GALAXIAS BREVIPINNIS IN NORTH-EASTERN VICTORIA 
23 
Zealand (McDowall 1970). However, land¬ 
locked (and usually lacustrine) populations of 
normally diadromous galaxiids frequently exhi¬ 
bit shifts in spawning period and migration pat¬ 
terns compared with riverine populations and 
migrate upstream in spring to spawn in streams 
(Pollard 1971, Andrews 1982, Humphries 
1989). Populations of G. brevipinnis in north¬ 
eastern Victoria are effectively land-locked 
because of their distance from the sea (2,225 km 
from Lake Hume to the mouth of the River Mur¬ 
ray) and because of the barriers impeding a 
return migration of juvenile fish. It is not known 
whether larvae of G. brevipinnis in north-eastern 
Victoria move downstream after hatching as do 
those in coastal streams, but such a movement 
would lead them eventually to Lake Hume (or 
Lake Mulwala for Kiewa River populations). 
Consequently, recruitment probably depends on 
juveniles which remain upstream of Lake 
Hume. Regular recruitment from the Snowy 
catchment or from the Murray seems unlikely. 
G. brevipinnis has not yet been found in Lake 
Hume and details of any migratory movements 
of these populations are unknown. 
Origin of parent stock 
G. brevipinnis from South Australian popu¬ 
lations near the mouth of the River Murray may 
have moved upstream into north-eastern Vic¬ 
toria, but such a movement would have involved 
an upstream migration of more than 2,000 km, 
past two large dams (Yarrawonga Weir and 
Hume Dam) and 13 smaller weirs. It seems im¬ 
probable that such a migration would be success¬ 
ful in recent times but not during the thousands 
of years prior to European settlement. More¬ 
over, if such a movement had occurred we would 
expect that G . brevipinnis would have been 
found in other tributaries entering the River 
Murray further downstream. 
It is more plausible that G. brevipinnis was 
deliberately or inadvertently released into the 
catchment of the upper River Murray. We do 
not regard G. brevipinnis as a natural part of the 
fish fauna of north-eastern Victoria. There are 
several potential sources and modes of introduc¬ 
tion for the G.brevipinnis now found in this area 
(and these origins are not mutually exclusive). 
Consignments of trout from hatcheries within 
the natural range of G. brevipinnis have been 
released into streams in the upper Murray, and 
these consignments may have included some 
specimens of G. brevipinnis. The species may 
have been illegally stocked into streams, other 
public waters or private waters from which es¬ 
cape occurred. Specimens of G. brevipinnis may 
also have been transported inland for aquarium 
specimens or as live bait and escaped into the 
wild. Evidence for any of these events is lack¬ 
ing. 
The most likely route by which G. brevipinnis 
gained access to north-eastern Victorian waters 
is via the Snowy Mountains Scheme by which 
water is diverted from the upper Snowy River 
catchment to the upper River Murray for irri¬ 
gation and the generation of hydro-electricity 
(Fig. 4). G. brevipinnis has been recorded at sites 
in the upper Snowy River catchment, including 
the Snowy River below Eucumbene Dam 
(McDowall & Frankenberg 1981; collected 
1973, J. Paxton pers. comm.), and four creeks 
draining to Lake Eucumbene (Tilzey 1976, as G. 
coxii = G. brevipinnis). 
Snowy Mountains Scheme 
The volume of water diverted annually by the 
Snowy Mountains Scheme is about 580 GL, but 
this volume may almost double in dry years and 
comprise about one-third of the total inflow to 
Lake Hume (Jacobs 1989). Some of the water 
flows by gravity from the Island Bend Pondage 
on the upper Snowy River to the Geehi Reser¬ 
voir in the River Murray catchment through the 
14.4 km-long Snowy-Geehi tunnel (Fig. 4). Ad¬ 
ditional water flows by gravity from Lake 
Eucumbene to the same tunnel via the 23.5 km- 
long Eucumbene-Snowy Tunnel, or is pumped 
from Lake Jindabyne via the 9.9 km long Jinda- 
byne-Island Bend Tunnel (Snowy Mountains 
Hydro-electric Authority 1982). Fish would 
have little difficulty moving through these tun¬ 
nels: screens on tunnel entrances are large rela¬ 
tive to fish size, and the tunnel gradients are 
generally small (flow can be in either direction in 
some tunnels depending on levels in the stor¬ 
ages). 
The rock wall of Geehi Dam (slope 2:1 and 
vertical height difference between full supply 
level and dam crest of 5.49 m) would not be a 
significant barrier to G. brevipinnis gaining ac¬ 
cess to the river downstream. The species is 
noteworthy for its ability to climb steep water¬ 
falls and moist rocky faces by using its pectoral 
and pelvic fins, and to move great distances in¬ 
land past formidable barriers (McDowall & 
Frankenberg 1981). 
Another route by which G. brevipinnis or other 
species of fish could move into the Murray- 
Darling drainage basin is via the Eucumbene- 
