Journal of Animal Ecology 2011, 80, 990–998 doi: 10.1111/j.1365-2656.2011.01834.x Introduced brown trout alter native acanthocephalan infections in native fish Rachel A. Paterson1*, Colin R. Townsend1, Robert Poulin1 and Daniel M. Tompkins2 1 Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand; and 2Landcare Research, Private Bag 1930, Dunedin 9054, New Zealand Summary 1. Native parasite acquisition provides introduced species with the potential to modify native host–parasite dynamics by acting as parasite reservoirs (with the ‘spillback’ of infection increasing the parasite burdens of native hosts) or sinks (with the ‘dilution’ of infection decreasing the parasite burdens of native hosts) of infection. 2. In New Zealand, negative correlations between the presence of introduced brown trout (Salmo trutta) and native parasite burdens of the native roundhead galaxias (Galaxias anomalus) have been observed, suggesting that parasite dilution is occurring. 3. We used a multiple-scale approach combining field observations, experimental infections and dynamic population modelling to investigate whether native Acanthocephalus galaxii acquisition by brown trout alters host–parasite dynamics in native roundhead galaxias. 4. Field observations demonstrated higher infection intensity in introduced trout than in native galaxias, but only small, immature A. galaxii were present in trout. Experimental infections also demonstrated that A. galaxii does not mature in trout, although parasite establishment and initial growth were similar in the two hosts. Taken together, these results support the hypothesis that trout may serve as an infection sink for the native parasite. 5. However, dynamic population modelling predicts that A. galaxii infections in native galaxias should at most only be slightly reduced by dilution in the presence of trout. Rather, model exploration indicates parasite densities in galaxias are highly sensitive to galaxias predation on infected amphipods, and to relative abundances of galaxias and trout. Hence, trout presence may instead reduce parasite burdens in galaxias by either reducing galaxias density or by altering galaxias foraging behaviour. Key-words: dilution effect, disease, galaxias, introduced species, New Zealand, parasite, salmonid Introduction Introduced species have the potential to modify native host– parasite dynamics through the acquisition of native parasites. Recent reviews have demonstrated that introduced animals may acquire about five native parasites on average, with some gaining more than 16 (Torchin et al. 2003; Kelly et al. 2009b). These reviews highlight that not only do a great variety of introduced species acquire native parasites, but also that the acquired parasites are highly diverse. The frequency with which introduced species acquire native parasites has led to increasing attention to the ways they can modify native parasite dynamics. Parasite ‘spillback’ to a native host may result from the acquisition of native parasites by an introduced species, provided that it becomes a *Correspondence author. E-mail: patra374@student.otago.ac.nz competent host that can act as a reservoir of infection (Daszak, Cunningham & Hyatt 2000; Tompkins & Poulin 2006; Kelly et al. 2009b). The spillback of parasites may act as an additional threat to native biodiversity via increased disease impacts on species already facing mounting pressures from other sources such as competition, predation, habitat loss, pollution and climate change. Recent studies have also documented how the acquisition of native parasites by introduced species may lead to parasite ‘dilution’, whereby native hosts experience lower parasite burdens in the presence of introduced hosts (Telfer et al. 2005; Kelly et al. 2009a; Thieltges et al. 2009). While reduced parasite burdens in native hosts may appear of less concern, alterations to processes that regulate population densities may have flow-on effects on native species dynamics, such as the modification of competitive advantage between conspecifics (Kopp & Jokela 2007), which could result in subsequent alterations to  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society Trout alter native host–parasite dynamics 991 community dynamics and ecosystem functioning. The frequency with which introduced species acquire native parasites and attain high population densities means there is much potential for either parasite spillback or dilution to occur. Our current understanding of the potential impacts of introduced species on native host–parasite interactions relies largely on field-based observations of introduced populations, using a combination of parasite fitness parameters at individual (e.g. size, fecundity) and population scales (e.g. prevalence, infection intensity, abundance; Bush et al. 1997) to assess the competency of introduced species as native parasite hosts (Rauque, Viozzi & Semenas 2003). Such assessments have been made with or without sympatric native host populations (Galli et al. 2005), and occasionally native parasite dynamics have been surveyed within native-only host populations to provide baselines from which the effects of an introduced host are judged (Dubois, Marcogliese & Magnan 1996; Kelly et al. 2009a). However, native populations free from the effects of introduced species are not always available for such comparisons to be made. While field-based observations are essential to ascertain which native parasites have been acquired by introduced species and to identify patterns, differences in parasite fitness parameters between native and introduced hosts may not provide a true reflection of a parasite’s overall performance. Field observations can obscure differences in establishment success, survival ⁄ mortality, maturation, growth and relative fecundity. For example, poor host competency of an introduced host in terms of parasite establishment could be masked if this species was more likely to encounter infectious stages, resulting in similar levels of prevalence and ⁄ or infection intensity in native and introduced populations. Similarly, comparable numbers of mature parasites in native and introduced hosts might suggest equal compatibility for a native parasite in both hosts, but field observations of parasite maturity cannot take into account differing maturity rates between native and introduced hosts. Recognizing the potential limitations of field observations, some researchers have used experimental infection trials to allow differences in parasite fitness between hosts to be verified under standard conditions (King, van Oosterhout & Cable 2009; Thieltges et al. 2009). However, even this approach fails to account for the dynamics of the whole host–parasite system in terms of relative abundances of host populations, infection encounter rates by intermediate hosts and potential host behavioural changes caused by the presence of an introduced species. For example, hosts that occur at high densities are more likely to influence parasite dynamics than less common hosts. Here, we provide an in-depth assessment of parasite fitness in native and introduced hosts in a system previously identified as one in which the acquisition of native parasites by an introduced species results in the dilution of parasites in native hosts (Kelly et al. 2009a). We utilize a multiple-scale approach of field observations, experimental trials and dynamic popula- tion modelling to examine this complex system and identify mechanisms responsible for observed field patterns. Materials and methods STUDY SYSTEM In New Zealand, the introduced brown trout (Salmo trutta L.) is a dominant freshwater fish considered responsible for the decline and local extinction of native fish species (Townsend 2003). Since its introduction as parasite-free eggs in 1867 (MacCrimmon & Marshall 1968), more than 10 native parasite species have been acquired by brown trout, including the generalist freshwater acanthocephalan Acanthocephalus galaxii Hine (Hine, Jones & Diggles 2000). In the Otago province of the South Island of New Zealand, this acanthocephalan has been reported from sympatric stream populations of native roundhead galaxias (Galaxias anomalus Stokell) and brown trout (Kelly et al. 2009a). The parasite matures in the intestine of the definitive fish host and releases eggs into the water column via fish faeces. Larval cystacanth stages develop in the intermediate amphipod host Paracalliope fluviatilis Thomson after egg ingestion, with the life cycle of A. galaxii completed following predation of infected amphipods by the definitive host (Hine 1977). Brown trout are a major predator of galaxiids (Crowl, Townsend & McIntosh 1992; Townsend 1996) and thus may acquire additional A. galaxii infections via postcyclic transmission when roundhead galaxias are eaten. This additional transmission route has been reported in many acanthocephalan species, including native acanthocephalans transmitted from native galaxias to introduced trout in Argentina (Rauque, Semenas & Viozzi 2002), but has not been experimentally demonstrated for A. galaxii. Preliminary field observations suggest that A. galaxii attains equal or higher prevalence and infection intensity in brown trout (Kelly et al. 2009a), indicating the potential for modification of parasite–host dynamics by this introduced host. FIELD SURVEYS Field surveys were conducted to determine the prevalence (the percentage of infected hosts) and infection intensity (the number of worms per host; sensu Bush et al. 1997) of A. galaxii in its definitive and intermediate host species, and the relative densities of these hosts at three sites previously identified by Kelly et al. (2009a) as ‘sympatric’ for roundhead galaxias and brown trout infected with A. galaxii: Swin Burn Q1 (2296¢90¢¢E, 5558¢70¢¢N), Swin Burn Q2 (2298¢30¢¢E, 5559¢40¢¢N) and Old Hut Creek (2294¢00¢¢E, 5568¢73¢¢N), all in the upper Taieri River catchment in Otago. Each site was visited on 4–5 occasions between June 2007 and August 2008 to monitor temporal changes in host and parasite populations. Fish densities at each site were assessed by single-pass electric fishing of the same reach (230–416 m2) on each occasion. Random subsamples of roundhead galaxias (n = 5–27) and brown trout (n = 1–16) from each site were euthanized, measured for fork length (FL) to the nearest 1 mm and preserved in 10% buffered formalin. In the laboratory, the alimentary canal from oesophagus to anus of each fish was removed and split longitudinally. The abundance of acanthocephalans was noted, and the length (lm) and sex of each worm were recorded. In addition, for female worms, developmental status (F1 – immature with ovarian balls only, F2 – maturing eggs, F3 – fully mature eggs) and proportion of mature eggs in a subsample of 50 eggs were also recorded. The number of P. fluviatilis and prey fish  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 992 R. A. Paterson et al. present in fish stomachs was also enumerated, to assess the relative importance of cystacanth and postcyclic infection pathways. The prevalence of infection in the intermediate amphipod host P. fluviatilis was measured on each site visit from a 3–5-min kick sample spread diagonally across the stream to sample all microhabitats. Amphipods were preserved in 10% formalin and up to 1000 amphipods from each sample were examined for the presence of cystacanths. INFECTION EXPERIMENT An infection experiment was conducted to determine the fitness of A. galaxii worms in both definitive hosts, from which relative rates of parasite establishment, maturation and mortality could be estimated. Brown trout (131–191 mm) and roundhead galaxias (60–75 mm) were collected by electric fishing from the Cap Burn (2295¢56¢¢E, 5546¢22¢¢N) and a tributary of the Kye Burn (2293¢23¢¢E, 5563¢67¢¢N), and acclimatized in the University of Otago’s controlled climate facilities (13-h day ⁄ 11-h night period, 10C, 15% daily water change, fed ad libitum with commercial pellets) for 6 months prior to experimentation. All experimental fish were treated with the anthelmintic Tetramisole HCl (10 mg L)1 48 h · 2 repeat) at the beginning of the acclimatization period to remove any naturally acquired parasites prior to artificial infection. Infection status prior to the experiment was investigated by autopsy of five-treated and five-untreated fish per species, during which no intestinal parasites were found in treated or untreated fish. Six experimental fish of each species were randomly assigned to two infection time-periods (2 or 6 weeks). During the 48 h prior to experimental infection, each fish was measured (FL) and placed without food into an individual tank (tank size: 15 L – roundhead galaxias, 30 L – brown trout). Intermediate infective stages of A. galaxii were obtained by collecting naturally infected P. fluviatilis amphipods by handnet from the Swin Burn (2298¢30¢¢E, 5559¢40¢¢N). Infection status was determined under a dissection microscope. A subsample of 15 amphipods identified as infected was autopsied to assess observer error in correctly identifying the presence of A. galaxii, with single infections of A. galaxii correctly identified in all amphipods. Ten live infected A. galaxii amphipods were placed into each tank. The number of amphipods consumed by each fish was monitored by checking for uneaten individuals 24- and 48-h postexposure. All unconsumed amphipods were removed at 48-h postexposure, and fish resumed their diet of ad libitum commercial fish pellets. At the completion of each infection period (2 or 6 weeks), fish were euthanized and immediately examined for parasites prior to fixing all acanthocephalans in 10% buffered formalin. Worm measurements were made as in the field survey. STATISTICAL ANALYSES Statistical analyses of field survey and experimental infection data were conducted using spss Statistic 15.0 (SPSS 2006). Parametric tests were used for all analyses, for which data were transformed when necessary to meet the normality assumptions of analysis of variance (anova). POPULATION MODEL A population model was constructed using model maker 4.0 (Cherwell 2000) to dynamically simulate the influence of brown trout presence on A. galaxii in its definitive hosts. Model formulation The life cycle of A. galaxii was modelled using a series of linked differential equations that simulate (on a daily basis, scaled to 1 m2) the populations of uninfected (ANP) and infected (AP) amphipods, the populations of immature and mature A. galaxii worms in roundhead galaxias (IG, WG) and trout (IT, WT), and the number of A. galaxii eggs in the environment (E), in discrete time steps: dANP ¼ ðkA S þ rexðANP þAP Þ  1  CT T  CG G  vEÞANP dt eqn 1a   dAP ¼ vEANP þ ðrexðANP þAP Þ  1ÞX þ ðCT T þ CG GÞY AP dt eqn 1b dIG ¼ CG GYc AP  ðmG þ aG þ bG þ DTÞIG dt eqn 2a dIT ¼ cTðCT YAP þ DIG Þ  ðmT þ aT þ bT ÞIT dt eqn 2b dWG ¼ mG IG  ðaG þ bG þ DTÞWG dt eqn 3a dWT ¼ mT IT þ cDTWG  ðaT þ bT ÞWT dt eqn 3b dE ¼ ðWG þ WT ÞkW  ðe þ vANP ÞE dt eqn 4 The amphipod population was dynamically modelled to reflect changing amphipod densities throughout the year while fish densities were held constant (T and G for trout and galaxias, respectively) because field surveys indicate the latter remain relatively constant throughout the year, and there is no evidence of parasite impact on fitness (Kelly et al. 2009a). The population of uninfected amphipods ANP (eqn 1a) increases as a function of amphipod fecundity (kA) (toggled by S, a logic switch denoting the breeding season) and decreases as a function of natural mortality (r ) 1, where r is survival) influenced by density dependence (x), predation by trout (CT) and galaxias (CG), and infection (m). The population of infected amphipods AP (eqn 1b) increases as a function of infection (v), and is modified by a combination of natural mortality (r ) 1) modified by density dependence (x) and infection (X), and the predation of infected amphipods by trout (CT) and galaxias (CG), also modified by infection (Y). The parameter Y acknowledges that the presence of the parasite can be expected to increase the likelihood that an amphipod will be eaten. If the rate at which uninfected amphipods ingest acanthocephalan eggs exceeds the number of uninfected amphipods, all amphipods are considered to be infected. The population of immature worms in galaxias IG (eqn 2a) increases with the predation of infected amphipods (CGY), modified by establishment success (c). The number of immature worms decreases with worm maturation (mG), parasite mortality (aG),  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 Trout alter native host–parasite dynamics 993 natural host mortality (bG) and postcyclic transmission of worms from galaxias to trout (D). The population of immature worms in trout IT (eqn 2b) increases with the predation of infected amphipods (CTY) and postcyclic transmission (D), both modified by establishment success (c), and decreases with worm maturation (mT), worm mortality (aT) and natural host mortality (bT). The population of mature worms in galaxias WG (eqn 3a) increases as a function of worm maturation (mG) and decreases with parasite mortality (aG), natural host mortality (bG) and postcyclic transmission (D). The population of mature worms in trout WT (eqn 3b) increases as a function of worm maturation (mT), the establishment (c) of postcyclically transmitted worms from galaxias to trout (D), and decreases because of the effects of parasite mortality (aT) and natural host mortality (bT). A density dependence parameter influencing worm survival was not incorporated into immature or mature worm equations, as very low field infection intensities of A. galaxii were observed in the species modelled in contrast to burdens recorded from other galaxias hosts in New Zealand (cf. Galaxias maculatus mean infection intensity 127 per fish; Hine 1977). The number of A. galaxii eggs in the environment E (eqn 4) increases as a function of the number of eggs produced (kW) from mature worms in both galaxias (WG) and trout (WT), and decreases with natural mortality (e) and the ingestion of eggs by amphipods (v). Model parameterization Parameter estimates were obtained, where possible, from the field surveys and experimental infections reported here, and from the published literature and unpublished data (see Tables 1 and 2). Host densities (G, T) were calculated as the average number of fish per m2 at Swin Burn Q1 and Q2 combined (Old Hut Creek was not included in the parameterization of the model as only a single A. galaxii infection was recorded there). The daily mortality rate of brown trout was based on a maximum life span of 5–6 years for nonmigratory brown trout (Huryn 1996), whereas the mortality rate of roundhead galaxias was estimated from a closely related species, Galaxias paucispondylus (Bonnett 1990). Mean amphipod density was estimated as 250 amphipods per m2 (Weller 2003), with natural survival estimated from the maximum life span of 1 year (F. Wilhelm, unpublished). The daily fecundity rate of 0Æ058 new individuals per amphipod during the breeding season (September to March, austral spring to end of austral summer) was estimated from a maximum of 3Æ5 broods of five eggs per female and a 0Æ5 sex ratio (F. Wilhelm, unpublished). However, as a proportion of females are known to breed year round (Towns 1981), breeding from April to August (austral autumn and winter) was set at 20% of the maximum. The probability of amphipods being consumed by each definitive host was calculated as the average number of amphipods in the host gut (with an assumed gut clearance time of 48 h; Tekinay, Guner & Davies 2003) relative to amphipod density (field survey). The potential for postcyclic transmission of parasites from galaxias to trout was calculated from the average number of roundhead galaxias present in the stomachs of brown trout in the field surveys. Although acanthocephalans have been shown to markedly alter intermediate host behaviour (Baldauf et al. 2007), often increasing the likelihood that infected individuals will be eaten (Moore 1984; Lagrue et al. 2007), no information is available about how A. galaxii alters the risk of being eaten (Y) or mortality (X) of its amphipod host. Likewise, no information is available about amphipod density dependence (x) or parasite egg ingestion rate by amphipods (m). These parameters were estimated by model optimization to the values Table 1. Definitive host parameter definitions and estimates used in the population model Galaxias Trout Parameter Symbol Value Symbol Value Units Source Density Predation of amphipods G CG 0Æ0763 0Æ00047 T CT 0Æ0117 2e)05 This study This study Parasite establishment Parasite mortality Parasite maturation Natural host mortality c aG mG bG 0Æ5078 0Æ01869 0Æ006 0Æ0007 c aT mT bT 0Æ5078 0Æ0416 0Æ000 0Æ0004 m)2 amphipod per host per day per worm per worm per day per worm per day per host per day Predation of galaxias – – D 0Æ01 per galaxias per trout per day This study This study This study Bonnett (1990), Huryn (1996) This study Table 2. Intermediate host and parasite parameter definitions and estimates used in the population model Parameter Symbol Value Units Source Amphipod breeding rate Amphipod natural mortality Parasite fecundity Ingestion rate of eggs by amphipods Egg mortality Amphipod density dependence Parasite influence on amphipod mortality Parasite influence on amphipod predation kA r kW M E x X Y 0Æ058 0Æ0027 807Æ5 6e)06 0Æ005 0Æ0001 26Æ68 40Æ36 per amphipod per day per amphipod per day per worm per day per egg per amphipod per day per egg per day per amphipod per day per amphipod per day per amphipod per day F. Wilhelm, unpublished data F. Wilhelm, unpublished data Crompton & Whitfield (1968) This study Crompton (1970) Model optimization Model optimization Model optimization  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 994 R. A. Paterson et al. that result in predicted levels of prevalence and infection intensity, in both intermediate and definitive hosts, equivalent to those observed in the field (a mean population size of 250 m)2 for uninfected amphipods, a mean prevalence of infected amphipods of 0Æ1%, and a mean worm population size of 0Æ1 m)2 at equilibrium dynamics). Once an amphipod becomes infected with a single A. galaxii, we assumed it does not consume additional eggs. This assumption does not affect model dynamics as eggs are highly abundant in the environment, and prevalence of amphipod infection is low. The establishment, maturation and mortality rates of A. galaxii in each definitive host were estimated from the infection experiment. Parasite establishment at day one and the daily mortality rate of A. galaxii in galaxias were estimated from the slope of the line calculated from the percentage of worms present at 2 and 6 weeks postinfection. No worms were present in trout at 6 weeks postinfection, so parasite establishment rate in trout was assumed to be the same as in galaxias, while parasite mortality rate was estimated from the decline to zero parasites at 6 weeks. No mature worms were present in trout from either the experimental infection or field survey; thus, the maturity rate in trout was set at 0. The daily rate of parasite egg production was estimated from the acanthocephalan, Polymorphus minutus (1700 eggs per female per day; (Crompton & Whitfield 1968), divided by the overall parasite male:female ratio observed in the current study. Egg mortality was based on aquatic acanthocephalan eggs being known to survive a maximum of 6–9 months (Crompton 1970). Model simulation Although brown trout are known to have altered freshwater fish communities in New Zealand, resulting in the declining abundance and localized extinction of some Galaxias species (Townsend 2003), their influence on native host–parasite dynamics is largely unknown. Here, we use model simulations to examine the influence of brown trout on native parasite dynamics in roundhead galaxias by investigating varying ratios of fish community composition that represent (i) fish communities prior to trout introduction (galaxias scenario), (ii) trout presence in native communities (galaxias + trout scenario) and (iii) localized extinction of native fish (trout scenario). To understand how the presence of the exotic host might be influencing observed levels of infection in the native host, we use sensitivity analyses to test the validity of potential contrasting hypotheses in the galaxias + trout scenario: (i) brown trout out-compete roundhead galaxias for infected amphipods (varying CT), (ii) brown trout prey on roundhead galaxias (varying D) and (iii) brown trout modify the behaviour of roundhead galaxias, resulting in altered consumption of infected amphipods (varying CG). Simulations were run varying the above parameters ± 100% around their estimated value in 10% increments, to investigate the influence of each parameter on the infection intensity of A. galaxii in roundhead galaxias. For each scenario, we monitored peak density of A. galaxii in the roundhead galaxias population (m)2) after 3000 daily iterations. Results FIELD SURVEY Roundhead galaxias was the most abundant definitive host in surveyed streams, with average densities of 0Æ05–0Æ20 fish per m2 for roundhead galaxias and 0Æ005–0Æ07 fish per m2 for brown trout. A total of 219 of 537 roundhead galaxias (39– 91 mm) and 84 of 126 brown trout (63–189 mm) were examined for A. galaxii. At Old Hut Creek, 72 galaxias and 34 trout were examined, but only a single A. galaxii infection was recovered from one roundhead galaxias on a single occasion (in June 2007). This site was removed from further analysis of field survey observations. Infection prevalence did not differ significantly between sites or host species (GLM; P > 0Æ05), with the percentage of fish infected ranging from 12Æ5% to 85% for roundhead galaxias and 0–100% for brown trout (Table 3). Infection intensities were highest in brown trout (natural log transformed: GLM species; F1,93 = 4Æ977, P = 0Æ028) but did not Table 3. Seasonal prevalence, infection intensity and length Acanthocephalus galaxii in naturally infected roundhead galaxias and brown trout from sites Q1 and Q2 on the Swin Burn Site Q1 Host Species Roundhead galaxias Brown trout Q2 Roundhead galaxias Brown trout Season No. of hosts Jun 07 Oct 07 Mar 08 May 08 Aug 08 Jun 07 Oct 07 Mar 08 May 08 Aug 08 Oct 07 Mar 08 May 08 Aug 08 Oct 07 Mar 08 May 08 Aug 08 10 27 16 20 19 0 6 3 1 0 20 10 20 5 14 4 2 2 Prevalence % Infection intensity (mean ± SD) Male length lm (mean ± SD) Female length lm (mean ± SD) No. of parasites 40 40Æ7 12Æ5 45 52Æ6 – 83Æ3 100 0 – 80 80 85 60 85Æ7 75 50 100 2Æ8 ± 6Æ6 ± 5 16Æ6 ± 11Æ2 ± – 12Æ6 ± 13Æ7 ± – – 11Æ2 ± 1Æ5 ± 4Æ1 ± 5Æ7 ± 10Æ8 ± 6Æ3 ± 27 10 1725 ± 2376 ± 2600 ± 1699 ± 2057 ± – 1210 ± 1039 ± – – 1851 ± 2346 ± 2108 ± 1642 ± 943 ± 1378 ± 1254 ± 1478 ± 4800 ± 3002 ± 4350 ± 2447 ± 3403 ± – 1141 ± 1162 ± – – 2520 ± 2720 ± 3349 ± 3255 ± 967 ± 1295 ± 1405 ± 1475 ± 11 72 10 149 112 – 63 41 – – 179 12 65 17 129 19 27 20 1Æ5 6Æ3 19Æ7 16 14Æ6 16Æ8 22Æ4 0Æ5 3Æ8 4Æ7 12Æ1 3Æ8 527 1023 650 563 393 323 239 713 940 716 391 193 370 192 311 566 1844 1222 963 1221 266 232 1374 952 1665 900 180 341 208 293  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 Trout alter native host–parasite dynamics 995 Q1 Q2 Worms in galaxias 100% 80% 60% 40% 20% 0% 80% 60% 40% 20% 0% Jun 07 Oct 07 differ between sites or seasons. Acanthocephalans were larger in roundhead galaxias than brown trout (GLM species; F1,220 = 24Æ803, P < 0Æ001), with female worms larger than male worms (GLM sex; F1,220 = 5Æ965, P = 0Æ015). All female worms recovered from brown trout were immature with ovarian balls present only, while on average 18Æ6% of female worms in roundhead galaxias contained developing eggs (F2) or fully mature shelled eggs (F3) (Fig. 1). Examination of stomach contents showed that 71% of brown trout stomachs contained amphipods, with an average of 14Æ9 amphipods per fish, while 17% of roundhead galaxias stomachs contained amphipods, with an average of two amphipods per fish. Fourteen per cent of brown trout stomachs also contained roundhead galaxias, with between one and three galaxias present. Prevalence of infection in amphipods collected from kick samples was very low at all sites, with a maximum of 0Æ1% of amphipods infected. INFECTION EXPERIMENT Similar percentages of worms had established in each host species at 2 weeks postinfection, with an average of 38Æ6% of worms establishing in roundhead galaxias and 28Æ0% in brown trout (one-way anova species; P = 0Æ465; Fig. 2). The intensity of infection declined in both species between 2 and 6 weeks postinfection, with 19Æ2% of worms remaining in galaxias and no worms in trout (Fig. 2). Male A. galaxii were larger in brown trout than in roundhead galaxias at 2 weeks postinfection (one-way anova male; F1,14 = 5Æ948, P = 0Æ030), while female worms were of similar length (P = 0Æ136, Table 4). Worm size increased between 2 and 6 weeks postinfection, with average male and female lengths in galaxias at 6 weeks postinfection greater than in trout at 2 weeks postinfection. No gravid female Mar 08 May 08 Aug 08 Jun 07 Oct 07 Time Mar 08 May 08 Aug 08 Time Male Non-gravid female Gravid female 60 Trout Establishment % Fig. 1. Seasonal composition of Acanthocephalus galaxii sex and reproductive status in roundhead galaxias and brown trout at sites Q1 and Q2 on the Swin Burn. Gravid female worms with maturing eggs (F2) or fully mature shelled eggs (F3) were absent from brown trout throughout the study period. Worms in trout 100% Galaxias 40 20 0 2 6 Week Fig. 2. Establishment (2 weeks) and survival (6 weeks) of Acanthocephalus galaxii in experimentally infected roundhead galaxias and brown trout. Establishment success did not differ between hosts (P > 0Æ05). worms were present in either host species at 2 weeks postinfection, while one of four female worms present in galaxias at 6 weeks postinfection was gravid. POPULATION MODEL Model simulations of varying fish community composition indicated that a minimum proportion of 70% galaxias in the total fish community is the threshold for the survival of A. galaxii, below which the native parasite rapidly goes extinct (Fig. 3). At the current field densities of the total fish community (0Æ088 individuals per m2), densities of A. galaxii infection in galaxias are predicted to be only slightly lower than would be expected if trout were absent from the system, as the ability of trout to act as an infection sink is weakened by the low relative density of this introduced host. Changing the rate of predation by brown trout on either amphipods or roundhead galaxias had little influence on the  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 996 R. A. Paterson et al. Table 4. Prevalence, infection intensity, length and reproductive status Acanthocephalus galaxii in experimentally infected roundhead galaxias and brown trout. Gravid females represent females with maturing eggs (F2) and fully mature shelled eggs (F3) Host Week No. of hosts Prevalence % Infection intensity (mean ± SD) Roundhead galaxias 2 6 100 3Æ3 ± 2Æ5 6 6 50 2Æ3 ± 1Æ5 2 6 100 2Æ6 ± 1Æ9 6 6 0 0 Length lm (mean ± SD) No. of parasites M F1 M:F ratio M F1 F2 M:F ratio M F1 M:F ratio – 1806 2097 1:1Æ6 2191 2783 2425 1:1Æ3 2128 2338 1:1Æ3 – ± 84 ± 108 8 12 0 ± 141 ± 65 3 3 1 25 ± 103 ± 103 7 9 0 – – Gravid females % (a) 0·4 0·2 0·3 0·2 0·1 0·1 0·0 0 0 20 40 60 80 100 % Galaxias Fig. 3. Density of Acanthocephalus galaxii infections in roundhead galaxias (filled circle) and brown trout (open circle) in relation to percentage of galaxias in the combined total fish population (0Æ088 fish per m2) predicted from model simulations. The dashed line represents the current percentage of galaxias in the total fish population as recorded in the field survey. Acanthocephalus galaxii density m–2 Acanthocephalus galaxii density m–2 Brown trout Parasite sex (b) 0·4 0·3 0·2 0·1 0·0 (c) 0·4 density of A. galaxii in the modelled roundhead galaxias population (Fig. 4a,c). In contrast, A. galaxii densities in roundhead galaxias were highly sensitive to changes in the rate of predation on amphipods by this fish (Fig. 4b), with predation rates below 0Æ004 per amphipod per host per day ()20%) resulting in extinction of the parasite. 0·3 0·2 0·1 0·0 –100 –50 0 50 100 % Change Discussion We used a combination of field observations, experimental infections and dynamic population modelling to examine the influence of an introduced species on a native host–parasite system. Our field observations indicated that although the prevalence and infection intensity of A. galaxii were similar in introduced trout and native galaxias, the native parasite attained smaller sizes and failed to mature in the introduced host. Our experimental infection results showed trout to be similar to galaxias in its competence as a host in terms of parasite establishment and initial growth, but trout was found to be unsuitable for maintaining A. galaxii populations, as Fig. 4. Simulating the effect on the density of Acanthocephalus galaxii infections in roundhead galaxias of altering (a) trout predation on amphipods CT, (b) galaxias predation on amphipods CG and (c) trout predation on galaxias (D). Each estimate was modified ± 100% of current estimates in 10% increments. worms did not reach maturity. Hence, the field and experimental results suggested that trout were likely to act as an infection sink because of poor host competency. However, the population model indicated that the influence of this mechanism on the dynamics of A. galaxii in its native host  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 Trout alter native host–parasite dynamics 997 would be limited because of only a small proportion of the total fish community being comprised of trout. In contrast, our model suggested that trout presence may have more substantial consequences for the native parasite population by altering amphipod consumption by roundhead galaxias, with such consequences being greater at higher trout densities. FIELD OBSERVATIONS Field observations are often the first source of information used to determine the influence of introduced species on host ⁄ parasite dynamics. For example, Telfer et al. (2005) used field surveys to determine the influence of the introduced bank vole (Myodes glareolus) on Bartonella sp. infections in native wood mice (Apodemus sylvaticus), concluding that infection prevalence in wood mice decreased with increasing bank vole density. Moreover, Rauque, Viozzi & Semenas (2003) used the results of field surveys to demonstrate that parasites in the introduced rainbow trout (Oncorhynchus mykiss) contributed 20% of total egg output by the native acanthocephalan parasite Acanthocephalus tumescens. Field observations may provide useful insights into the potential impacts of introduced species on native parasite–host dynamics. However, as demonstrated by our study, assessment of parasite maturation is needed in addition to observations of infection prevalence and intensity to determine the impact of an introduced species. INFECTION EXPERIMENT Our experimental infections generally supported the fieldbased conclusions that trout are inferior hosts of A. galaxii, yet interestingly A. galaxii initially experienced similar establishment success and equivalent or greater growth in this host. Although worms in galaxias were relatively slower growing, in the long term they attained larger sizes than in trout. Most importantly, A. galaxii worms failed to reach maturity in brown trout; thus, the introduced host was unsuccessful at supporting this stage of the parasite’s life cycle. POPULATION MODEL The population model revealed that the density of A. galaxii in roundhead galaxias may only be slightly reduced as a result of parasite dilution by trout. In sensitivity analyses, modifying either the rate at which trout feed on roundhead galaxias or on amphipods caused little change to the native parasite’s density in its native roundhead galaxias host. However, decreasing the rate of amphipod predation by roundhead galaxias markedly decreased the parasite’s adult population. The model also revealed that parasite burdens were highly sensitive to relative host abundances. Hence, a correlation between increasing trout presence and decreasing native parasitism of roundhead galaxias, as observed for the parasite Deretrema philippae by Kelly et al. (2009a), may just as likely be caused by trout reducing galaxias density or altering their foraging behaviour, rather than trout serving as an infection sink. This conclusion is supported both by demonstrations that trout predate galaxias (Crowl, Townsend & McIntosh 1992; Townsend 2003) and experimental studies of habitat use by native galaxiids in the presence of brown trout (McIntosh, Townsend & Crowl 1992), which showed that trout out-compete Galaxias spp. in streams, forcing them to forage in areas with lower invertebrate drift. Native parasite loss through dilution or via other mechanisms, such as altered predation on intermediate hosts by native species, is a largely unrecognized consequence of the invasion of introduced species. Most commonly, invasion biology research has focused on the impacts of introduced species on native conspecifics via predation and ⁄ or competition and through the introduction of novel parasites. However, as native parasites play major roles in the structuring of native communities, neglecting the potential for native parasite loss subsequent to the arrival of exotic species overlooks a potentially important consequence of invasions. LIMITATIONS AND STRENGTHS OF DYNAMIC POPULATION MODELS We acknowledge that the strength of population simulation models in depicting ‘real-world’ dynamics is only as good as the data used to parameterize them. In the present study, we utilized field information from multiple sites and seasons, experimental infections and published literature to generate the majority of parameter estimates. However, we were unable to draw on these sources to estimate the amphipod’s mortality and predation risk, although the dynamic model suggested that A. galaxii infections are strongly influenced by both parameters. Manipulation of intermediate host behaviour to enhance transmission success via predation is a common strategy of parasites that utilize hosts at different trophic levels and is frequently adopted by acanthocephalans (Moore 1984; Poulin 1995). Acanthocephalan infections in amphipods have also been reported to alter host immune responses to bacterial infections, potentially increasing intermediate host mortality (Cornet et al. 2009). Experimental studies are required to fully quantify the mechanism and degree to which A. galaxii infection alters the mortality and predation risk of amphipods, and to identify whether behavioural responses of infected amphipods to predator presence differ between native and exotic fish. Acknowledgements We thank D. Kelly, M. Green and A. Lal for field and laboratory assistance, and F. Wilhelm for P. fluviatilis life-history information. We are also grateful to R. Scott and D. Crutchley for site access. Permission was obtained from the Department of Conservation, NZ Fish and Game and the University of Otago Animal Ethics Committee (ET 16 ⁄ 7). Funding for this research was provided by the Royal Society of New Zealand Marsden Fund. References Baldauf, S.A., Thunken, T., Frommen, J.G., Bakker, T.C.M., Heupel, O. & Kullmann, H. (2007) Infection with an acanthocephalan manipulates an  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998 998 R. A. Paterson et al. amphipod’s reaction to a fish predator’s odours. International Journal for Parasitology, 37, 61–65. Bonnett, M.L. (1990) Age and growth of alpine galaxias (Galaxias paucispondylus Stokell) and longjawed galaxias (Galaxias prognathus Stokell) in the Rangitata River, New Zealand. New Zealand Journal of Marine and Freshwater Research, 24, 151–158. Bush, A.O., Lafferty, K.D., Lotz, J.M. & Shostak, A.W. (1997) Parasitology meets ecology on its own terms: Margolis et al. revisited. Journal of Parasitology, 83, 575–583. Cherwell (2000) ModelMaker User Manual. Cherwell Scientific Limited, Oxford. Cornet, S., Franceschi, N., Bauer, A., Rigaud, T. & Moret, Y. (2009) Immune depression induced by acanthocephalan parasites in their intermediate crustacean host: consequences for the risk of super-infection and links with host behavioural manipulation. International Journal for Parasitology, 39, 221– 229. Crompton, D.W. (1970) An ecological Approach to Acanthocephalan Physiology. University Press, Cambridge. Crompton, D.W. & Whitfield, P.J. (1968) Course of infection and egg production of Polymorphus minutus (Acanthocephala) in domestic ducks. Parasitology, 58, 231–246. Crowl, T.A., Townsend, C.R. & McIntosh, A.R. (1992) The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Reviews in Fish Biology and Fisheries, 2, 217–241. Daszak, P., Cunningham, A.A. & Hyatt, A.N. (2000) Emerging infectious diseases of wildlife: threats to biodiversity and human health. Science, 287, 443– 449. Dubois, N., Marcogliese, D.J. & Magnan, P. (1996) Effects of the introduction of white sucker, Catostomus commersoni, on the parasite fauna of brook trout, Salvelinus fontinalis. Canadian Journal of Zoology – Revue Canadienne De Zoologie, 74, 1304–1312. Galli, P., Stefani, F., Benzoni, F. & Zullini, A. (2005) Introduction of alien host-parasite complexes in a natural environment and the symbiota concept. Hydrobiologia, 548, 293–299. Hine, P.M. (1977) Acanthocephalus galaxii n.sp. parasitic in Galaxias maculatus (Jenyns, 1842) in Waimeha Stream, New-Zealand. Journal of the Royal Society of New Zealand, 7, 51–57. Hine, P.M., Jones, J.B. & Diggles, B.K. (2000) A checklist of the parasites of New Zealand fishes, including previously unpublished records. National Institute of Water and Atmospheric Research, Wellington. Huryn, A.D. (1996) An appraisal of the Allen paradox in a New Zealand trout stream. Limnology and Oceanography, 41, 243–252. Kelly, D.W., Paterson, R.A., Townsend, C.R., Poulin, R. & Tompkins, D.M. (2009a) Has the introduction of brown trout altered disease patterns in native New Zealand fish? Freshwater Biology, 54, 1805–1818. Kelly, D.W., Paterson, R.A., Townsend, C.R., Poulin, R. & Tompkins, D.M. (2009b) Parasite spillback: a neglected concept in invasion ecology? Ecology, 90, 2047–2056. King, T.A., van Oosterhout, C. & Cable, J. (2009) Experimental infections with the tropical monogenean, Gyrodactylus bullatarudis: potential invader or experimental fluke? Parasitology International, 58, 249–254. Kopp, K. & Jokela, J. (2007) Resistant invaders can convey benefits to native species. Oikos, 116, 295–301. Lagrue, C., Kaldonski, N., Perrot-Minnot, M.J., Motreuil, B. & Bollache, L. (2007) Modification of hosts’ behavior by a parasite: field evidence for adaptive manipulation. Ecology, 88, 2839–2847. MacCrimmon, H. & Marshall, T.L. (1968) World distribution of brown trout Salmo trutta. Journal of the Fisheries Research Board of Canada, 25, 2527– 2548. McIntosh, A.R., Townsend, C.R. & Crowl, T.A. (1992) Competition for space between introduced brown trout (Salmo trutta L.) and a native galaxiid (Galaxias vulgaris Stokell) in a New Zealand stream. Journal of Fish Biology, 41, 63–81. Moore, J. (1984) Altered behavioral responses in intermediate hosts – an acanthocephalan parasite strategy. American Naturalist, 123, 572–577. Poulin, R. (1995) ‘‘Adaptive’’ changes in the behaviour of parasitized animals: a critical review. International Journal for Parasitology, 25, 1371–1383. Rauque, C.A., Semenas, L.G. & Viozzi, G.P. (2002) Post-cyclic transmission in Acanthocephalus tumescens (Acanthocephala: Echinorhynchidae). Folia Parasitologica, 49, 127–130. Rauque, C.A., Viozzi, G.P. & Semenas, L.G. (2003) Component population study of Acanthocephalus tumescens (Acanthocephala) in fishes from Lake Moreno, Argentina. Folia Parasitologica, 50, 72–78. SPSS (2006) SPSS for Windows Version 15.0.0. SPSS, Inc., Chicago. Tekinay, A.A., Guner, Y. & Davies, S.J. (2003) Influence of dietary energy level on stomach emptying and appetite revival rates in rainbow trout, Oncorhynchus mykiss. Turkish Journal of Veterinary and Animal Sciences, 27, 1077– 1084. Telfer, S., Bown, K.J., Sekules, R., Begon, M., Hayden, T. & Birtles, R. (2005) Disruption of a host-parasite system following the introduction of an exotic host species. Parasitology, 130, 661–668. Thieltges, D.W., Reise, K., Prinz, K. & Jensen, K.T. (2009) Invaders interfere with native parasite-host interactions. Biological Invasions, 11, 1421–1429. Tompkins, D.M. & Poulin, R. (2006) Parasites and biological invasions. Biological Invasions in New Zealand (eds R.B. Allen & W.G. Lee). pp. 67–86, Springer, Berlin, Heidelberg. Torchin, M.E., Lafferty, K.D., Dobson, A.P., McKenzie, V.J. & Kuris, A.M. (2003) Introduced species and their missing parasites. Nature, 421, 628–630. Towns, D.R. (1981) Life histories of benthic invertebrates in a kauri forest stream in northern New Zealand. Australian Journal of Marine and Freshwater Research, 32, 191–211. Townsend, C.R. (1996) Invasion biology and ecological impacts of brown trout Salmo trutta in New Zealand. Biological Conservation, 78, 13–22. Townsend, C.R. (2003) Individual, population, community, and ecosystem consequences of a fish invader in New Zealand streams. Conservation Biology, 17, 38–47. Weller, F. (2003) Effects of increased fine sediment inputs due to different land use regimes on stream invertebrates. MSc thesis. University of Munich. Received 31 May 2010; accepted 22 February 2011 Handling Editor: Mike Boots  2011 The Authors. Journal of Animal Ecology  2011 British Ecological Society, Journal of Animal Ecology, 80, 990–998