two riffles com¬ 
prised boulder 10%, cobble 40%, pebble 30% 
and gravel 20%. One egg was collected from a 
pool with a “flat” streambed whose substrate 
comprised sand/silt 70%, boulder 10%, and 
cobble 10%. No eggs were found in four other 
pools and three other riffles “searched” in the 
same manner. 
We surmise that most eggs of G. olidus are 
attached to the underside of boulders, but some 
eggs come to rest in the substrate and a few drift 
downstream. 
% description and incubation period 
Unfertilised oocytes were spherical, demersal, 
l ^ eoro PaQue and adhesive. Fertilized, water 
ar ened eggs were spherical, demersal, trans- 
^ rcn J an ^ initially adhesive but single, un- 
hesne ^ ^ ecame P^S^ssively less ad- 
he diameters of oocytes and fertilized eggs at 
e earliest stages of development and at the 
eyed stage were: 
Mean oocyte diameter = 2.29 mm (N = 50, SE 
= 0.22 mm) 
Mean egg (early stage) diameter = 2.4 mm (N = 
8, SE = 0.11 mm) 
Mean egg (eyed) diameter = 2.29 mm (N = 20, 
SE = 0.20mm). 
Three eggs collected on 5 September 1990 
were at the earliest stage of embryological devel¬ 
opment (i.e. no embryo in nucleus) and took 21 
days to hatch at temperatures of 12.9-14.8°C. 
All other viable eggs collected were at later stages 
of development and hatching times were 9-14 
days at 14.0-15.0°C. 
Larvae 
The lengths of larvae measured 0-6 hours after 
hatching ranged from 9.0 mm to 9.8 mm TL. 
Mean length was 9.39 mm (N = 16, SD = 0.26). 
The larvae had small yolk sacs 1.4 mm long. In 
the light, these larvae were active swimmers 
throughout the water column including at the 
surface, but they also lay motionless on the bot¬ 
tom of the tank for several minutes. Larvae 
appeared to have absorbed their yolk sacs after 5 
days and commenced feeding 3 days later. 
DISCUSSION 
Our study provided information on G. olidus 
from relatively natural surroundings unaffected 
by major habitat alterations or by other fish 
species including introduced predators. The 
major habitat preference of G. olidus in Bruces 
Creek was clearly for pool areas with slow- 
flowing, deeper water in a stream containing 
abundant instream habitat of wood debris, sub¬ 
merged tree roots and undercut banks and intact 
riparian vegetation. Such attributes in streams 
have been recognised as important for 
freshwater fishes (Cadwallader & Backhouse 
1983, OCE 1988, Lloyd & Walker 1988, Koehn 
& O’Connor 1990a, 1990b). Riparian vegetation 
prevents erosion, provides instream habitat and 
shading, as well as terrestrial invertebrates 
which form an important element in the diet of 
G. olidus (Cadwallader et al. 1980). McDowall 
(1980b) has suggested that removal of riparian 
vegetation correlates with the reduction in 
population numbers of several galaxias species 
in New Zealand. 
The relatively high population densities of G. 
olidus (1.31-1.70 fish/m 2 ) in Bruces Creek are 
comparable to those of 0.79 and 2.83 fish/m 2 
reported by Fletcher (1979) in Watchbox Creek 
and at Mt Buffalo. Tilzey (1976) reported den- 
