Does predation by the fishGambusia holbrooki(Atheriniformes: Poeciliidae) contribute to declining frog populations?

9 Pages·1997·1 MB·English

by  Cameron Webb

Does predation by the fish Gambusia holbrooki
(Atheriniformes: Poeciliidae) contribute to
declining frog populations?
Cameron and Jean Jossl
'School rrf Bi18logic~Sl ciences. \12cciuar1r Un!versnly. Nt:w S0u111 \\'ales 2109
'I'roent arldress: >ledacal F.ntomologv, \Vesttnead Hosl,~ral, \Vcs~n~cadN,c * Sourh M1al(:s 214.5
ABSTRACT
Gambusia holbmki is a small, aggressive fish introduced into Australia in 1925 to control
mosquitoes. It has been suggested that this species may also prey on tadpoles and be a contributing
factor in the decline of some Australian frog populations. We examined the influence of tadpole body
size (10 mm. 15 mm and 20 mm). predator prey ratio (1 :I, 1: 2 and 1: 4) and nuvitional status of the
predator on the level of predation by G. holbrooki on tadpoles of Limnodynasfas peronii. Unfed fish
attacked all three size classes of tadpoles without any significant preference but did so more vigorously
when the predatorlprey ratio was densest. Predatorlprey ratio had no significant overall effect on attack
of tadpoles by fed fish but there was a preference for the smaller tadpoles during the first six hours
of the experiment. Unfed fish also consumed more of the tadpolesthey attacked than did the fed fish.
Field surveys were carried out on 10 permanent water bodies in nolth-west Sydney to examine
any correlation between the abundance of G. holbrooki and fmg species richness or abundance. The
most abundant species was C. signifera. Regression analysis showed a negative relationship between
density of G. holbmoki and the abundance of frogs. A positive relationship between the abundance
of tree frog species and the cover of aquatic vegetation was also found. While this study did not
investigate other factors conb.ibuting to declining frog populations, the results of the laboratory experi-
ments and field surveys are consistent with the hypothesis that G. hoibrooki is contributing to declining
frog populations.
INTRODUCTION species of amphibians have developed charac-
teristics that reduce their vulnerability to preda-
A decline in frog populations has been
tion; such as, unpalatability, reduced activity,
reported across all continenw of the world,
and selection of fish-free ponds for breeding
including Australia (Blaustein and Wake 1990;
(Grubh 1972; Smith 1983; Crump 1984; Kats et
Baringa 1990; Martin 1990; Ferraro and Burgin
al. 1988; Petranka 1989; Resetaritset al. 1989).
1993). Of Australia's 203 species of frogs,
approximately 10% are suspected to be now Gumbusin holbrooki (previously G. affinis) is a
extinct (Phillips 1990) and at least 34 species are member of the Poedliidae family. It is commonly
considered endangered (Tyler 1994). On a called the Mosquito Fish because of its reputa-
global scale, a range of factors have been tion as a biological control for pest mosquitoes.
suggested as possible causes for this decline. A native of South America, G. holbrooki was
These include climate change, habitat degrada- introduced to Sydney's waterways in 1925
tion and ozone layer depletion (Jutterbuck following concern over the transmission of
1990). On a local scale there have been examples mosquito-borne diseases (Myers 1965). Gambusia
of increased mortality due to insecticide use holbrooki quickly adapted to the Australian
(Tyler 1991), altered water flow regimes (Hero waterways and presently occupies permanent
1991), reduced water quality (Beattie and Tyler- waterways in all major geographical areas of
Jones 1992; Warner et al. 1993), habitat disturb- Australia (Memck and Schmida 1984). It has
ance (Tyler and Davis 1985; Barinaga 1990) and become so abundant in some areas that it is now
the introduction of exotic flora and fauna considered a pest (Allen 1989). Mosquito Fish
(Bradford 1989; Blaustein and Wake 1990; tolerate a broad range of environmental con-
Bronmark and Edenhamn 1994). In this latter ditions, can exist in high densities and have a
category, it has been suggested that introduced non-specific diet. Their high fecundity, vivi-
fish may influence the survivorship of frogs by parity and low fry mortality result in rapid
preying on their eggs and larvae (Grubb 1990; population growth (Moyle and Cech 1982).
Ferraro and Burgin 1993).
Hurlbert et al. (1972) have suggested that G.
For gape-limited fish predators, prey body holb~ookii ntroduced to non-native sites adversely
size has a significant influence on the level of affects the aquatic environment by altering
predation (Semlitsch and Gibbons 1988). Some invertebrate and phytoplankton communities. It
316 Australian Zmiogist 30(3)
may also reduce native fish populations through The tadpoles of each species were maintained
predation of eggs and fry as well as competitive in tanks under conditions similar to the experi-
interaction with adults (Leger Moss 1988). Lloyd mental tanks and were regularly fed boiled
(1989) has proposed that 35 fish species world- lettuce and commercial fish-food flakes.
wide are in decline due to predation by G.
holbrooki. Predation Experiments
In pristine environments, G. holbrooki may be Limnodynastes peronii tadpoles of 10 mm,
controlled by inter-specific competition with 15 mm and 20 mm SV length were tested with
native species such as the ornate Rainbow a predatorlprey ratio of 1:2 (5 fish and 10
Fish Rhadinocentrw omtus (Leger Moss 1988). tadpoles). Three predatorlprey ratios, 1:l (10
However, Lloyd (1989) reported that Mosquito fish and 10 tadpoles), 1:2 and 1:4 (5 fish and 20
Fish gain ascendancy when streams become tadpoles) were tested with L. peronii tadpoles of
degraded. Native fish are much less tolerant of 15 mm SV length. Crinia signi/era tadpoles of
polluted water, and are unable to outcompete 10 mm SV length were tested at a predatorlprey
G. holbrooki in degraded habitats. ratio of 1:2.
Although G, holbrooki is primarily insectivor- For each experimental trial, ten tanks (600 X
ous (McDowall 1980; Booth 1980), it is very 300 X 260 mm) were used with 15 litres of aged
aggressive and may attack and prey upon water and a 20-30 mm depth substrate of
tadpoles (Ferraro and Burgin 1993; Mahony medium-coarse gravel. The tanks were not pro-
1993). Information regarding predation by G. vided with filtration or aeration and after each
holbrooki upon tadpoles is mainly anecdotal but trial they were cleaned and the water changed.
its predation of amphibian spawn has been
The fish were introduced to the experimental
documented (Grubb 1972). Knowledge of the
tanks 48 hours before each trial. In five of the
factors influencing predation are important pre-
tanks, the fish were fed 4 hours prior to intro-
requisites for the development of appropriate
duction of the tadpoles while the fish in the
conservation strategies (Bronmark and
other five tanks were not fed during the 48
Edenhamn 1994).
hours prior to the trials. Each trial commenced
This study investigated whether G. holbrooki mid-morning. Tadpoles were released into each
could be contributing to the decline in native tank simultaneously and were not provided with
frog populations. Of particular interest was to food during the trial.
identify the factors influencing the level of G.
The number of tadpoles either lethally
holbrooki predation on tadpoles and the influ-
attacked or consumed were recorded after 10
ence of the presence of G, holbrooki on the com-
minutes, 60 minutes, 6 hours, and 24 hours.
position of anuran pond communities. This
Predation was defined as death of the tadpole
study did not investigate other factors contrib-
resulting from lethal attack by G, holbrooki, or as
uting to frog population decline or how such
consumption, when all traces of the tadpole had
other factors may interact with predation by G.
gone from the tank. Both values were recorded
holbrooki.
as a percentage of total tadpoles released into
each tank at the beginning of each trial. No
MATERIALS AND METHODS
tadpole was removed from the tank until 24
Experimental Animals hours had elapsed. After 24 hours all remaining
tadpoles and fish were removed from the tank.
Gumblcsia holbrooki was collected from the
These tadpoles and fish were not used in
Macquarie University Dam. Only those in the
subsequent experiments.
size range 2535 mm SL were used in order to
restrict the influence of predator body size on A two fixed factor analysis of variance was
the predation experiments. used to analyse the results of the predation
experiments. The assumption of homogeneity
The Striped Marsh Frog Limnodynasles peronii
was evaluated using Cochran's test.
is a relatively large frog reaching 65-70 mm in
length and it is common in the coastal regions
Field Surveys
of eastern Australia (Tyler 1992). The tadpoles
of L. peronii used in the predation experiments Sile descriptiom
were raised from a single egg mass collected
Ten ponds were selected in the northwestern
from the lungfish ponds at Macquarie Univer-
region of Sydney and surveyed over the period
sity. The Common Brown Froglet Crinia signfma
September-October, 1994:
is a small frog reaching 35 mm in length and is
widely distributed throughout southeastern Pond A: Macquarie University Dam
Australia (Tyler 1992). Tadpoles of C. signifea (33"46'30"S, 15l006'9OE). A lake created by the
used in predation experiments were collected damming of Mars Creek in 1991. There is little
from the Lungfish ponds with a 400 mm tree cover by the side of the pond but upslope of
diameter dip net. the northern bank is a small area of woodland.
Australian Zoologist 30(3) 317
The bulrushes are regularly cut-back and density (number of G. holbrooki per sweep per
following periods of heavy rain there is a site). Although this technique did not yield an
significant input of storm water. estimate of total abundance at each site it did
give a measure of comparison among the 10
Pond B: Top Lungfish Pond (33'46'10"S,
ponds.
151°06'70"E), and,
Pond C: Bottom Lungfish Pond (33'46'10"S, Estimating frog populations
15lo06'70E). Ponds B and C were approxi-
Frogs were either identified morphologically
mately two-years old at the time of sampling.
using Robinson (1993), with further reference
Each has a capacity of 750 m3. Both were
to Tyler (1994) or by their calls using Grigg and
covered with netting to inhibit entry by birds
Barker (1983). Since frequency, duration and
during the course of this study. The top pond
pattern of calls act as isolating mechanisms
contained nine adult lungfish and an unknown
between sympatric species (Tyler 1994),t hey are
number of juveniles.
a good means of taxonomic identification
Pond D: Top Tondall Dam (33'37'90"S, (Littlejohn 1968).
150°58'30"E). A dam approximately 6 years old
Although gathering richness (number of
with some tree cover.
species) data is relatively simple for frogs, quan-
Pond E: Bottom Tondall Dam (33'37'60"S, tifying their abundance is more difficult (Slater
150°58'30"E). The pond was created by 1978). Southwood (1966) suggested that catch
damming a small creek more than six years per unit effort in a particular habitat provides a
ago. The pond is at the bottom of a gully good estimate of the actual population. A modifi-
surrounded by eucalypt woodland. cation of this method as described by Ferraro
Pond F: API Country Club (33"41'30"S, (1993) was used. The perimeter of each pond
15W55'30"E). A dam situated in a golf course was searched on three occasions, at night for
with no tree cover, and an adjacent, recently one hour, twice following rain and once follow-
constructed dam. ing 10 nights without rain. For each replicate
search all ponds were visited within three days
Pond G: Castlebrook Gardens Dam (33"41'30"S,
of each other to minimize variation due to
15Oo55'30"E). A dam with a small island in the
weather conditions.
middle. There is a small stream emptying into
the dam which was barely running at the time The number of individuals identified during
of study. Some portions of the pond bank have the three visits was averaged to obtain a relative
been covered with concrete. measure of abundance (number of frogs per
hour search). Regression analysis was used to
Pond H: Market Dam (33"43'30"S,
determine the relationships between the frog
150e56'30"E). A dam adjacent to Caddies Creek.
population size and the abundance of G. holbrooki
There is no emergent aquatic vegetation and the
and the per cent cover of aquatic vegetation.
bank is lined with Casurina spp.
Pond I: Top Parklea Dam (33"44'30"S,
RESULTS
150°56'60"E). A dam used for irrigation of the
adjacent sports fields. The western boundary is Predation Experiments
bordered by farms.
The mean per cent mortality of the three body
Pond J: Bottom Parklea Dam (33'44'30"S, size cIasses of L. peronii tadpoles exposed to G.
15Oo56'60"E). A semi-natural water body holbrooki over four consecutive time periods is
situated at the bottom of a small gully. This shown in Figure 1. Fed fish attacked only 10 mm
pond is in a grazing paddock so there were some tadpoles over the first hour and 20 mm tadpoles
cattle present and a few overhanging trees. were attacked only during the G24 hour period
There is a similar-sized dam to the west with less (Fig. la). Unfed fish readily attacked all the
aquatic vegetation. three size classes of tadpoles presented over the
experimental period (Fig. lb). After 10 minutes,
The percentage cover of water body circum-
there was a significant difference in the mean
ference by aquatic vegetation was recorded for
percent mortality of tadpoles between the diet-
each pond and identification of aquatic plants
ary status of the fish (F,,, = 6.75, P = 0.02) and
was carried out using Sainty and Jacobs (1981),
the body size of the tadpoles (F2,24= 6.30, P =
Beadle et al. (1982) and Robinson (1991).
0.01). After 24 hours, however, there was a
Estimating Gamhusia holbrooki densities significant difference due to dietary status (FL,24
= 22.53, P < 0.01) but not to tadpole body size
Each site was inspected twice for G. holbrooki
(FZS4= 2.27, P = 0.125).
using a visual abundance classification based on
the spatial distribution of the fish throughout The mean per cent mortality of L. pmoii
the pond. Their density was estimated by taking tadpoles exposed to G. holhooki at three
10 random sweeps in each pond. The data were predator:prey ratios over four consecutive time
then averaged to obtain a relative measure of periods is shown in Figure 2. Both unfed and
378 Australian Zwlogist 30(3) March 1997
consumed between predator:prey ratio (Fz,z+:
tr10m m 6.74, P : 0.01), more tadpoles were consumed
0.9 a15mm
at predator:prey ratio l: I than either l:2 or l:4.
120 mm
0.8 Unfed fish consumed a significantly greater pro-
portion of tadpoles regardless of tadpole body
G size (F1,2a:2 7.08, P < 0.01) or predator:prey
o ub ratio (F2,2a: 28.26, P < 0.01) than fed fish.
The mean per cent mortality of L. peronii and
o
Z ^r C. signiftra tadpoles (body size 10 mm) exposed
o to G. holbrooAi( predator:prey ratio 1:2) is shown
in Figure 3. After 10 minutes, there was a signifi-
o.2
cant difference in the mean per cent mortality
of tadpoles between the dietary status of fish
(Fr,ro : 10.32, P : 0.01) but not the tadpole
species( Fr,ro: 2.74,P : 0.117).H owever, after
24 hours, there was a significant difference in
mean per cent mortality between the dietary
statuso f the fish (Fr,ro: 7. 78,P : 0.01)b ut not
tadpole species( Fr,ro : 0.98, P : 0.34). How-
1
ever, due to the substantial variance in the data
tr10m m
0.9 @15mm (Cochran's Test : 0.89, P < 0.01), these results
f20mm should be approached with caution.
0.8
=
o u.o
0.9
c
0.8
o
o 0.4
do
o ub
o.2
o
0.1 oq ^A
exposuret ime o.2
Figure 1. The mean per cent mortality (+ st. dev) of Lirn-
nodynastesp erone:it adpoles (body size 10, 15 and 20 mm SVL)
when exposed to (a) fed and (tr) unfed Gambusia holbrooki
(predator:prey : 1:2) over four consecutive time periods.
fed fish attacked tadpoles at the three
predator:prey ratios over all time periods. After
10 minutes there was a significant difference in 0.9
the mean per cent mortality of tadpoles between 0.8
the dietary status of the fish (Fr,z+ : 18.44,
P < 0.01) and predator:prey ratio (F2,24: 9.75, =
o
P < 0.01). The highest mean per cent mortality o ub
of tadpoles (80%) occurred during the first ten 5 o.s
minutes, by unfed fish at a predator:prey ratio g
of l:1. After 24 hours there was a significant 4c- o4
difference in mean per cent mortality between oE UJ
the dietary status of the fish (Fr,z+ : 20.13,
4.2
P < 0.01) but not to predator:prey ratio (Fz,z+:
3.18,P : 0.06). 0.1
After 24 hours, there was a significant differ- 0
10min th 6h 24h
ence in the mean per cent of L. peronii tadpoles
exoosurelim e
consumed by G. holbrooli between tadpole body
size (F2,2a: 4O.32,P < 0.01). A greater number Figure 2. The mean per cent mortality (+ st. dev) of Lim-
of the smallest tadpoles were consumed at nodynastesp eronii tadpoles (body size : l0 mm SVL) when
exposed to (a) fed and (b) onfed Gambusia holbroohi
predator:prey ratio l:2. There was also a signifi-
(predator:prey l:1, l:2 and 1:4) over four consecutive rime
cant difference in the mean percent of tadpoles penocls.
March 1997 AustralianZ oalogist 30(3) 319
(bottom Tondall dam) which had an overhang-
ing tree cover of >75%, which was predomin-
antly Eucalyptur spp.
The mean abundance of particular frog
species at the 10 survey sites is given in Table 2.
There was a total of 10 species identified, six of
which were "ground frogs" (Myobatrachidae)
and four "tree frogs" (Hylidae). These have all
been previously recorded in the Sydney area !
(Robinson 1994). The four most common frogs
were Criniu signifem, which occurred at all sites,
Litoriu peronii (seven sites), Litoriu fallax (five
sites) and Litaria uerreawi (five sites). Sites B and
D had the greatest species richness (six species
each). Site F, where the highest mean frog
10 min 1 h 6 h 24 h , -
abundance was recorded (9.0 k 4.55 frogs1
."llOsYm time hour), had only three species contributing to
this abundance. The lowest number of species
(two) was recorded for sites E and H while the
site w*it h the lowest mean frog abundance
(2.67 0.94) was site A.
Regression analysis revealed no significant
relationship between frog richness and either G.
holbrooki density (P = 0.16) or cover of aquatic
vegetation (P = 0.81). There was no significant
relationship between the mean frog abundance
and cover of aquatic vegetation (P = 0.065).
However, there was a significant negative relation-
ship between mean frog abundance and the sweep
density of G. holbrooki (P = 0.01, 3 = 0.59). No
significant relationship was found between the
richness or abundance of either "tree frog"
10 mi" I h 6 h 24 h or "ground frog" species and the density of
'xp0SY.e time G. holbrooki.
There was no significant relationship between
Figure 3. The mean per cent mortality (+ st. dev) of Lim- the abundance of "ground frogs" and the cover
nodymfcs pnmii and Crinio signiJera tadpoles (body size =
10 mm SVL) when exposed to (a) fed and (b) unfed Gambwia of aquatic vegetation (P = 0.12). However, there
holbrooki (predator:prey = 1:2) over four consecutive rime was a significant positive relationship between
periods. the abundance of "tree frog" species and the
per cent cover of emergent aquatic vegetation
Field Surveys (P = 0.03, r2 = 0.49).
G. holbrooki were recorded for seven of the 10
DISCUSSION
sites (Table 1). The highest mean sweep density
+
was 3.9 2.34 fishlsweep at site A (Macquarie The aim of this study was to determine
University dam). In four sites (B, C, D and J) G. whether predation by Gambusiu holbrooki could
holbrooki was not detected in sweep samples. be contributing to declining frog populations.
However, at site C, fish were reported in the Therefore, the most important finding from the
qualitative population estimates. There was a laboratory experiments was that G. holbrooki will
correlation (P = 0.856) between the visual and attack and kill tadpoles. Gambzlria holbrooki is well
sweep estimates of G. holbrooki density. documented as an insectivorous predator
(Booth 1980; Merrick and Schmida 1984; Lloyd
The coverage of aquatic vegetation ranged
et al. 1986)b ut the results described here suggest
from nothing for site H (Market dam) to 100%
that G. holbrooki is also likely to be a predator of
cover at site G (API golfcourse). The dominant
tadpoles.
emergent aquatic plant at most sites was the
Bulrush Typha orientalir. At sites B and C Since unfed G. holbrooki preyed more heavily
(Macquarie University lungfish ponds) there was on tadpoles than did fed fish, tadpoles in
a good cover of Water Hyacinth Eichornia disturbed habitats, where insects are less avail-
crmsipes. The majority of the sites had very few able, may be more susceptible to predation. G.
overhanging trees since they were located on holbrooki is commonly found in disturbed
farming properties. The exception was site E habitats because it is relatively tolerant to
320 kstraiian Zoologist 30(3)
e e
m 10 replicat 1 1 O.O(O.0) 90 ~~ prriods at th I 3.00 4 8.00(2.16)
fish ber of per sweep fro H I 5 3 3(2.05) 1.2(1.60) 0 10 dev) from three sarriplir H I 2.66 7.37 . 2 4 4.00(0.82) ti.OO(2.94)
e (mean num G 4 W1.24) 2. 100 ndance st. (? (1 2.00 4 4.60(1.25)
5 4 = ommon, abundant and abundant) and quantitativvery = mergent, aquatic vegetation at the 10 ponds. survey Survey Ponds E C D F 2 2 4 I O.O(O.0) O.O(O.0) 0.3(0.64) l.l(l.13) 0. 70 10 10 20 he frog richness (number of species) and the mean total frog abupond individuals were collected during any search. bur no a Survey Ponds E B C D F 2.33 1 .OO 4.33 3.33 2.66 3.66 2.66 3.33 3.33 + + 0.33 1.00 + 0.33 0.66 1.33 I33 Ofifi 6 4 6 2 3 8.67(1.25) 8.00(0.00) 8.30(3.30) 4.67(1.25) g.OO(4.55)
2 I. Table = = = Qualitative (visual abundance; 1 absent, sparse, 3 cof holhooki per of Garnbusia cover sweeps) estimates density and cent e Variables B A C. holbrooki qualitative estimate 1 5 mean sweep density (st. dev) 3.9r2.34) O.O(O.0) Emergent aquatic vegetation 60 30 per cent cover Tablc 2. The mean abundance (per hour search) of each frog specics, t10" survey ponds. indicares species was recorded calling from a A (+) A Swcies Common Name Family: Myobatrachidae Crinia hmelli Haswell's Frogler rignfwa Brown Froglet 1.67 Grim Limnodynartes pmnii Striped Marsh 0.66 Frog ~imnod~tes>armnnicll~is spotted Marsh ~r6g Prnuiophyne bibroni Bibron's Toadlet lornipgala Upcrolca Smooth Toadlet 0.33 Hylidae Family: Tree Litoria dmtata Bleating Frog Liloria/allax Tree Dwarf Frog Litoriapwonii Peron's 'Tree Frog Liloria vcneauri Verreaux'r Froe Species richness 3 Mean total frog abundance (st. dev) 2.67(0.94)
" L - 2- .5
pollution (Lloyd el al. 1986) as are some of the presumably through selective predation of par-
more common species of tadpoles. The fact that ticular tadpole species. There have been no
the size of the tadpoles and the ratio of tadpoles reports of direct predation of adults. During
to fish had little effect on the attacks made on spawning, it is possible that frogs exhibit a
tadpoles by hungry fish, suggests that G. preference for ponds without fish (Bronmark
holbrooki initially responds to hunger rather than and Edenhamn 1994) but no specific studies
aggression. This is supported by the work of have been undertaken to examine this proposi-
Bence and Murdoch (1986), who also found that tion. If a species did exhibit a preference for a
G. holbrooki responded to its own hunger level fish-free spawning habitat, then the presence of
rather than prey density. Fed fish however, also G. holbrooki could influence the abundance of that
attack tadpoles but not necessarily as food. frog. Predation by G. holbrooki upon tadpoles
Thus, even where preferable foods are in could reduce the abundance of fmgs by reducing
abundant supply, tadpoles may still be severely the number reaching metamorphosis. The impact
attacked by G. holbrooki. of G. holbrooki will depend on the abundance of
tadpoles, the length of larval period and the
Handling time is an important factor in-
susceptibility of the species.
fluencing size limited predation (Brodie and
Formanowics 1983; Begon et al. 1990). Gambusia Gambwia holbrooki are indicative of disturbed
holbrooki can easily consume the 10 mm sized waterbodies. Although G. holbrooki could be
tadpoles, usually in one or two strikes, but a reducing the abundance of frogs, there are
longer handling time is required f~."la rger other factors associated with disturbed habitats
tadpoles. Previous research has shown small that may be of greater importance. There is a
tadpoles to be more susceptible to predation by number of factors that could be influencing
vertebrate (Morin 1981; Semlitsch and Gibbons frog communities. Pollution due to suburban
1988; Alford 1989; Petranka 1989; Semlitsch stormwater, disturbance of the aquatic andlor
1993) and invertebrate (Caldwell et al. 1980; terrestrial vegetation or the presence of non-fish
Crump 1984; Travis et al. 1985; Richards and predators would influence frog populations.
Bull 1990) predators. However, our results indi-
cate small tadpoles were at no greater risk than Studies have indicated that pesticides and
larger tadpoles from predation during extended other chemicals can adversely affect frog popu-
exposure to G. holbrooki. lations (Ferraro and Burgin 1993). As reported
in Ferraro and Burgin (1993) studies carried out
Tadpoles are particularly susceptible to pre-
by Johnson in 1976 found that the reaction of
dation immediately following hatching when
tadpoles to different herbicides and pesticides
they hang vertically, immobile beneath the egg
varied depending on the chemical concentra-
mass. During this time the tadpoles are small
tion, frog species and tadpole age. A number of
and are not able to actively avoid predation. In
the survey sites were located in the proximity to
the case of Limnodynastm peronii, there may be
agricultural areas, on which pesticides andlor
more than 800 tadpoles suspended under one
herbicides are used. The leaching or runoff of
egg mass. Although these tadpoles are at risk
these chemicals could alter community structure
from predation because of their small size, the
by reducing the abundance of sensitive species
results of the predatorlprey ratio experiments
andlor indirectly increasing the inter-specific
suggest predation is initially significantly lower
competition.
at higher prey densities. However, both tadpole
body size and abundance will affect the level of The natural variation that is found among
predation since a single fish may consume a ponds makes extrapolation of ecological con-
greater number of small tadpoles. Regardless of clusions difficult. Future studies should address
the density of G. holbrooki in a water body, a greater range of factors that may he influenc-
tadpoles are likely to be exposed to predation at ing frog populations such as, pond area, pond
some time during their larval period. vegetation, pond water pH, pesticidelherbicide
concentration, distance from nearest permanent
Some tentative conclusions can be drawn from
water body and number of ephemeral water
the field studies which suggest that the presence
bodies. A study of this kind could test the null
of G. holbrooki may directly affect the frog com-
hypothesis that the density of G. holhrooki does
munities of permanent ponds. The presence of
not have a significantly greater impact on frog
G. holbrooki did correlate with a lower abundance
populations than other biotic or abiotic factors.
of frogs in general but appeared to have little,
if any, affect on the composition of the frog
For species that are particularly susceptible to
community.
G. holbrooki, the occurrence of ephemeral water-
The richness of a pond's frog community will bodies surrounding permanent water bodies
depend on the ability of the species to locate the would be valuable refuge sites. G a m k h olbrooki
waterbody and the presence of species in the and frogs could co-exist during periods of high
local area. If G. holbrooki are altering the species rainfall due to the presence of ephemeral water-
composition of permanent ponds it would be bodies. These water bodies provide a refuge
322 Ausfralian Zoologist 30(3)
from fish predation. For species shown to be Crump, M. L., 1984. Ontogenetic changes in vulnerability
preyed upon by G. holbrooki, provision of to predation in tadpoles of Hyla pseudopuma. Hn;
petologca 40(3): 265-71.
ephemeral sites will ensure maintenance of their
population. Such pro-active actions are needed Ferraro, T. J., 1993. Amphibian decline: a case study in
to forestall further declines in native frog western Sydney. Pp. 197-204 in Herpetology in Awlralia:
populations. A Diverse Discipline ed by D. Lunney and D. Ayers. Royal
Zoological Society of New South \+'ales: Mosman.
Ferraro, T. J. and Burgin. S., 1993. Review of environ-
ACKNOWLEDGEMENTS
mental facton influencing the decline of Australian
frogs. Pp. 205-18 in Herpetology in Australia: A Diuerse
We thank all those who allowed access to
Discipline ed by D. Lunney and D. Ayers. Royal Zoo-
their properties, especially Marcus and Pauline
logical Society of New South Wales: hlorman.
Tondall whose enthusiasm and assistance was
much appreciated. Arthur White for his Grigg, G. and Barker, J., 1983. Frog Calk ufS.E. Au(t~a/io,
(Audio). Australian Museum: Sydney.
valuable advice regarding experimental
methodology. Ian Oliver and Janice Lord pro- Gruhb, J. C., 1972. Differential predation by Gomhuiaaflnis
vided expert statistical advice. David Webb and on the eggs of seven species of anuran amphibians.
Amer. Midland A'atur 88: 102-08.
the 1994 School of Biological Sciences Honours
students for assisting with the field work. Animal Hero, J., 1991. A froggy forecast. Wlldl. Aust. 28: 14-15.
Ethics and Experimentation Authorization for
Hurlbert, S. H., Zedler, J..and Fairhank, D., 1972. Eco-
Macquarie University: 94038 :Effects of preda-
system alteration by mosquitofish (Gambwia afinis)
tion on frog distribution in Sydney. predation. Science 175: 63941.
Jutterbuck, J. E., 1990. Comments about the amphibian
REFERENCES decline symposium. He~pctolR. ev. 21: 75-76.
Alford, R. A,, 1989. Variation in predator phenology affects Kats, L. B., Petranb, J. W. and Sih, A,, 1988. Antipredator
predator performance and prey community compo- defenses and the persistence of amphibian larvae with
sition. Ecology 70: 20619. fishes. Ecology 69: 1865-870.
Allen, G. R., 1989. Fmhwoter Fishes uf Australia. T. F. H. Leger Moss, J. T., 1988. A preliminary report for Queens-
Publications: Brookvale. land National Parks and Wildlife Service on the fresh-
-
water fishes of Morton Island. Oueensland with soedal
Barinaga, ht., 1990. Where have all the froggies gone?
comment on the exotic "mosquno fish, Gombwvl afinu.
Scimir 247: 1034-034.
Dept. Recreation and Health: Brishane City Council.
Beadle. N. C., 1982. The uegetation o f Aurtmlia. Cambridge
Littlejohn, M. J., 1968. Frogcalls and the species problem.
University Press: Sydney.
Aurt. Zool. 14: 25944.
Beattie, R. C. and Tyler-Jones, R., 1992. The effect of low
pH and aluminium on breeding success in the frog Lloyd, L. N., Arthington, A. H. and Milron. D. A,, 1986.
Rena tempom&. J. Herpetof. 26: 35340. The mosquitofish - a valuable mosquito-control agent
or a pest? Pp. 7-25 in The Ecology of Erotic AnimaLr and
Begun, M., Harper, J. L. and Townsend, C. R., 1990. Plants: somr Australian case histo& ed by R. L. Kitching.
Ecoloa: Indiuiduak, populationr and communities. John Wiley and Sons: Gladesville.
Blackwell Scientific Publications: Boston.
Lloyd, L., 1989. Ecological inreractions of Gambusia holbrooki
Bence, J. R. and Murdoch, W. W., 1986. Prey sire selection with Australian native fishes. Pp. 94-97 in Introduced
by the mosquitofish: relation to optimal diet theory. and Trum.locatedfrches and their ecological effects ed by D.
Ecok 67: 24-36. A. Pollard. Australian Government Publishing Service:
Blaustein, A. R. and Wake, D. B., 1990. Declining Amphi- Canberra.
bian Populations: A Glohal Phenomenon? TREE 5:
hlahony, M. J., 1993. The status of frogs in Watagan Moun-
20344.
tains area of the Central Coast of New South Wales. Pp.
Booth, D. J., 1980. Investigations into the extent and 25744 in Herpetology in Autralia: A Diverrr discipline ed
mechanisms of prey-size selection by the fish species by D. Lunney and D. Ayers. Royal Zoological Society of
P~domugilsi gn* (KNER) and Cambuia enis (Baird New South Wales: Mosman.
and Girard), unpublished B. Sci (Hons) thesis. Univer-
hlanin, G., 1990. Froggy Bottom. hcui,er 1990: 3tL37.
sity of Sydney.
Bradford, D. F., 1989. Allotropic distribution of native frogs McDowall, R. M., 1980. F~e~hwoteFri che1 of South-eartern
and introduced fishes in hi"zh Sierra Nevada Lakes of Aut~alia.R eed Publishing: Terrey Hills.
California: Implication of the negative effecw of fish
Merrick,J. R. andSchmida, G. E., 1984AwtralianFrahwater
introductions. Copein 1989(3): 775-78.
Fishes: Biolag). and Management. Griffin Press Ltd:
Brodie, E. D., Formanowicz, D. R. and Jr., 1983. Prey size Netley.
preference of predators: Differential vulnerability of
larval anurans. Herpetologica 39: 67-75. Morin, P. J., 1981. Predatory salamanders reverse the out-
come of competition among three species of Anuran
Bronmark, C. and Edenhamn, P., 1994. Does the presence tadpoles. Scimce 41% 1284-286.
of fish affect the distribution of tree frogs (Hylaarborea)?
Comew. Bial. 8: 84145. Moyle, P. B. and Cech, J. J., 1982. Fisher: An introducttan to
IchIhyology. Prentice-Hall Inc: New Jersey.
Caldwell. I. P.. Thoro. I. H. and lervev. T. 0.. 1980.
Myers, G. S., 1965. GambusLL, the fish destroyer. AM. Zool.
13: 102.
March 1997 Ausbalian Zoologist 30(3) 323
Petranka, J. W.. 1989. Response of toad tadpoles to con- Slater, K., 1978. Fauna:reptiler and amphibians. Pp. 80-86
flicting chemical stimuli: Predator avoidance versus in Land we on the South Coat of NmS mth Waler, Vol2.
"optimay foraging. Herpetolog'ca 45: 283-92. ed by R. H. Gunn. CSIRO: Melbourne.
Phillips, K., 1990. Where have all the frogs and toads gone? Smith, D. C., 1983. Factors controlling tadpole populations
BioScunCe 40: 422-24. of the Chorus Frog (PI& trireriata) on Isle Royale,
Michigan. Ecology 64: 501-10.
Resetarits, W. J., Wilbur, Jr. and H. M., 1989. Choice of
Oviposition Site by Hyla Chrysoscelis: Role of Predators Southwood, T. R. E., 1966. Ecological Methods with parliculnr
and Competitors. Ecology 70: 22&28. re/erence lo the stud? ojimecl populnttm. Chapman Hall:
London.
Richards, S. J. and Bull, C. M., 1990. Size-Limited Predation
on Tadpoles of Three Australian Frogs. Cap- 1990: Travis, J., Keen, H. and Juilianna, J., 1985. The role of
1041-046. relative body sire in a predator-prey relationship
between dragonfly naiads and larval anurans. Oikm 45:
Robinson, L., 1991. Field Guide to the Native Planti oJSydwy. 59-65.
Kangaroo Press: Kenthurst.
Tyler, M. J., 1991. Our vanishing frogs. Hahilot 19: 2&25.
Robinson, M., 1993. A FkM Guide to Frogs oJAvstraiia. Reed
Publications: Chatswood. Tyler, M. J., 1992. Enqclopedia of Auhalian Anid: Frogs.
Angus and Robinson: Pymble.
Sainty, G. R. and Jacobs, S. W. L., 1981. Watnplanu of New
Swth Wales. C. C. Merritt Pty Ltd: Chiswick. Tyler, M. J. and Davis, hi., 1985. The gastric brooding frog
Rheobohahw elm. Pp. 469-70 in The Biology ojAmtra-
Semlitsch, R. D., 1993. Effects of different predators on the lian Frogs and Reptiles ed by G. Grigg, R. Shine and H.
survival and development of tadpoles from the Ehmann. Surrey Beatty & Sons: Chipping Norton.
hybridogeneticR am esnrlenta complex. Oikos 67: 4W6.
Warner, S. C., Travis, J. and Dunson, W. A., 1993. Effect of
Semlitsch, R. D,a nd Gibbons, J. W., 1988. Fish predation in pH an interrpecificc ompetition between two species of
size-structured populations of treefrog tadpoles. Hylid radpoles. Ecology 74: 183-94.
Oecologica 75: 321-26.
BOOK REVIEWS - BOOK REVIEWS
-
A Shadow and a Song the shuggke to save an conservation is no guarantee of successful outcomes.
endangered species by M. J. Walten (1992). If there is a villain of the story it is the US Fish and
Chelsea Green Publishing Company. Wildlife Service but Walters is at pains to emphasize
ISBN 0-930031-58X. that the agency did not act with malicious intent,
rather the Dusky Seaside Sparrow fell victim to the
There are few taxa whose time and cause of demise prevailing corporate wisdom which arose out of the
can be accurately documented. It is of the nature of a-zenc.y's h istory.
rare species that we know little about them when they
are alive, and we may be uncertain for many years as Some would argue that sad as the tale is, the loss of
theDusk~S easideS~arrowwasofnoFeatconsequence
to whether or not they are extinct. The death, in
captivity, of the last of several species, has been as it was not a distinct species, and even its subspecific
recorded, examples include :he passenger pigeon and status within a difficult taxonomic complex could be
the thylacine, T~ [his list can be added the ~~~k~ doubted. This would be to miss the point; there is no
sparrow evidence that even if it were a dearly distinct full
seaside ~~~~d~~~~ -ritimu %jgrescmt,h e
last known individual dying on June 16, 1987. species appropriate conservation measures would have
been applied. In addition the current biodiversity
The causes of extinction are many and various, paradigm rightly stresses the importance of conserva-
although ultimately most can be ascribed to the cupidity tion of genetic diversity; genotypes, regardless of
and stupidity of Homo snpiens. The Dusky Seaside whether they are afforded formal taxonomic status,
Sparrow was a victim of the American way of life, are proper subjects for conservation management.
being destroyed by the space race and real estate However, the taxonomic distinctiveness of the Dusky
development with the last bird dying in Disney World. Seaside Sparrow did, at the eleventh hour, become of
great significance. With the population reduced to a
Despite, or perhaps because of, being recognized handful of individuals, all male, the suggestion was
as endangered, and thus subject to special protection made to crossbreed with females of a related subspecies
by conservation agencies, extinction was not averted. and then backcross the offspring so as to conserve
The story of the decline of the Dusky Seaside Sparrow "almost pure" Dusky Seaside Sparrows (a similar
is an object lesson in how bureaucratic confusion and strategy to that currently being implemented with the
vacillation can undermine the most noble of objectives. Norfolk Island Boobook Owl). Unfortunately the Fish
With increasing concern for the survival of threatened and Wildlife Service could not decide whether this
species, and a burgeoning industry in the production was permissible under the Endangered Species Act,
of recovery plans, Walters' book should be required and after a period of indecision and lost opportunity
reading by all conservationists. The existence of the agency washed its hands of the problem and
legislation (in this case the US Endangered Species transferred the surviving captive males to Disney
Act) and apparently appropriate mechanisms for World.
324 hstralian Zoologist 30(3) February 1997