michael d. mills, russell b. rader, and mark c. belk department of … · 2009. 12. 4. · 1...
TRANSCRIPT
1
Complex Interactions between Native and Invasive Fish:
The Simultaneous Effects of Multiple Negative Interactions
Michael D. Mills, Russell B. Rader, and Mark C. Belk
Department of Integrative Biology, Brigham Young University
Department of Integrative Biology, 401 WIDB, Brigham Young University, Provo, UT 84602
E-mail: [email protected] Fax: 801-422-0090
2
Abstract
We suggest that the ultimate outcome of interactions between native species and invasive species
(extinction or coexistence) depends on the number of simultaneous negative interactions
(competition and predation), which depends on relative body sizes of the species. Multiple
simultaneous interactions may constrain the ability of native species to trade fitness components
(i.e. reduced growth for reduced risk of predation) causing a spiral to extinction. We found
evidence for five types of interactions between the adults and juveniles of introduced western
mosquitofish (Gambusia affinis) and the juveniles of native least chub (Iotichthys phlegethontis).
We added 10 large (23-28 mm) and 7 small (9-13 mm) young-of-the-year (YOY) least chub to
replicate enclosures with zero, low, and high densities of mosquitofish in a desert spring
ecosystem. Treatments with mosquitofish reduced the average survival of least chub by one-
third. No small YOY least chub survived in enclosures with high mosquitofish densities. We
also performed 2 laboratory experiments to determine mortality to predation, aggressiveness, and
habitat selection of least chub in the presence of mosquitofish. Mean mortality of least chub due
to predation by large mosquitofish was 69.7 % over a 3-hour trial. Least chub were less
aggressive, selected protected habitats (Chara sp.), and were more stationary in the presence of
mosquitofish where the dominance hierarchy was: large mosquitofish >> large least chub ≈ small
mosquitofish >> small least chub. Least chub juveniles appear to be figuratively caught in a
vice. Rapid growth to a size refuge could reduce the risk of predation, but the simultaneous
effects of competition decreased least chub growth and prolonged the period when juveniles
were vulnerable to mosquitofish predation.
3
Key words: Biological invasions, competition, predation, body size, scale, refugia, Gambusia
affinis, Iotichthys phlegethontis
4
Introduction
Invasive species often cause rapid population declines and the extinction of native
species. Our ability to minimize the effects of invasive species depends on understanding the
mechanisms underlying their interactions with native populations. Many invasive species are
successful because they are released from natural controls (e.g. competitors, predators, parasites)
that regulate population growth within their native range (Brown 1989; Simberloff 1989).
Invasive species also tend to be superior competitors with a high reproductive potential and
predation strategies that were absent during the evolution of isolated native populations (e.g.
Savidge 1987; Ogutu-Ohwayo 1993; Townsend 1996; Sakai 2001). Competition and predation
between introduced and native species, and the introduction of novel parasites can have obvious
direct negative effects (Ashton and Mitchell 1989; Brown 1989). However, efforts to preserve
declining native species can also be frustrated by the complexity of combined direct and indirect
biotic interactions.
Recent research has revealed complex interactions between invasive and native species
because of size-structured competition and predation (Kotler and Holt 1989; Wissinger 1992;
Belk et al. 2001). For example, large invasive predators that feed on large prey can consume the
adults of native populations while providing young native fish with refuge zones from
intermediate-sized predators, thereby increasing recruitment and survival of the entire population
(e.g. Mittelbach and Chesson 1987; Winkelman and Aho 1993). The overall effects of invasive
species on native populations can be difficult to predict.
Individuals can trade-off multiple components of fitness and continue to persist if
confronted by a single threat (Soderback 1995; Kiesecker et al. 2001; Bertolino and Genovesi
5
2003). Prey may respond to competition, predation, or parasitism by altering their morphology,
behavior, or life history to minimize negative effects (reviewed by Tollrian and Harvell 1999).
For example, small fish of many species typically inhabit vegetated or otherwise structured
habitats with poor food resources as a refuge from predation (e.g. Mittelbach 1988). Although
anti-predator responses often result in increased costs to other fitness components (e.g., growth
rate or fecundity), the prey population can persist because decreased growth is balanced by
decreased mortality. Similarly, increased risk of mortality can be traded for increased growth
when confronted with competition from a new species. For example, competing individuals may
forage in habitats that increase vulnerability to predators, but are energetically more profitable
than safer habitats (Wellborn 2002). Overall, native populations can often successfully respond
to single threats from invasive species and continue to persist (Tollrian and Harvell 1999;
Gurevitch et al. 2000).
Given the complexity of size-structured interactions, invasive species may affect native
species via multiple interactions. Native species may be unable to respond to multiple threats,
especially when the intensity of interactions is size-dependent (e.g., Wissinger 1992).
Competition and predation can exert simultaneous pressure on both large and small size classes
reducing the effectiveness of trade-offs. Trade-offs that ameliorate the effects of predation on
smaller size classes (e.g. habitat shifts) may intensify the harmful effects of competition.
Similarly, trade-offs that ameliorate the effects of competition between small size classes (e.g.
shifts in foraging time or place) may exacerbate the harmful effects of predation (reviewed by
Gurevitch et al., 2000). For example, when small fish shift to refuge habitats to avoid predation,
they may be confronted by intense competition from other species, including refuging size
classes of invasive species (Mittelbach 1986; Mittelbach and Chesson 1987). We propose that
6
multiple interactions can constrain the ability of native species to trade fitness components
causing a spiral to extinction.
We suggest that the coexistence of native species with invasive species depends on the
number and strength of simultaneous negative interactions. We emphasize the need to
understand the complex effects of multiple interactions when attempting to minimize the impact
of invasive species. Identifying the relative strength of multiple negative interactions may help
resource managers predict which species are most at risk, and may indicate effective
management actions for conservation of native species.
We conducted one field experiment and two laboratory experiments to examine multiple
interactions between small young-of-the-year (YOY) least chub (Iotichthys phlegethontis), native
to the Bonneville Basin of Utah, and introduced western mosquitofish (Gambusia affinis). We
tested three hypotheses: 1) large mosquitofish can prey on small least chub, 2) competition
(interference and/or exploitative) between mosquitofish and least chub will reduce least chub
growth, and 3) mosquitofish will force small least chub into refuge habitats (macrophytes), thus
increasing their exposure to invertebrate predators.
Conceptual Background
To evaluate the full range of interactions between two species it is necessary to consider
the size range of each species. Table 1 shows potential interactions for both juvenile and adult
life stages between native and invasive species with similar adult body sizes. This table was
made assuming that invasive species are more aggressive predators and competitors than native
7
species, and that they can reach a high population density because of a high reproductive
potential (Sakai 2001).
We have identified 8 interactions between similar sized invasive and native species that
could have detrimental effects on native populations (Table 1). We decided to focus our research
on the YOY of native species because 5 of these interactions are specific to that size class, thus
suggesting the greatest potential for examining the effects of simultaneous multiple negative
interactions. Although the occurrence of each interaction is contingent upon various traits of
invasive and native species (i.e. competition may depend on whether native fish occupy the same
refugia as invasive fish), this high number of potential negative interactions suggests that enough
interactions may often be realized that trade-offs among fitness components are not possible,
thus resulting in extinction.
Historical Background
Fifty years ago, least chub were widely distributed in the Bonneville Basin in a variety of
habitats including rivers (Provo River), lakes (Utah Lake), small streams, ponds, and marshes
scattered throughout central and southern Utah (Sigler and Miller 1987). Currently the
distribution of least chub is limited to ten isolated spring-fed pools in the deserts of central Utah
(Perkins et al. 1998). Five of these ten springs are infested with abundant mosquitofish
populations. Causes of declining least chub populations are not well documented, but habitat
loss and interactions with mosquitofish are the primary suspects (Holden et al. 1974; Hickman
1989; Crist and Holden 1990; Perkins et al. 1998).
8
Mosquitofish are members of the Poeciliidae family and give birth to live young.
Although they are native to the southeastern United States, their range has been expanded
throughout North America for mosquito control. Increases in mosquitofish have been correlated
with a decline in native species, but the mechanisms are rarely known (e.g. Galat and Robertson
1992; Rincon et al. 2002). However, mosquitofish are aggressive predators that feed on a variety
of aquatic organisms including their own young and the young of other fish (Meffe 1985; Sigler
and Miller 1996; Barrier and Hicks 1994; Gamradt and Kats 1996; Goodsell and Kats 1999;
Komak and Crossland 2000). Mosquitofish are also known to be competitively superior to some
native fish (Lydeard and Belk 1993; Crivelli 1995).
Our field research was conducted at Walter Spring (113.402° West, 31.871° North; Juab
County, Utah, USA), an impounded, spring-fed pond managed by the U.S. Fish and Wildlife
Service at the Fish Springs National Wildlife Refuge. The Fish Springs National Wildlife
Refuge is a large spring complex, consisting of 15 to 20 springheads and associated marshes
connected by surface and groundwater flows spread over approximately 40 km2. Walter Spring
is an averaged-sized spring for the refuge with a surface area of 320 m2 and maximum depth of
3.0 m. Variations in flow and depth are small on both a seasonal and annual basis. Although
temperature at the inflow is constant (18 C), daily and seasonal variations increase away from the
wellhead.
In 1996, mosquitofish were removed (via draining and Rotenone) and least chub were
introduced to Walter Spring as part of restoration efforts. Least chub rapidly increased during
1997 through 2000 reaching thousands of individuals. Until recently, state surveys (minnow
traps) found numerous juveniles indicating healthy recruitment of least chub in Walter Spring
(Wilson unpublished reports). However, mosquitofish managed to re-invade this spring in 1998,
9
and two surveys (July 2002 and May 2003) using 100 minnow traps at various depths for 24
hours found a total of 7 least chub smaller than 30 mm.
Methods
Competition and Predation Experiment
We placed 18 enclosures (1 m2 by 1.5 m in height) consisting of a polyvinyl chloride pipe frame
(PVC) and dark nylon netting (2 mm mesh) in Walter Spring at depths ranging from 29 to 66 cm
in areas where both species of fish were known to occur. Each enclosure was embedded at least
20 cm into the soft sediment and extended at least 20 cm out of the water. We used poles to
gently probe through the vegetation to remove fish before the area was enclosed. Any remaining
fish were regularly removed with a hand-held net for one week before the start of our
experiment. Every effort was made to minimize disturbance to the enclosed area, including the
dominant plants, Potamogeton spp., Chara sp., and metaphyton, which is suspended tufts of
filamentous green algae (Goldsborough and Robinson 1996).
We used a target-neighbor design to determine potential harmful effects of mosquitofish
on least chub (Goldberg and Werner 1983). Each enclosure was randomly assigned one of three
initial mosquitofish densities: no mosquitofish (control), 6 mosquitofish, or 100 mosquitofish.
There were six replicates of both treatments and the control. Treatment densities were based on
estimates of high and low mosquitofish densities at Walter Spring. Although males and females
were included in each treatment, large female mosquitofish were most common (> 30 cm). Two
distinct cohorts of YOY least chub were identified: larger, early-spawned (23-28 mm SL) and
smaller, late-spawned (9-13 mm SL). Based on these size differences, we added either 7 or 10
10
YOY least chub to each enclosure. Four replicates out of the six for each treatment received 10
small YOY least chub, whereas two replicates of each treatment received 7 large YOY least
chub.
We measured the initial length of each least chub at the start of the experiment using a
digital camera raised to 30 cm over the water in a white bucket with a plastic ruler fixed to the
bottom. We used the digital image of each fish to measure standard length (SL) to the nearest
millimeter using image analysis software (SigmaScan Pro 5.0, SPSS Inc.). Water depth was
always 2.5 cm, and only images where the fish was resting directly on the tape were used to
measure length. We used a length-mass regression (Wet Mass = 0.0029 e 0.1763 [Standard Length]; R2
= 0.95) to estimate the initial mass of least chub (to the nearest 0.001 g) in each enclosure based
on fifty individuals varying in size from 7 to 30 mm.
After eight weeks (between July and September, 2001) we removed and counted all fish
from each enclosure. We measured the final length and wet mass of all least chub to estimate
their growth in body length and mass. We calculated mean percent survivorship by dividing the
final number of least chub retrieved by the initial number placed in each enclosure. We used a
removal-sampling scheme to determine the number of surviving individuals. The same person
would sweep the entire interior of an enclosure with a hand-held net (45 cm X 40 cm; mesh size
= 2 mm) for five 1-minute periods, with a one-minute break between periods. This procedure
resulted in all the vegetation being removed and searched within the first 2 sampling periods.
Typically, no additional fish were collected after 3 sampling periods, and no fish were ever
collected on the fifth sampling period.
Mosquitofish reproduce multiple times during the summer (Courtenay and Meffe 1989),
so we expected YOY mosquitofish abundance in enclosures to increase during the course of the
11
experiment. Consequently, the number of large predaceous mosquitofish assigned to treatments
did not change, but the number of small mosquitofish that cannot prey on least chub increased,
especially in the low-density treatment (Table 2). Therefore, differences in levels of predation
were maintained between treatments throughout the study, whereas levels of competition
converged between the treatments containing mosquitiofish as the study progressed. This
difference between predation and competition treatments necessitated a difference in the
statistical analysis of survival and growth (see below).
Predation and Habitat Use Experiment
We used a 35-liter aquarium to determine the direct and indirect effects of mosquitofish
predation on least chub. We examined the potential effects of refuge habitat and invertebrate
predation by adding the most abundant types of vegetation (metaphyton, Chara sp., and
Potamogeton spp.) and invertebrate predators (Aeshna sp.: Aeshnidae, Notonecta unifasciata:
Notonectidae, and Belostoma flumineum: Belostomatidae) found at Walter Spring. The
experiment was a fully crossed factorial design manipulating two treatment factors: mosquitofish
presence (absent, or 3 large (> 30 mm) female mosquitofish present), and invertebrate predator
presence (absent, or 3 dragonfly nymphs, 3 adult backswimmers, and 3 giant waterbugs present).
Four small YOY least chub (11 to 14 mm SL) were present in each of the four treatments. Equal
amounts of each vegetation type were used in each trial, and each treatment was replicated four
times. We conducted the treatments in a randomly selected order, and conducted two trials per
day. All fish were used in only one trial and were haphazardly netted from 190-liter aquaria
containing separate populations of least chub and mosquitofish. Invertebrates were haphazardly
12
selected from a 35-liter aquarium containing 8 to 10 individuals of each species. A few
invertebrates could have been used in more than one trial because they were returned to the same
aquarium between trials. The response variables were percent survival of least chub and the
percent of time least chub spent in each habitat type. The percent time spent in habitat types was
determined using a standard scan sample procedure (e.g. Altmann 1974).
Mosquitofish and invertebrate predators were starved for 48 hours and allowed to
acclimate 1-hour before each trial. Least chub were placed in the aquarium at the same time as
mosquitofish and/or invertebrate predators, but they were enclosed in a glass container during the
acclimation period. Following acclimation, least chub were released and an observer recorded
all aggressive interactions and habitat use for each least chub every 15 minutes for 3 hours. All
trials were run between 0800 and 1600 hrs. In order to minimize bias associated with human
intrusion, we made observations through a port (30 cm X 30 cm) behind a blind, and covered the
aquarium on three sides with opaque paper.
Interference Competition Experiment
We compared the activity and habitat use of two size classes of least chub with and without two
size classes of mosquitofish to determine their behavioral interactions. The behavior of each
group of four least chub (two 20 mm– 30 mm SL and two 30 mm – 40 mm SL) was recorded
with no other fish present and with four mosquitofish present (two 20 mm - 35 mm SL and two
40 mm – 45 mm SL). We recorded the mean percentage of time least chub spent actively
swimming versus stationary, and the mean percentage of time both species spent in the open
water versus cover every two minutes throughout each 20-minute trial. Additionally, we
13
recorded every instance of aggressive interactions by noting the species and size of the initiator
and the recipient, and which individual dominated. We identified the initiator of an interaction
as the fish that approached rapidly and directly at another fish and the victor as the fish that
retained the occupied space following the confrontation (see McDonald et al. 1968). We did not
differentiate between size classes of least chub in terms of activity or habitat use. However, we
did note the size of each individual during aggressive interactions. We completed four replicates
of each treatment (least chub alone, and least chub with mosquitofish). This is a test of
interference competition because the smallest least chub were too large to be consumed by
mosquitofish.
We used a 35-liter aquarium containing two habitat types: open water and cover as
simulated vegetation consisting of green rope (5 mm diameter) attached to the bottom. We
covered three sides of the aquarium with opaque paper, and made observations from behind a
blind. Both species were allowed a 20-minute acclimation period before being allowed to feed
and interact for 20 minutes. All fish were used in only a single trial consisting of both treatments
in a paired design. That is, the same least chub used in the control with no mosquitofish were
also used in the treatment with mosquitofish present, which immediately followed the control.
Statistical Analyses
We used 2-way completely crossed analysis of variance (ANOVA) to determine the
effects of mosquitofish density (none, low, and high) and least chub size (23-28 mm SL and 9-13
mm SL) on the mean percentage of least chub that survived in each treatment of the predation
and competition field experiment. This analysis included a mosquitofish density by least chub
size interaction. Because the design was unbalanced we used Type III sums of squares for
14
interpretation (SAS 1997). We used the same design to analyze the effects of mosquitofish
density and least chub size on the growth of least chub, however none of the small least chub
survived in the high mosquitofish density treatment resulting in a missing cell. To solve this
problem and the problem of convergence between competition treatments described above, we
collapsed the low and high mosquitofish density levels and created a 2X2 design with two levels
of least chub size crossed with the presence and absence of mosquitofish. This design also
included the interaction between mosquitofish presence/absence and least chub size. Growth
was measured as the change in least chub length and mass. We used multivariate analysis of
variance (MANOVA, SAS 1997) with both change in mass and change in length as dependent
variables, and then used separate univariate ANOVA’s to analyze the effects of YOY least chub
size and the presence of mosquitofish on each dependent variable.
For the predation part of the laboratory experiment, we used a 2-way ANOVA to
determine the effects of predators on the short-term survival of small least chub in a 2X2
factorial design crossing the presence/absence of mosquitofish and invertebrate predators. This
analysis included the potential effects of the interaction between mosquitofish and invertebrate
predators on least chub. The response variable was the arcsine-square root transform of the
proportion of least chub surviving after each trial.
Determining the effects of predators on small least chub habitat choice was more
complicated because least chub could occupy only one of four habitat types (open water, Chara
sp. Potamogeton sp., and metaphyton) during each scan sample. This resulted in four response
variables for each trial as the percent time spent in each habitat. However, these are not
independent measurements because these fish could only be located in one of the four habitats
during each scan. To resolve this problem we followed the procedure outlined by Manly (1993),
15
and subtracted one response variable from the next for each trial to create differences between
successive response variables. The mean of these three differences was then used as response
variables in a multivariate analysis of variance (MANOVA) to determine the effect of predators
on least chub habitat use. Presence/absence of predators were main effects, and the interaction
between mosquitofish and invertebrate predators was included.
We used a paired t-test to determine the effect of mosquitofish on the mean percent time
least chub spent actively swimming versus stationary, and the mean percent time they spent in
open versus structured habitats in the laboratory interference competition experiment. Because
there were only two mutually exclusive choices for each response variable (active or inactive and
in or out of cover) these measurements were not independent. Thus, we used percent time of one
of the choices for each response variable, time spent stationary and time spent in the open
habitat, as dependent variables in the analysis (e.g. Manly 1993).
Results
Field Competition and Predation Experiment
The presence of mosquitofish reduced both the survivorship and growth of YOY least
chub and the magnitude of effects differed between the small and large cohorts. On average,
survival of both cohorts of least chub was reduced by about one-third (F2,12 = 4.44, P = 0.036)
when mosquitofish were present (Fig. 1). Survival of large YOY least chub was greater than the
survival of small least chub (F1,12 = 14.56, P = 0.003). None of the small YOY least chub
16
survived in any of the enclosures with high mosquitofish densities. The interaction between
mosquitofish presence and least chub size was not significant (F2,12 = 0.89, P = 0.436).
Young-of-year least chub growth (length and mass combined) was lower in the presence
of mosquitofish, (Wilk’s λ = 0.46, F2,8 = 4.69, P = 0.045; Fig. 2), growth differed depending on
the size of least chub (Wilk’s λ = 0.041, F2,8 = 94.14, P < 0.0001; Fig 2), and the least chub by
mosquitofish interaction was significant (Wilk’s λ = 0.44, F2,8 = 5.19, P < 0.036). Growth in
mass of large least chub was 6-fold less with mosquitofish present than in the control (F1,9 =
7.27, P = 0.024), whereas growth in small least chub was only 1.3-fold less than the control
(chub size x mosquitofish interaction: F1,9 = 5.36. P = 0.046, Fig. 2a). Growth in length was
greater for small compared to large least chub (F1,9 = 40.86, P = 0.0001; Fig. 2b), but neither the
presence of mosquitofish (F1,9 = 2.13, P = 0.178) nor the interaction between the mosquitofish
and least chub size (F1,9 = 0.61, P = 0.455) was significant. Even though growth in length of
large least chub was 7-fold lower in the presence versus the absence of mosquitofish (Fig. 2b),
high variance among replicates contributed to a lack of statistical significance.
Predation and Habitat Use Experiment
Mosquitofish readily consumed YOY least chub while invertebrates made only four
predatory attempts. Survival of least chub was lower in the presence of mosquitofish (F1,12 =
106.8, P = 0.0001), but neither the presence of invertebrate predators (F1,12 = 0.11, P = 1.00) nor
the interaction was significant (F1,12 = 0.11, P = 1.00). The combined mean survival of small,
YOY least chub in treatments with mosquitofish was 31.3 % (± S.E. = 6.25). Conversely, 100 %
of the small least chub survived in treatments without mosquitofish.
17
Mosquitofish presence affected habitat use of small least chub (Wilk’s λ1,0.5,4 = 0.225,
F3,10 = 11.47, P = 0.0014). Least chub spent the majority of their time in open habitat with
mosquitofish absent, but in Potamogeton spp. with mosquitofish present (Fig. 3). Habitat use of
small least chub was unaffected by the presence of invertebrate predators (Wilk’s λ1,0.5,4 = 0.82,
F3,10 = 0.74, P = 0.55) or, the interaction between invertebrate predators and mosquitofish
(Wilk’s λ1,0.5,4 = 0.73, F3,10 = 1.22, P = 0.35). Least chub rarely used metaphyton and only
occasionally used Chara sp. in treatments with or without mosquitofish (Fig. 3).
Interference Competition Experiment
Least chub spent more time stationary (t3 = -2.71, P = 0.07) and were more likely to
occupy cover habitat (t3 = 2.58, P = 0.08) in the presence of mosquitofish. Least chub spent an
average of 53 min (± S.E. = 8.8) in cover when mosquitofish were present and 27 min (± S.E. =
3.9) when they were absent. Similarly, they spent, on average, 7.3 min (± S.E. = 2.2) versus 0.75
min (± S.E. = 0.5) stationary when mosquitofish were present or absent, respectively.
Large mosquitofish were the dominant aggressor in all interactions. Large least chub and
small mosquitofish were approximately equal, whereas small least chub occupied the lowest
position in the dominance hierarchy. Although a total of 34 aggressive interactions were
recorded during all trials with mosquitofish, no aggressive interactions were observed between
least chub when mosquitofish were absent, and only two aggressive interactions occurred
between least chub when mosquitofish were present. Of the 34 interactions in treatments
containing mosquitofish, 22 (65 %) were initiated by mosquitofish, with 13 being directed
towards least chub. Mosquitofish (small or large) were successful, (i.e. they did not retreat and
18
retained the occupied space following a confrontation), in all interactions they initiated against
least chub. Large least chub were successful in all eight confrontations initiated against
mosquitofish. Small least chub initiated two aggressive interactions with mosquitofish (both
large) and were not successful in either attempt.
Discussion
Least chub and western mosquitofish provide an excellent example of the possible
interactions between invasive and native species with similar adult body sizes. We found
evidence for all five of the potential negative interactions between invasive species and native
juveniles (Table 1), although the evidence was stronger for interactions that are easier to
demonstrate (predation and interference competition). Together these interactions are complex
and depend on the life stage of both species. Large mosquitofish can eat smaller YOY least chub
and are aggressively dominant over all least chub size classes. Plus, smaller mosquitofish are
equal in dominance with the largest adult least chub. While predation probably accounted for
most of the mortality of small YOY least chub in our field enclosures, the effects of competition
were also apparent. Growth of surviving large YOY least chub was reduced by an average of
107% in treatments with mosquitofish compared to the control without mosquitofish. Our
laboratory experiments showed that YOY least chub responded to mosquitofish aggression by
remaining stationary for long periods, which would reduce their foraging opportunities (e.g.
Werner and Anholt 1996). It was not possible to separate the effects of interference competition
(increased time in plant refugia and remaining stationary) from exploitation competition,
especially because we did not measure differences in resource levels between enclosures with
19
and without mosquitofish. However, the reduced growth of large least chub in the presence of
mosquitofish indicates that one or both forms of competition had a strong effect.
When mosquitofish are present, perhaps least chub spend long periods in refugia (e.g.
dense vegetation) where food resources are reduced compared to open water habitats. This
hypothesis was supported by the observation that least chub in our aquarium experiments seldom
left refuge habitat (Potamogeton sp.) in the presence of mosquitofish, but seldom used it when
mosquitofish were absent. This shift in habitat use in the presence of large mosquitofish may
cause YOY least chub to compete with YOY mosquitofish that take refuge in the same habitat.
Future research should compare food availability and YOY least chub growth in and out of
refuge habitat with and without YOY mosquitofish.
This shift in habitat use of least chub in the presence of mosquitofish led us to
hypothesize that increased utilization of submersed aquatic vegetation (refugia) may increase
least chub exposure to invertebrate predators. However, there was no evidence that invertebrate
predators could efficiently attack and consume least chub in or out of aquatic vegetation under
laboratory conditions. The invertebrates were inactive and usually remained stationary in almost
all trials. Although evidence of an effect of invertebrates on least chub in the laboratory would
provide convincing evidence of the importance of this interaction, the absence of a response is
not sufficient to abandon the hypothesis that mosquitofish, in the field, could indirectly effect
least chub survival by increasing their exposure to invertebrate predators. It is possible that
invertebrate predators could actively forage on small fish at night, but remain inactive during the
day, especially if larger fish are present. Future studies could test this hypothesis by repeating
our laboratory experiments during the day and the night.
20
The overall effects of mosquitofish (predation and competition) depended on least chub
size. The effects of predation were most pronounced on small YOY least chub, whereas reduced
growth was evident for large YOY least chub. These data indicate that YOY least chub pass
through a period in which their small size makes them vulnerable to predation by mosquitofish.
This size window most likely starts at emergence and continues until they grow large enough in
depth that they exceed the gape-width of large mosquitofish. Further research is needed to
determine the exact sizes that define this period of vulnerability to predation. However, once
least chub reach a size refuge from predation they continue to be confronted by mosquitofish
aggression and interference competition. Although the threat of predation may decline with
growth in least chub, the effects of aggression might continue throughout its life cycle because
large mosquitofish can aggressively dominate all size classes of least chub. This aggression may
restrict habitat utilization and food availability throughout the life cycle of least chub.
Although we could not independently test for all single effects (e.g. we could not create
conditions where mosquitofish behaved as a predator but not a competitor), this study does
support our original hypothesis that the ability of native species to trade fitness components and
adjust to invasive species is limited when multiple negative interactions simultaneously affect
multiple size classes. Least chub juveniles appear to be figuratively caught in a vice between
predation and competition. Rapid growth and a larger juvenile body size could reduce the risk of
predation, but this requires energy that is probably not available because of the simultaneous
effects of competition. Instead, competition reduces growth and prolongs the period when
juvenile least chub are vulnerable to mosquitofish predation. A spatial or temporal refuge from
negative interactions would allow the time and/or energy needed for development of
characteristics (e.g. rapid growth) that would increase the survival of native species (e.g.
21
Chapman et al., 2003). However, this is often not possible, especially when native and invasive
species are similar in size and thus, view the world on similar spatial and temporal scales. The
result is often a rapid decline to extinction for native species. Least chub have shown such a
decline, and our results suggest mosquitofish are one of the primary causes.
The number of negative interactions between native species in the absence of invasive
species will vary among communities (e.g. Schoener 1986), and the effects of invasive species
(none, positive, and negative) on all species in a community can be difficult to predict (e.g.
Mittelbach and Chesson 1987). However, we suggest that the total number of negative
interactions that native species experience will increase with the colonization of invasive species
when native and invasive species are similar in body size. Because species perceive the world at
the same spatial and temporal scales if they are similar in size and mobility (MacArthur and
Levins 1964, Levins 1992, Addicott et al. 1987), the number of negative interactions between
invasive and native species should increase with similarity in size and mobility. Other studies
provide evidence that native species failed to coexist with invasive species of similar size and
mobility (see Fuller et al. 1999). For example, invasive mosquitofish have had their greatest
impact on similar-sized native fish of various species while having little impact on larger native
taxa (Myers 1967, Courtenay and Meffe 1989). Certainly, invasive species that are larger or
more mobile than native species can cause a reduction in native species through predatory
interactions (e.g. Elton 1958). But this is a single interaction and native species may trade
various fitness components to reduce the effects of predation and promote coexistence. For
example, if least chub were ten times smaller, then they might exploit smaller prey and habitats
unavailable to mosquitofish, and they might be able to achieve a higher reproductive potential.
Thus, although they would be vulnerable to predation by mosquitofish throughout their life, their
22
small size should allow them to find refuge from predation or competition in habitat unavailable
to the larger invasive species, which should allow coexistence. Multiple interactions are most
likely between similar-sized species. When native and invasive species view the world on
different scales, the number of simultaneous negative interactions decrease and the availability of
refugia increase, thus promoting coexistence and the time needed to make necessary
evolutionary adaptations.
Management efforts should focus on methods that reduce the number of negative
interactions between invasive and native species so that impacted life stages acquire sufficient
time to adjust to the pressures imposed by invasive species. When native juveniles are faced
with predation and competition an obvious management action would be to reduce the number of
invasive adults. Eradication of invasive species is a common management goal, but it is often an
overly daunting task, and in the case of some invasive species (i.e. mosquitofish), it is nearly
impossible. By focusing on efforts that reduce the number of negative interactions, management
goals may shift to practices that promote coexistence between native and invasive species. In the
case of least chub and mosquitofish, removal of large mosquitofish in the spring when least chub
are small and vulnerable could increase the potential of coexistence. Plus, habitat manipulations
that reduce the optimal habitat for mosquitofish, warm shallow water, could decrease
mosquitofish growth and the rate at which their populations increase in the spring and summer.
Although least chub and mosquitofish view the world on similar scales, they have different
fundamental niches because they have different evolutionary histories. Western mosquitofish
thrive in warm water because they evolved in subtropical environments (Courtenay and
Meffe1989), whereas least chub historically occupied a variety of habitats with both cool and
warm temperatures (Sigler and Miller 1987). Habitat manipulations that decrease warm waters
23
and increase cool temperatures, and the removal of large mosquitofish may provide sufficient
time for least chub to grow and reach a size refuge before predation and competition from
mosquitofish can diminish recruitment.
Acknowledgements
We gratefully acknowledge funding support from the Utah Division of Wildlife Resources and
logistical support from Jay Banta and the staff at the Fish Springs National Wildlife Refuge. We
thank Benjamin Shettell, Adrian Bell, and Amy Schweitzer for assistance in setting up our field
experiment. This study was conducted under the approval of the Institutional Animal Care and
Use Committee at Brigham Young University. Permits to collect the fish used in this study were
provided by the Utah Division of Wildlife Resources.
Literature Cited
Ashton PJ, Mitchell DS (1989) Aquatic plants: patterns and modes of invasion, attributes of
invading species and assessment of control programmes. In: Drake JA, Mooney HA, Di
Castri F, Groves RH, Kruger FJ, Rejmánek M, and Williamson M (eds) Biological
invasions: a global perspective. John Wiley & Sons, Chichester, pp 111-154
Altmann, J 1974. Observational study of behaviour: sampling methods. Behaviour 49:227-267
Barrier RFG, Hicks BJ (1994) Behavioral interactions between black mudfish (Neochanna
diversus) and mosquitofish (Gambusia affinis). Ecol Freshwat Fish 3:93-99
24
Belk MC, Whitney MJ, Schaalje GB (2001) Complex effects of predators: determining
vulnerability of the endangered June sucker to an introduced predator. Anim Conserv
4:251-256
Bertolino S, Genovesi P (2003) Spread and attempted eradication of the grey squirrel (Sciurus
carolinensis) in Italy, and consequences for the red squirrel (Sciurus vulgaris) in Eurasia.
Biol Conserv 109:351-358
Brown JH (1989) Patterns, modes, and extents of invasions by vertebrates. In: Drake JA,
Mooney HA, Di Castri F, Groves RH, Kruger FJ, Rejmánek M, and Williamson M (eds)
Biological Invasions: A Global Perspective. John Wiley & Sons, Chichester, pp 85-110
Chapman LJ, Chapman CA, Schofield PJ, Olowo JP, Kaufman L, Seehausen O, Ogutu-Ohwayo
R (2003) Fish faunal resurgence in Lake Nabnugabo, East Africa. Conserv Biol 17:500-
511
Courtenay, WR and GK Meffe. (1989) Small Fishes in Strange Places: A Review of Introduced
Poeciliids. In: GK Meffe and FF Snelson, Jr. (eds) Ecology and Evolution of Livebearing
Fishes (Poeciliidae). Prentice Hall, Englewood Cliffs, New Jersey, USA, pp 319-331
Crist L, Holden PB (1990) Aquatic biology study of a spring complex in Snake Valley, Utah
final summary report Pr-36-1. Bio/West Inc, Logan
25
Crivelli AJ (1995) Are fish introductions a threat to endemic freshwater fishes in the Northern
Mediterranean region. Biol Conserv 72:311-319
Elton, CS (1958) The ecology of invasions by animals and plants. University of Chicago Press,
Chicago
Galat DL, Robertson B (1992) Response of endangered Poeciliopsis-Occidentalis sonoriensis in
the Rio-Yaqui drainage, Arizona, to introduced Gambusia affinis. Environ Biol Fishes
33:249-264
Gamradt SC, Kats LB (1996) Effect of introduced crayfish and mosquitofish on California
newts. Conserv Biol 10:1155-1162
Goldberg DE, Werner PA (1983) Equivalence of competitors in plant communities: a null
hypothesis and a field experimental approach. Am J Bot 70:1098-1104
Goldsborough LG, Robinson GC (1996) Pattern in wetlands. In: RJ Stevenson, ML Bothwell,
RL Lowe (eds). Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, New
York, New York, USA, pp 77-117
Goodsell JA, Kats LB (1999) Effect of introduced mosquitofish on Pacific treefrogs and the role
of alternative prey. Conserv Biol 13:921-924
26
Gurevitch J, Morrison JA, Hedges LV (2000) The interaction between competition and
predation: a meta-analysis of field experiments. Am Nat 155:435
Holden P, White W, Sommerville G, Duff D, Gervais R, Gloss S. (1974) Threatened fishes of
Utah. Utah Acad Sci Arts Let 2:46-65
Hickman TJ (1989) Status report of the least chub (Iotichthys phlegethontis) prepared for the
U.S. Fish and Wildlife Service. Western Ecosystems, St. George
Kiesecker JM, Blaustein AR, Miller CL (2001) Potential mechanisms underlying the
displacement of native red-legged frogs by introduced bullfrogs. Ecology 82:1964-1970
Komak S, Crossland MR (2000) An Assessment of the introduced mosquitofish (Gambusia
affinis holbrooki) as a predator of eggs, hatchlings, and tadpoles of native and non-native
anurans. Wildl Res 27:185-189
Kotler BP, Holt RD (1989) Predation and competition: the interaction of two types of species
interactions. Oikos 54:256-260
Lydeard C, Belk MC (1993) Management of indigenous species impacted by introduced
mosquitofish: an experimental approach. Southwest Nat 38:370-373
27
Manly, BFJ (1993). Comments on the design and analysis of multiple-choice feeding-preference
experiments. Oecologia 93:149-152
McDonald AL, Heimstra NW, Damkot DK (1968) Social modification of agonistic behaviour in
fish. Anim Behav 16:437-441
Meffe GK (1985) Predation and species replacement in American Southwestern fishes: a case
study. Southwest Nat. 30:173-187
Mittelbach GG (1986) Predator-mediated habitat use: some consequences for species
interactions. Environ Biol Fishes 16:159-169
Mittelbach GG (1988) Competition among refuging sunfishes and effects of fish density on
littoral zone invertebrates. Ecology 69: 614-623
Mittelbach GG, Chesson PL (1987) Predation risk: indirect effects on fish populations. In: Sih A,
Kerfoot CW (eds) Predation: direct and indirect impacts on aquatic communities.
University Press of New England, Hanover, pp 315-332
Myers, GS (1967) Gambusia, the fish destroyer. Australian Zoologist 13:102
28
Ogutu-Ohwayo R, (1993) The effects of predation by Nile perch, Lates niloticus L., on the fish
of lake Nabugabo, with suggestions for conservation of endangered endemic cichlids.
Conserv Biol 7:701-711
Perkins MJ, Lentsch LD, Mizzi J (1998) Conservation Agreement and Strategy for Least Chub
(Iotichthys phlegethontis). Utah Division of Wildlife Resources, Salt Lake
Rincon PA, Correas AM, Morcillo F, Risueno P, Lobon-Cervia J (2002) Interaction between the
introduced Western mosquitofish and two autochthonous Spanish toothcarps. J Fish Biol.
61:1560-1585
Sakai AK (2001) The population biology of invasive species. Annu Rev Ecol Syst 32:305-332
SAS, I. (1997). SAS/STAT software: changes and enhancements through release 6.12. Cary,
North Carolina, USA, SAS Institute.
Savidge JA (1987) Extinction of an island avifauna by an introduced snake. Ecology 68:660-668
Schoener TW (1986) Overview: kinds of ecological communities--ecology becomes pluralistic.
In: Diamond J and TJ Case (eds) Community Ecology. Harper and Row, New York, pp
467-479
29
Sigler WF, Sigler JW (1987) Fishes of the Great Basin, a natural history. University of Nevada
Press, Sigler WF, Miller RR (1996) Fishes of Utah. University of Utah Press, Salt Lake
Simberloff D (1989) Which insect introductions succeed and which fail. In: Drake JA, Mooney
HA, Di Castri F, Groves RH, Kruger FJ, Rejmánek M, and Williamson M (eds)
Biological invasions: a global perspective. John Wiley & Sons, Chichester, pp 61-76
Soderback B (1995) Replacement of the native crayfish Astacus astacus by the introduced
species Pacifastacus leniusculus in a Swedish lake- possible causes and mechanisms.
Freshwat Biol 33:291-304
Tollrian R, Harvell CD (1999) The ecology and evolution of inducible defenses. Princeton
University Press, Princeton
Townsend CR (1996) Invasion biology and ecological impacts of brown trout Salmo trutta in
New Zealand. Biol Conserv 78:13-22
Wellborn GA (2002) Trade-off between competitive ability and antipredator adaptation in a
freshwater amphipod species complex. Ecology 83:129-136
Werner EE, Anholt BR (1996) Predator-induced behavioral indirect effects: consequences to
competitive interactions in anuran larvae. Ecology 77:157-169
30
Winkelman DL, Aho JM (1993) Direct and indirect effects of predation on mosquitofish
behavior and survival. Oecologia 96:300-303
Wissinger SA (1992) Niche overlap and the potential for competition and intraguild predation
between size-structured populations. Ecology 73:1431-1444.
31
Table 1. Potential negative interactions between native and invasive species with similar adult
body sizes. A minus sign (-) indicates interactions that could have a negative impact on
native species, while a zero (0) indicates an improbable interaction.
Invasive Adults Invasive Juveniles Mechanism of Interaction Native
Juveniles Native Adults
Native Juveniles
Native Adults
Predation - 0 0 0 Interference Competition - - - 0 Exploitation Competition - - - -
32
Table 2. Mean number of mosquitofish and standard deviations (in parentheses) in each
mosquitofish density treatment at the conclusion of the field experiment.
Treatment Initial Size of Least
Chub
Mean Number Large
Mosquitofish (SL >25mm)
Mean Number Small
Mosquitofish (SL <25mm)
Mean Total Number of Mosquitofish
Least Chub Alone Large 2 (0) 4.5 (0.7) 6.5 (0.7) Small 1 (2) 6.25 (6.7) 7.25 (5.7)
Low Mosquitofish Large 15 (9.9) 141 (188.1) 156 (197.9) Small 8.5 (6.8) 191 (170.9) 199.5 (176.8)
High Mosquitofish Large 107 (31.1) 46.5 (23.3) 153.5 (7.8) Small 116.5 (18.9) 46.5 (15.7) 163 (34.5)
33
Figure Legends
Figure 1. Mean survivorship of small (9-13 mm standard length (SL)) and large (23-28 mm SL)
young-of-the-year (YOY) least chub (Iotichthys phlegethontis) cohorts in enclosures with
different mosquitofish (Gambusia affinis) densities. Bars represent one standard error. Data
points are covering some standard error bars.
Figure 2. Mean growth as the increase in (A) wet mass and (B) length of small (9-13 mm SL)
and large (23-28 mm SL) YOY least chub cohorts in enclosures with mosquitofish present or
absent. Bars represent one standard error.
Figure 3. Mean percent time that small (11 to 14 mm SL) YOY least chub spent in four habitat
types (open water, Potamogeton sp., Chara sp., and metaphyton). Bars represent one standard
error.
34
Mosquitofish density
None Low High
Perc
ent s
urvi
val
0
20
40
60
80
100Large least chubSmall least chub
35
Treatment
Incr
ease
in le
ngth
(mm
)
-4
-2
0
2
4
6
8
10
12
14
Incr
ease
in w
et m
ass
(g)
0.0
0.1
0.2
0.3
0.4
0.5
Large least chubSmall least chub
Mosquitofish absent Mosquitofish present
A)
B)
36
Habitat type
Perc
ent t
ime
(min
)
0
20
40
60
80
100 No predatorsInvertebrate predatorsMosquitofish predatorsInvertebrate and mosquitofish predators
Open Chara Metaphytonsp.Potamogeton spp.