Comparisons of the food niches of three native and two introduced fish species in an Australian river

Synopsis The breadth, correspondence and overlap of the diets of the small and large size classes of the three native species (Galaxias occidentalis, Bostockia porosa and Edelia vittata) and two introduced species (Gambusia holbrooki and Perca fluviatilis) of fishes found in the shallows of the main channel and in the tributary creeks of a south-western Australian river have been compared in each season. Classification and ordination were used to examine the overall interrelationships of the diets across species, size groups, seasons and the location where the fish were caught (channel or creek). The smaller fish had a narrower dietary breadth than larger fish in the spring and summer, presumably reflecting the size limit imposed on prey size by their possession of a relatively small mouth in these seasons. Intraspecific dietary overlap between large and small size classes was usually high in G. occidentalis, but generally low in G. holbrooki and P. fluviatilis, and also in B. porosa and E. vittata when the difference between the lengths of the two size groups was greatest. Dietary overlap was least in autumn when the main prey taxa were most abundant. During winter, the diets of the three native species in the tributaries converged, probably reflecting a relatively low faunal diversity in these highly seasonal water bodies. The only relatively consistent interspecific overlap in diet was between B. porosa and E. vittata. Classification and ordination of the dietary samples separated G. occidentalis (which fed extensively on terrestrial organisms from the water surface) from the smaller P. fluviatilis (that concentrated on copepods in the plankton) and from B. porosa and E. vittata (which ingested primarily benthic organisms). Furthermore, B. porosa tended to ingest larger prey taxa than E. vittata. The diet of Gambusia holbrooki is sometimes dominated by terrestrial insects and at other times by benthic organisms, demonstrating that this species is an opportunistic carnivore. It is concluded that food partitioning by the three native and two introduced fish species in the Collie River is likely to be one of the principal factors facilitating the coexistence of substantial populations of these species in this system.     

Effects of flooding on recruitment and dispersal of the Southern Pygmy Perch (Nannoperca australis) at a Murray River floodplain wetland

Summary With limited evidence linking Australia's Murray‐Darling Basin fish species and flooding, this study assessed annual variation in abundance and recruitment levels of a small‐bodied, threatened floodplain species, the Southern Pygmy Perch (Nannoperca australis), in floodplain habitats (creeks, lakes and wetlands) in the Barmah‐Millewa Forest, Murray River, Australia. Spring and summer sampling over a 5‐year period encompassed large hydrological variation, including 1 year of extended floodplain inundation which was largely driven by an environmental water release, and 2 years of severe regional drought. Recruitment and dispersal of Southern Pygmy Perch significantly increased during the floodplain inundation event compared with the other examined years. This study provides valuable support for an environmental water allocation benefiting a native species, and explores the link between flooding and its advantages to native fish. This suggests that the reduced flooding frequency and magnitude as a result of river regulation may well be a major contributing factor in the species’ decline in the Murray‐Darling Basin.     

Net design for selective control of the “plague minnow” Gambusia holbrooki that minimises impact on native Australian fishes

Gambusia holbrooki is one of the world's most environmentally damaging introduced species, being notoriously difficult to control once established. A composite double-winged fyke net comprising four vertically stacked compartments was developed to determine the potential to control G. holbrooki, while reducing negative interactions of this aggressive species with small threatened fishes. The stacked fyke net captured three times as many G. holbrooki as a conventional fyke net while maintaining consistent catches of native fishes relative to that from a conventional fyke net, and detected species-specific vertical distributions. This stratified net design represents a valuable management option for controlling this agonistic species or for limiting antagonistic interactions between G. holbrooki and native species during typical fyke sampling of native ecosystems.     

The influence of planktivorous fish on zooplankton communities in riverine slackwaters

1. Contrary to that for lakes and ponds, our knowledge of the influence of planktivorous fish on zooplankton communities in rivers is slight, largely because of the general assumption that such communities are overwhelmingly regulated by physical conditions. 2. In two separate but concurrent in situ enclosure experiments, we investigated the effects of carp gudgeon (Hypseleotris spp.) and Eastern Gambusia (Gambusia holbrooki) on zooplankton communities in slackwaters of a temperate Australian floodplain river. 3. A high biomass of Hypseleotris suppressed the density of daphniid microcrustaceans, but enhanced the total density of rotifers. A high biomass of Gambusia, on the other hand, suppressed the total density of both microcrustaceans and rotifers. 4. A high biomass of planktivorous fish also reduced the density of many of the ovigerous (egg‐carrying) zooplanktonic taxa. Indeed, ovigerous cyclopoid copepods were suppressed in the presence of a high biomass of Hypseleotris, even though there was no significant effect on overall (ovigerous plus non‐ovigerous) density. 5. Our results imply that a high biomass of planktivorous fish can potentially influence zooplankton communities in riverine slackwaters, as in many lakes and ponds.     

Diversity and fine-scale spatial genetic structure of Cyperus papyrus populations in Lake Naivasha (Kenya) using microsatellite markers

Species introduced outside their native range to serve anthropogenic purposes may have unintended consequences on native organisms. Mosquitofish (Gambusia affinis and Gambusia holbrooki) have been introduced throughout the world to control larval mosquito populations in aquatic environments, but they have also been implicated in the decline of native fish, amphibian, and aquatic invertebrate populations. We investigated the roles of introduced western mosquitofish (G. affinis), and two naturally colonizing predators (crayfish and dragonfly naiads) in inflicting tail injuries observed on ranid frog tadpoles in experimental constructed wetlands. We also examined the influence of vegetation in reducing tail injury severity. We found that mosquitofish significantly increased the prevalence and severity of tail injuries, but crayfish and dragonfly naiads contributed much less to increased injury levels. Furthermore, increased vegetation cover did not significantly attenuate tail injuries. However, after chemical removal of Gambusia, injury prevalence and severity was significantly reduced. Although previous investigations have concluded that mosquitofish prey upon the eggs and larvae of some amphibians, our results illustrate that these fish can cause substantial but apparently sub-lethal injuries to large numbers of larval amphibians in a wetland. Further investigations are needed to determine if these injuries impede the fitness of victims and lead to population reductions.     

Effects of a non-native cichlid fish (African jewelfish, Hemichromis letourneuxi Sauvage 1880) on a simulated Everglades aquatic community

In an 8-month mesocosm experiment, we examined how a simulated Everglades aquatic community of small native fishes, snails, and shrimp changed with the addition of either a native predator (dollar sunfish Lepomis marginatus) or a non-native predator (African jewelfish Hemichromis letourneuxi) compared to a no-predator control. Two snail species (Planorbella duryi, Physella cubensis) and the shrimp (Palaemonetes paludosus) displayed the strongest predator-treatment effects, with significantly lower biomasses in tanks with Hemichromis. One small native fish (Heterandria formosa) was significantly less abundant in Hemichromis tanks, but there were no significant treatment effects for Gambusia holbrooki, Jordanella floridae, or Pomacea paludosa (applesnail). Overall, there were few treatment differences between native predator and no-predator control tanks. The results suggest that the potential of Hemichromis to affect basal food-web species that link primary producers with higher-level consumers in the aquatic food web, with unknown consequences for Florida waters.     

Effects of flooding on recruitment and abundance of Golden Perch (Macquaria ambigua ambigua) in the lower River Murray

Flooding is often considered a stimulus for production of fish in floodplain rivers. In the southern Murray–Darling Basin (MDB), Australia, however, few native fish species have been shown to use the floodplain for spawning, and recruitment has been positively and negatively associated with flooding. In 2010/11, extensive flooding in the lower River Murray provided an opportunity to investigate the recruitment response of Golden Perch (Macquaria ambigua ambigua) following 10 years of drought and floodplain isolation. Annual variation in Golden Perch abundance and recruitment were investigated in anabranch and main channel habitats at Chowilla in the floodplain geomorphic region of the lower River Murray over a 7‐year period incorporating the flood and 6 years of in‐channel flow. Spatial variation in recruitment in the lower River Murray was also investigated by comparing the age structure of Golden Perch in the swamplands/lakes, gorge and floodplain geomorphic regions. Golden Perch abundance in the Chowilla region increased significantly postflooding compared with drought years. Age structures indicated that increased abundance was due predominantly to fish spawned during the flood (2010/11) and the previous year (2009/10), which was characterised by in‐channel flows. Age structure was similar in the nearby Katarapko Anabranch system indicating a uniform postflood recruitment response in the floodplain geomorphic region. Juvenile Golden Perch from the 2010/11 and 2009/10 cohorts were less apparent in the gorge and swamplands/lakes regions. Golden Perch have flexible life histories and will spawn and recruit in association with in‐channel rises in flow and overbank flows, but significant increases in abundance in the lower River Murray may result from overbank flooding. Contemporary approaches to flow restoration in the MDB emphasise overbank flows and floodplain processes. We suggest, however, that environmental flow management that incorporates floodplain and in‐channel processes, at appropriate spatio‐temporal scales, will result in more robust populations of Golden Perch.     

Response of Endangered Desert Fish Populations to a Constructed Refuge

The U.S. Bureau of Reclamation created a shallow, 110-m channel to provide habitat for two endangered fishes, Cyprinodon elegans (Comanche Springs pupfish) and Gambusia nobilis (Pecos gambusia), at the site of the fishes' former natural habitat. The ciénega (marsh) associated with Phantom Lake Spring in Jeff Davis County, Texas, was destroyed by the creation of an irrigation canal system. In 1993, the endangered fishes were stocked into the refuge with individuals from the irrigation canals, and in the case of C. elegans, hatchery stocks. The condition of habitat, status of fish populations, and fish ecology within the refuge were then monitored for two years. The abundance and density of both species increased in accordance with aquatic plant development. Cyprinodon elegans abundance peaked after one year and stabilized at an average density of 14.7/m² by the end of our study. Juvenile C. elegans were always rare, which may indicate that the population reached the refuge's carrying capacity and that recruitment is low. Gambusia nobilis was the most abundant fish in the refuge (average density 96/m²), used the entire refuge, and outcom-peted nonindigenous G. geiseri. The two Gambusia species used similar habitats but showed almost no dietary overlap. High densities of aquatic plants reduced the amount of open water areas necessary for C. elegans. The refuge will sustain the two endangered fishes at this historic site of endemism while maintaining flow to the irrigation system; however, the refuge is not equivalent to a restored ciénega.     

Summer co‐existence of small‐sized cyprinid and percid individuals in natural and impounded stretches of a lowland river: food niche partitioning among fishes

Due to changes of discharge regime downstream of a dam reservoir, an alluvial natural stretch of the Warta River changed to a macrophyte‐dominated ecosystem. Large patches of submersed, aquatic macrophytes appeared in summer and their effect is analysed in this study. These patches contained enriched macroinvertebrate assemblages (epiphyton and benthos) and they were refuge for both zooplankton and young fishes released from the reservoir. Despite these altered conditions in this stretch, roach Rutilus rutilus, perch Perca fluviatilis and ruffe Gymnocephalus cernua dominated, as they did in the natural backwater. Fishes were sampled every 2 weeks from June to August, together with their food resources to assess the partitioning of the diet among small individuals of the three species in both stretches (the natural and affected ones). The aim of the analysis was to answer how animal food associated with water plants was partitioned between the species. In both stretches, G. cernua were primarily benthivorous, but epiphytic fauna, zooplankton and large‐sized benthic chironomid larvae replaced lack of many large, benthic insects in the tailwater. Levins' food breath index decreased from 0·36 in the backwater to 0·29 in the tailwater. An opposite trend was observed for P. fluviatilis occurring among macrophytes. Perca fluviatilis were competitors of R. rutilus and took food not only in or on the river bed, but also in the water column. They ate zooplankton and epiphytic fauna and Levins' index increased from 0·32 to 0·44 in the tailwater. Rutilus rutilus fed on adult insects, algae and plant fragments in the natural stretch. In the tailwater, these food types were chiefly complemented by zooplankton. Despite this, the niche breadth of R. rutilus was similar at the two sites. Abundance of food associated with the macrophytes appeared to facilitate cohabitation in the abundant fish populations, but P. fluviatilis appeared to benefit the most in the altered river stretch.     

Low diversity and high levels of population genetic structuring in introduced eastern mosquitofish (<Emphasis Type="Italic">Gambusia holbrooki</Emphasis>) in the greater Melbourne area,Australia

Eastern mosquitofish (Gambusia holbrooki) were introduced into Australia in 1925 and released to control mosquitoes. Gambusia holbrooki rapidly became invasive in recipient environments and now threaten native fauna. In this study, we used five polymorphic microsatellite loci and sequence from two mitochondrial genes, cytochrome b and cytochrome oxidase I, to evaluate genetic variation, colonisation and movement patterns of introduced G. holbrooki in the greater Melbourne area, and to assist in identifying the feasibility of local eradication. Microsatellite variation was consistently low within populations and there was evidence of bottleneck events for several populations. Populations displayed significant structuring associated with river basins rather than geographic distance, suggesting that habitat connectivity is important for dispersal. However, a few populations within river basins were more closely related to populations in other river basins than within their own basin, most likely reflecting a role of human-assisted dispersal in population establishment. Mitochondrial sequencing revealed only a single haplotype and suggested all populations were founded by individuals from a common source. These genetic data help delineate boundaries for local management strategies.     

Multi-species monitoring of rare wetland fishes should account for imperfect detection of sampling devices

Accuracy in estimating occupancy of a threatened species is important for conservation but false absences bias many monitoring programs. Imperfect detection is especially relevant to surveys of rare wetland fishes which are often small-bodied and cryptic. Many factors influence probability of detection, including fish size and abundance, habitat characteristics and sampling devices. Imperfect detection can be addressed by accounting for probability of detection when estimating occupancy by modelling detection/non-detection data collected in replicate surveys. Three ecological specialists were once common in habitats associated with Lake Alexandrina at the terminus of the Murray–Darling Basin, Australia. The threatened Murray Hardyhead (Craterocephalus fluviatilis), Southern Pygmy Perch (Nannoperca australis) and Yarra Pygmy Perch (N. obscura) are now rare in the region following population collapses during a prolonged drought, and ongoing monitoring aims to assess their statuses for management purposes. This study compares probability of detection of the rare wetland fishes and cohabiting species during 2 years of multi-species monitoring using contrasting sampling devices (fyke and seine). The findings suggest large variations in estimated probability of detection can occur between devices for Murray Hardyhead and Southern Pygmy Perch. Yarra Pygmy Perch was undetected during the study. Overall, the findings show multi-species monitoring programs using a single sampling device may wrongly estimate the occupancy of a target fish. By accounting for imperfect detection, multi-species monitoring programs will improve inferences regarding population status, recovery and habitat quality of fishes to more accurately inform wetland management.     

Behavioural plasticity in a native species may be related to foraging resilience in the presence of an aggressive invader

Competition between invasive and native species can result in the exploitation of resources by the invader, reducing foraging rates of natives. However, it is increasingly recognized that multiple factors can enhance the resilience of native species competing for limiting resources with invaders. Although extensively studied in terrestrial species, little research has focused on behavioural plasticity in aquatic ecosystems and how this influences native species resilience. Here, we examined the role of behavioural plasticity in interactions between a native Australian fish, Pseudomugil signifer, and a widespread invasive fish, Gambusia holbrooki. To determine whether P. signifer displays behavioural plasticity that may mitigate competition with G. holbrooki, we first quantified social behaviours (aggression, submission and affiliation) and shoal cohesion for each species in single- and mixed-species groups. Second, we compared the feeding rates of both species in these groups to ascertain if any modulation of social behaviours and cohesion related to foraging success. We found that aggressive and submissive behaviours of G. holbrooki and P. signifer showed plasticity in the presence of heterospecifics, but social affiliation, shoaling and, most importantly, foraging, remained inflexible. This variation in the degree of plasticity highlights the complexity of the behavioural response of a native species and suggests that both behavioural modulation and consistency may be related to sustaining foraging efficiency in the presence of an invader.     

Competitive and predatory interactions between invasive mosquitofish and native larval newts

Invasive fish have a high disruptive potential in aquatic ecosystems, in which amphibians may be highly impacted due to intense competition and/or predation on their eggs and larvae. Most studies have focused on the effect of large invasive fishes such as salmonids, whereas the effect of smaller fish on amphibians has been seldom investigated. We experimentally studied effects of the invasive Eastern mosquitofish (Gambusia holbrooki) on pygmy newts (Triturus pygmaeus), a species endemic to the Iberian Peninsula. We set up outdoor mesocosms in Doñana National Park with native aquatic flora and invertebrate fauna, and containing larval newts at two experimental densities. Density of larval newts was also crossed with presence or absence of mosquitofish, either free-swimming or caged, in order to distinguish consumptive and non-consumptive effects. Increased density of coexisting larval newts did not reduce their survival, but reduced their growth and development. Newt survival and size at metamorphosis were dramatically reduced in the presence of free-swimming mosquitofish, whether at low or high fish densities. Caged mosquitofish, however, had no effect on larval newts. In laboratory trials, mosquitofish preyed more efficiently on insect larvae than did larval pygmy newts, highlighting the high competitive potential of mosquitofish. This was confirmed by the depletion of zooplankton that free fish caused in the experimental outdoor mesocosms. Our study suggests that invasive mosquitofish exert a high negative impact on coexisting newt populations. Such effects can be explained by a combination of direct predation, injuries caused by predation attempts, and intense competitive exploitation of common food resources.     

Heat and hypoxia give a global invader, <Emphasis Type="Italic">Gambusia holbrooki,</Emphasis> the edge over a threatened endemic fish on Australian floodplains

Deciphering the mechanisms by which climate change interacts with invasive species to affect biodiversity is a major challenge of global change biology. We conducted experiments to determine whether the global invader, Gambusia holbrooki, was more resistant to high water temperature (heat) and low dissolved oxygen (hypoxia) than a threatened native fish, Nannoperca australis. Metabolic experiments conducted at 25 and 29 °C showed that G. holbrooki had at least four times the capacity for metabolic depression during hypoxia than N. australis. An increase in environmental temperature from 25 to 29 °C had no significant impact on the critical oxygen tension, P _crit, of G. holbrooki, but significantly and strongly increased P _crit of N. australis. Gambusia holbrooki also had a lower Q ₁₀ of standard metabolic rate than N. australis. Our results indicate that G. holbrooki have physiological traits conferring greater resistance to hypoxia than N. australis, and as temperature increases, the resistance of N. australis to hypoxia was more eroded than that of G. holbrooki. Intensive monitoring of the temperature and dissolved oxygen dynamics of wetlands showed that contemporary heat waves are already causing conditions that might give G. holbrooki the edge over N. australis on Australian floodplains. Our study adds weight to recent anecdotal reports of drought and heat waves causing localised extinction of N. australis, but the proliferation of G. holbrooki.     

River flow,zooplankton and dominant zooplanktivorous fish dynamics in a warm‐temperate South African estuary

The possible links between river flow, zooplankton abundance and the responses of zooplanktivorous fishes to physico‐chemical and food resource changes are assessed. To this end, the seasonal abundance, distribution and diet of the estuarine round‐herring Gilchristella aestuaria and Cape silverside Atherina breviceps were studied in the Kariega Estuary. Spatio‐temporal differences were determined for selected physico‐chemical variables, zooplankton abundance and zooplanktivorous fish abundance and distribution. Results indicated that, following a river flood event in winter (>30 m³ s^?1), altered physico‐chemical conditions occurred throughout the estuary and depressed zooplankton stocks. Abundance of G. aestuaria was highest in spring, with this species dominant in the upper and middle zones of the estuary, while A. breviceps was dominant in summer and preferred the middle and lower zones. The catch per unit of effort of both zooplanktivores also declined significantly following the flooding, thus suggesting that these fishes are reliant on zooplankton as a primary food source for healthy populations. Copepods dominated the stomach contents of both fish species, indicating a potential for strong interspecific competition for food, particularly in the middle reaches. Temporal differences were evident in dietary overlap between the two zooplanktivorous fish species and were correlated with river flow, zooplankton availability and fish distribution. The findings of this study emphasize the close trophic linkages between zooplankton and zooplanktivorous fishes under changing estuarine environmental conditions, particularly river flow and provide important baseline information for similar studies elsewhere in South Africa and the rest of the world.     

Experimental evidence for the use of artificial refugia to mitigate the impacts of invasive <Emphasis Type="Italic">Gambusia holbrooki</Emphasis> on an endangered fish

Habitat degradation has a major impact on freshwater ecosystems and also facilitates biological invasions thus intensifying the problem. The survival of native species under threat from invaders can be improved either by eradicating the invading species or by providing resources or conditions that benefit the native species. One such resource is shelter and in degraded habitats artificial refugia may be a viable option. However, the use of artificial refugia to promote coexistence between native and invasive species remains poorly understood. We assessed the potential for artificial refugia to ameliorate the disruption of social interactions in an endangered native Iberian toothcarp, Aphanius iberius, by the invasive mosquitofish, Gambusia holbrooki. We found that mosquitofish do not compete for access to refugia despite their higher level of aggression compared to native fish. Native fish use refugia more overall, particularly in the presence of mosquitofish. Despite this, the benefits of refugia are not clear cut: increases in refuge use by male toothcarp induced by mosquitofish aggression correspond to decreases in attention to conspecifics. However, changes in refugia use over time together with constant attention to conspecifics indicates that it is not refugia use itself that disrupts social interactions but the interrelated effects of mosquitofish aggression. Provision of artificial refugia in degraded freshwater ecosystems may thus be a viable management tool to protect native populations under imminent threat of invasion.     

Temporal genetic dynamics among mosquitofish (<Emphasis Type="Italic">Gambusia holbrooki</Emphasis>) populations in invaded watersheds

The temporal components of genetic diversity and geographical structure of invasive mosquitofish populations are poorly known. Through the genetic monitoring of four consecutive cohorts of Gambusia holbrooki from three different river basins we aimed to determine temporal patterns of regional genetic variation and dispersal rates within invasive populations. Despite showing evidence of strong population size fluctuations, genetic diversity levels were maintained among local cohorts. We only detected temporal allele frequency changes associated with seasonal flooding that did not modify major trends on population structure among cohorts. Downstream gene flow coupled with increased connectivity at lowland locations to increase genetic diversity levels in these areas. A large proportion of local fish (up to 50 %) were dispersers, often originated from locations within the same river basin. High dispersal capability, ecological tolerance, and reproductive traits likely promote river colonization. Finally, our results also confirmed that human-assisted translocations promote within and among basin gene flow and maintained levels of genetic diversity, particularly in upstream locations.     

Habitat preferences and habitat restoration options for small‐bodied and juvenile fish species in the northern Murray–Darling Basin

Degradation of instream habitats in the northern Murray–Darling Basin has occurred through numerous stressors, including siltation, clearing of bankside vegetation, intrusion of livestock and impacts of pest species. A better understanding of habitat preferences of native fish species could help guide future instream habitat restoration actions. The habitat choices of seven native fish species, juvenile Murray Cod (Maccullochella peelii), juvenile Golden Perch (Macquaria ambigua ambigua), juvenile Silver Perch (Bidyanus bidyanus), adult Murray–Darling Rainbowfish (Melanotaenia fluviatilis), adult Olive Perchlet (Ambassis agassizii), adult Un‐specked Hardyhead (Craterocephalus stercusmuscarum fulvus) and adult carp gudgeons (Hypseleotris spp.) were tested in preference troughs to help inform potential habitat restoration actions in the Condamine catchment. Each species was given a choice between pair combinations of open sandy habitat, submerged macrophytes, emergent plants and rocky rubble. Habitat preferences varied between species. Murray Cod, Golden Perch, carp gudgeons and Olive Perchlets preferred structure over open sandy habitat, whilst juvenile Silver Perch, Un‐specked Hardyhead and Murray–Darling Rainbowfish did not avoid open sandy habitats. Juvenile Murray Cod preferred rocky rubble habitat over all other habitat choices. Use of complex rock piles to provide nursery habitat for Murray Cod populations is a potential restoration option. Introduction of rock could also benefit Golden Perch and carp gudgeons. Use of emergent plants, submerged macrophytes and rocky rubble for habitat restoration all appear to have merit for one or more species of small‐bodied fishes or juvenile stages of larger sized fishes. Rocky rubble or floating attached macrophytes could be viable restoration options in areas too turbid to establish submerged macrophytes. These habitat interventions would complement existing actions such as re‐snagging and provision of fish passage to assist with sustainable management of native fish populations.     

Diet of Murray cod (<Emphasis Type="Italic">Maccullochella peelii peelii</Emphasis>) (Mitchell) larvae in an Australian lowland river in low flow and high flow years

Researchers have hypothesised that influxes of pelagic zooplankton to river channels after floods and high flows are necessary for strong recruitment of some native fish species, including Murray cod (Maccullochella peelii peelii) (Mitchell), in the Murray–Darling river system, Australia. This study investigated the composition of the diet and gut fullness of drifting Murray cod larvae weekly during two spawning seasons with contrasting flows, to determine if pelagic zooplankton comprised a greater proportion of the gut contents and guts were fuller in a high flow (2000) than in a low flow (2001) year. Gut fullness and yolk levels of 267 larvae were ranked, and prey identified to family level. Approximately 40 and 70% of individuals had been feeding in 2000 and 2001, respectively. Gut fullness increased with declining yolk reserves. Larvae in both the years had an almost exclusively benthic diet, irrespective of the flow conditions at the time. Substantial inundation of dry ground in 2000, albeit restricted to in-channel benches, anastomosing channels and oxbow lakes, did not lead to an influx of pelagic, floodplain-derived zooplankton subsequently exploited by Murray cod larvae. These results have the implications for the management of regulated temperate lowland rivers: high flows cannot automatically be assumed to be beneficial for the fish larvae of all species and their food resources, and caution should be exercised with the timing of flow releases.     

Far from home: responses of an American predator species to an American prey species in a jointly invaded area of Australia

Far from their native ranges in the Americas, two invasive species come into contact in Australian waterbodies. Cane toads (Rhinella marina) fatally poison many anurophagous predators, whereas eastern mosquito fish (Gambusia holbrooki) voraciously consume anuran larvae. As cane toads spread south along Australia’s east coast, they are colonizing areas where mosquito fish are abundant. What happens when these two American invaders encounter each other in Australia? We tested the responses to toad tadpoles of mosquito fish from populations that were sympatric versus allopatric with cane toads. Toad-sympatric fish generally ignored toad tadpoles. Toad-allopatric fish initially consumed a few tadpoles, but rapidly developed an aversion to these toxic prey items. The laboratory-reared progeny of toad-allopatric fishes were more likely to approach toad tadpoles than were the offspring of toad-sympatric fishes, but the two groups learned toad-avoidance at similar rates. Thus, mosquito fish show an innate aversion to cane toad tadpoles (perhaps reflecting coevolution with North American bufonid taxa), as well as an ability to rapidly learn taste-aversion. Our comparisons among populations suggest that several decades of toad-free existence in Australia caused a decline in the fishes’ innate (heritable) aversion to toads, but did not affect the fishes’ capacity to learn toad-avoidance after an initial exposure. Any impact of mosquito fish on cane toads thus is likely to be transitory. The rapid (<100-year) time frame of these shifts (the initial weakening of the fishes’ response during toad-allopatry, and its recovery after secondary contact) emphasizes the dynamic nature of faunal responses during biological invasions, and the interplay between adaptation and phenotypic plasticity.