Scholarly article on topic ' Feeding behaviour, predatory functional responses and trophic interactions of the invasive Chinese mitten crab ( Eriocheir sinensis ) and signal crayfish ( Pacifastacus leniusculus ) '

Feeding behaviour, predatory functional responses and trophic interactions of the invasive Chinese mitten crab ( Eriocheir sinensis ) and signal crayfish ( Pacifastacus leniusculus ) Academic research paper on "Biological sciences"

0
0
Share paper
Academic journal
Freshwater Biology
OECD Field of science
Keywords
{""}

Academic research paper on topic " Feeding behaviour, predatory functional responses and trophic interactions of the invasive Chinese mitten crab ( Eriocheir sinensis ) and signal crayfish ( Pacifastacus leniusculus ) "

Freshwater Biology (2016) 61, 426-443 doi:10.1111/fwb,12717

Feeding behaviour, predatory functional responses and trophic interactions of the invasive Chinese mitten crab (Eriocheir sinensis) and signal crayfish (Pacifastacus leniusculus)

PAULA J ROSEWARNE*, ROBERT J.G. MORTIMER1 -1, ROBERT J. NEWTON1, CHRISTOPHER GROCOCK*, CHRISTOPHER D. WING* AND ALISON M. DUNN*

*School of Biology, Faculty of Biological Sciences, University of Leeds, Leeds, U.K. ^School of Earth and Environment, University of Leeds, Leeds, U.K.

1School of Animal, Rural and Environmental Sciences, Nottingham Trent University, Southwell, Nottinghamshire, U.K. SUMMARY

1. Freshwaters are subject to particularly high rates of species introductions; hence, invaders increasingly co-occur and may interact to enhance impacts on ecosystem structure and function. As trophic interactions are a key mechanism by which invaders influence communities, we used a combination of approaches to investigate the feeding preferences and community impacts of two globally invasive large benthic decapods that co-occur in freshwaters: the signal crayfish (Pacifastacus leniusculus) and Chinese mitten crab (Eriocheir sinensis).

2. In laboratory preference tests, both consumed similar food items, including chironomids, isopods and the eggs of two coarse fish species. In a comparison of predatory functional responses with a native crayfish (Austropotamobius pallipes), juvenile E. sinensis had a greater predatory intensity than the native A. pallipes on the keystone shredder Gammarus pulex, and also displayed a greater preference than P. leniusculus for this prey item.

3. In outdoor mesocosms (n = 16) used to investigate community impacts, the abundance of amphipods, isopods, chironomids and gastropods declined in the presence of decapods, and a decapod >gastropod >periphyton trophic cascade was detected when both species were present. Eriocheir sinensis affected a wider range of animal taxa than P. leniusculus.

4. Stable-isotope and gut-content analysis of wild-caught adult specimens of both invaders revealed a wide and overlapping range of diet items including macrophytes, algae, terrestrial detritus, macroinvertebrates and fish. Both decapods were similarly enriched in 15N and occupied the same trophic level as Ephemeroptera, Odonata and Notonecta. Eriocheir sinensis 813C values were closely aligned with macrophytes indicating a reliance on energy from this basal resource, supported by evidence of direct consumption from gut contents. Pacifastacus leniusculus 813C values were intermediate between those of terrestrial leaf litter and macrophytes, suggesting reliance on both allochthonous and autochthonous energy pathways.

5. Our results suggest that E. sinensis is likely to exert a greater per capita impact on the macroinvertebrate communities in invaded systems than P. leniusculus, with potential indirect effects on productivity and energy flow through the community.

Keywords: invasive species, mesocosm, prey choice, stable isotopes, Type II functional response

Correspondence: Alison M. Dunn, School of Biology, Faculty of Biological Sciences, University of Leeds, Leeds LS2 9JT, U.K. E-mail: a.dunn@leeds.ac.uk

426 © 2016 The Authors Freshwater Biology Published by John Wiley & Sons Ltd.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Introduction

Freshwaters are particularly vulnerable to humanmediated introduction of invasive species due to their physical connectivity and high levels of human disturbance (Dudgeon et al., 2006), hence biological invasions are currently one of the most widespread and deleterious pressures on freshwater ecosystems (Leprieur et al., 2009; Ricciardi & Macisaac, 2011). Understanding and quantifying the impacts invasive species have on the communities and ecosystems they invade is crucial for effectively targeting the limited resources available for management and control (Parker et al., 1999; Keller et al., 2011). As the mechanisms by which invaders influence ecosystems are frequently complex and vary greatly between species, assessing invader impacts is often difficult (Ruiz et al, 1999; Simberloff et al, 2013). Further, the accelerating rate of biological invasions means many ecosystems already support several sympatric invasive species (Simberloff & Von Holle, 1999; Jackson & Grey, 2013). The potentially complex interactions between invaders can mediate ecological outcomes, for example, by facilitating subsequent establishment of introduced species in an 'invasion meltdown' (Simberloff & Von Holle, 1999; Ricciardi, 2001), or modifying structural and functional impacts on the invaded community through additive or counteractive effects (Strayer, 2010; Preston, Henderson & Johnson, 2012; Jackson et al., 2014).

Trophic interactions are a key mechanism by which invaders influence communities. Invaders can reduce the abundances of native species by direct predation and, through a variety of mechanisms, may exert a stronger predation pressure than functionally equivalent native predators (Noonburg & Byers, 2005; Salo et al., 2007). Invaders may compete with native species for food resources and are often more successful competitors due to r-selected traits such as boldness, faster growth rate and higher fecundity (Williamson & Fitter, 1996; Karatayev et al., 2009). As well as frequently reaching higher densities than native predators (Parker et al., 2013), invaders may exert a stronger per capita effect on prey, if naive native prey exhibit less effective predator-avoidance strategies for the exotic predator (Diamond, Ashmole & Purves, 1989) or if invasive predators capture or handle prey more effectively (Bollache et al., 2008; Haddaway et al, 2012; Dick et al, 2013). Due to the often complex nature of freshwater food webs, with high connectance (Polis & Strong, 1996; Woodward et al., 2005), predation by invaders at one trophic level can cause cascading effects. Further, where invaders that

exploit similar food resources occur in sympatry, they may switch prey or broaden their resource base (Jackson & Britton, 2014; Rothhaupt, Hanselmann & Yohannes, 2014), resulting in increased predatory pressure on these alternative prey species. To disentangle the effects of ; multiple invaders in an ecosystem, it is necessary to quantify impacts (e.g. predation) for each species separately, but also in combination to detect potential niche shifts and behavioural plasticity when invasive predators occur in sympatry.

The invasive North American signal crayfish (Pacifas-tacus leniusculus: Astacidae) and Chinese mitten crab (Eriocheir sinensis: Varunidae) are both listed within the top 100 worst invaders (Lowe et al., 2000). Pacifastacus leniusculus has spread rapidly through Europe since its introduction for aquaculture, largely extirpating and replacing native crayfish populations through its role as a vector of Aphanomyces astaci, the cause of crayfish pla-; gue (Alderman, Holdich & Reeve, 1990). Eriocheir sinen-sis is native to eastern Asia, but has spread, mainly via i ships' ballast, to Europe and more recently to North America (Cohen & Carlton, 1997; Dittel & Epifanio, 2009). Unlike P. leniusculus, which completes its lifecycle entirely within freshwater, E. sinensis is catadromous, whereby reproduction and larval development occurs in estuarine waters, with the main growth phase (~3 years in Europe) in freshwater (Panning, 1939; Gilbey, Attrill & Coleman, 2008). Although their life histories are quite different, P. leniusculus and E. sinensis increasingly over; lap in freshwaters as they undergo range expansion. In the UK, advancement of E. sinensis inland has created i overlap zones with P. leniusculus, which presently occurs : in 83% of sub-catchments in England and Wales (Rogers & Watson, 2011). Further, the extent of overlap is likely to be far greater than recorded due to underreporting of E. sinensis occurrence (Mitten Crab Recording Project, 2013). Despite their expanding populations and increasing sympatry, no study has compared the feeding ecology of the two species, and for E. sinensis, we currently lack any quantitative data on feeding rates, both of which hinder reliable impact assessment (Ojaveer et al., 2007). Interspecies comparison is a valuable tool for forecasting potential impacts of established invaders for which very little ecological information is available (e.g. E. sinensis), or indeed, those recently introduced or at high risk of future introduction, when there exist similar invaders (with likely functional equivalence) for which there are better documented impacts, in this case P. le-niusculus (Dick et al., 2013, 2014). Given the likely complexity of their trophic interactions in freshwaters, we

employed a variety of approaches to elucidate potential impacts of P. leniusculus and E. sinensis (both in allopa-try and sympatry), including quantitative comparison of predatory functional response (the relationship between prey density and prey consumption by a predator) with a functionally equivalent native crayfish species.

Our study had three main aims. First, to compare the dietary preferences and feeding habits of P. leniusculus and E. sinensis using a combination of laboratory prey choice experiments, along with gut analyses and stable-isotope analysis of wild-caught specimens. Second, to quantify the predatory impact of P. leniusculus and E. sinensis on a key prey species and keystone shredder in the community using predatory functional responses (Holling, 1959; Bollache et al, 2008; Dick et al, 2014), determined through laboratory experiments, and compared with that of the native crayfish species Austropota-mobius pallipes (Astacidae). Third, to examine the effects of P. leniusculus and E. sinensis on freshwater communities using a field mesocosm experiment. Mesocosms provide a more realistic representation of the natural environment than laboratory experiments, but still with a level of control and replication difficult to obtain in the field. This scaling of approaches from laboratory and mesocosm manipulation to field observations was used with a view to reduce potential bias created by the inherent limitations of each, and thereby strengthen interpretation.

Methods

Study species

As an omnivorous keystone consumer and ecosystem engineer, P. leniusculus has the potential to modify communities through trophic interactions (Nystrom, Bron-mark & Graneli, 1996; Crawford, Yeomans & Adams, 2006) and physical changes, for example as bioturbators modifying sediment transport and increasing turbidity (Harvey et al, 2011; Johnson, Rice & Reid, 2011). There is concern that it preys on the eggs and tadpoles of amphibians (Axelsson et al., 1997), and on the emerging fry of commercially important fish (Edmonds, Riley & Maxwell, 2011). In invaded systems, P. leniusculus causes a reduction in the biomass and species richness of macrophyte and macroinvertebrate communities (Sten-roth & Nystrom, 2003; Crawford et al., 2006), with an accompanying shift towards predation resistant (e.g. sediment-dwelling) taxa (Nystroom, 1999), although recent work in boreal lakes, suggests community

impacts may be habitat specific in some contexts (Ruokonen et al., 2012).

Global concerns about E. sinensis derive primarily from its burrowing activities which undermine river banks and flood defences causing huge economic cost (ca. € 80 million since 1912 in Germany alone) (Gol-lasch, 2006), and also from impediment of commercial fishing operations due to bait interference and clogging of fishing gear (Van Der Velde et al., 2000; Veld-huizen & Stansih, 1999, unpubl. data). Very little attention has been given to the potential ecological impacts of E. sinensis either through its role as an ecosystem engineer, or through trophic interactions. Evidence, mainly from estuarine habitats, suggests E. sinensis is omnivorous, exploiting a range of food sources including macrophytes, algae, detritus, aquatic invertebrates and small fish (Rudnick & Resh, 2005; Czerniejewski, Rybczyk & Wawrzyniak, 2010). Similar patterns of resource use in freshwater environments would render it likely to affect a range of trophic levels directly through consumption, and also indirectly through cascading effects.

Collection and maintenance of animals for experiments

Given the catadromy of E. sinensis, its residence and therefore impacts in freshwater mostly occur during the sexually immature juvenile stage; hence, juvenile decapods were used for all laboratory and mesocosm experiments. Decapods were collected from multiple locations in the UK and combined to form a laboratory stock. Eriocheir sinensis were collected from the estuarine River Thames at Chiswick Eyot (51°29'13.97"N, 0°14'44.81"W) using hand search, and from the tidal limit of the river Blackwater at Beeleigh (51°44'34.31"N, 0°39'41.85"E) as by-catch within an elver monitoring trap. Pacifastacus leniusculus were collected from freshwater reaches of the River Pant (51°55'28.14"N, 0°31'16.59"E), and the nearby River Glem (52° 5'33.44"N, 0°41'36.69"E) using hand search. Juvenile native A. pallipes were collected under license (Natural England #20122661) from Adel Beck (53°51' 20.80", -1° 34' 29.91") using hand search and were returned to the collection site after study completion. Eriocheir sinensis ranged from 20.4 to 30.5 mm carapace width, 3.6 to 10.8 g wet mass, corresponding to sexually immature juveniles of less than 2 years age (Dittel & Epifanio, 2009). Pacifastacus leniusculus ranged from 19.2 to 32.7 carapace length, 3.4 to 10.6 g (wet mass), corresponding to sexually immature crayfish of less than 2 years (Guan & Wiles, 1999). Austropotamobius pallipes ranged from 25.2 to 28.3 carapace length, 5.1 to 6.8 g

(wet mass), corresponding to 2-3 year old juveniles (Pratten, 1980).

Decapods were maintained in aquaria (38 L) filled with dechlorinated tap water (17°C, 16 h light: 8 h dark) and fed a diet of crab pellets (Hinari) and algal wafers (King) for a minimum of 3 weeks prior to the start of experiments. Species were maintained separately, with up to eight animals per tank. Shelters (plastic pipe sections) were provided to reduce aggressive interactions and risk of injury. To ensure animals had experience of encountering all the freshwater prey types to be offered in experiments, a kick sample of macroinvertebrates collected from Meanwood Beck, UK (53°490 51.60", -1°34' 37.19") was added to each tank weekly. Terrestrial leaf litter (20 g) as leaves of beech (~80%) and alder (~20%) soaked for over 2 months, and fresh algae (Cladophora sp.) (5 g) were also added each week.

Macroinvertebrates were collected from streams and ponds located within 100 km of Leeds, U.K., using a combination of kick sampling, hand searching and sweep netting, with the exception of chironomid larvae which were purchased live from a pet retailer. After collection, invertebrates were transported to the laboratory, sorted into taxa and maintained separately in aerated aquaria (8 L) until required.

The eggs of two common UK freshwater coarse fish species, roach (Rutilus rutilus) (Cyprinidae) and common bream (Abramis brama) (Cyprinidae), were collected on the day following fertilisation from a restocking facility. Eggs were retained on the spawning medium (Matala filter mat, CA, U.S.A.) in aerated water (17 °C) until use (<3 days).

Prey preference experiments

The prey preferences of E. sinensis and P. leniusculus were compared using four mobile prey items widespread in UK freshwaters: the amphipod Gammarus pulex (Gammaridae), the isopod Asellus aquaticus (Aselli-dae), the gastropod Radix peregra (Lymnaeidae) and chi-ronomid larvae (Chironomidae). Size-matched juvenile E. sinensis and P. leniusculus (5 ± 0.4 g, wet mass) were isolated in individual aquaria (8 L) filled with 2 L water (50 mm depth), aerated via an air stone and maintained at 17 °C, 16 h light: 8 h dark. The sides of the aquaria were covered in black plastic to reduce stress and promote foraging. Decapods were starved for 24 h prior to the start of the experiment at which point 20 individuals of each prey type were added to each aquarium, with one prey type per corner of the tank. To reduce the possibility of total prey depletion, the experiment lasted

; four hours in light conditions (Guan & Wiles, 1998; Jin et al., 2001). At the end of the experiment, the remaining prey items were counted. A total of 10 replicates were carried out per treatment group (E. sinensis and P. le-niusculus), along with five controls with no decapod present. Each decapod was used only once.

In a second experiment, predation by E. sinensis and P. leniusculus on the eggs of two common species of coarse fish was investigated by means of a simple pair-wise choice. Eriocheir sinensis and P. leniusculus (9.5 ± 1.5 g, wet mass) were isolated and starved as before, then 50 eggs of Rutilus rutilus and 100 eggs of Abramis brama were introduced to the aquaria. Twice as many A. brama eggs were used because they were 1 approximately half the size of the R. rutilus eggs. Due to the fragile and sticky nature of the eggs, it was not feasible to remove them from the spawning medium. Instead, this was cut into small squares (~6 cm2), ensuring the appropriate number of eggs were present on each. The experiment ran for 23 h, after which the remaining eggs were retrieved and counted. There were seven replicates per treatment group (E. sinensis and P. leniusculus), along with seven controls with no decapod present. Each decapod was used only once.

Predatory functional response experiments

The invasive decapods E. sinensis and P. leniusculus and the native A. pallipes were tested for differences in their i predatory functional response towards Gammarus pulex, a prey item widely distributed in both lotic and lentic water bodies and a keystone shredder. Size-matched decapods (6 ± 1 g) were isolated in individual aquaria (8 L) and starved for 24 h, as previously described, before prey was added at 17:00 hours. A section of plas-: tic pipe (50 mm diameter, 120 mm length) provided refuge. Gammarus pulex were size matched (12 ± 1 mm, TL) to both standardise biomass between trials and reduce cannibalism (Dick, 1995). Prey were introduced to each treatment group (E. sinensis and P. leniusculus and A. pallipes) at ten different densities (5, 10, 16, 20, • 30, 40, 60, 80, 120 and 160). These prey densities in the experimental arena corresponded to 120, 240, 385, 480, 720, 962, 1442, 1923, 2885 and 3840 individuals m~2. There were four replicates of each density per treatment group, yielding a total of 120 trials. Each trial lasted 24 h, after which the decapod was removed and the number of intact remaining prey items counted. Controls were five replicates of each prey density in the absence of decapods to assess natural mortality and cannibalism among the prey. In between trials, the decapods were

returned to the communal aquaria (38 L) after being marked on the carapace with non-toxic correction fluid to enable identification of individuals. Trials were conducted in a randomised order with each decapod used between one and five times for different prey densities with a recovery period of at least 2 days between successive uses; it was necessary to replace animals as they grew beyond the permitted mass range. To check that reuse of animals did not affect their behaviour, a generalised mixed-effects model (GLMM) was run to identify significant predictors of the proportion of prey eaten in each trial as a function of (i) initial prey density, (ii) species and (iii) number of previous trials in which animal was used, with individual as a fixed factor and weighted to take account of the total number of trials in which an individual was used. The number of previous trials was not a significant predictor of proportion of prey eaten during the trial (P = 0.42); the only significant predictors were species and initial prey density. The mean mass of decapods used was 5.62 ± 0.7, 6.14 ± 0.7 and 6.10 ± 0.7 g (± SD) for E. sinensis, P. leniusculus and A. pallipes, respectively, and did not vary between groups (linear mixed-effect model; V = 0.67, P = 0.41). Data from individuals that moulted within the 3 days following the experiment were excluded because crabs and crayfish reduce or desist from feeding prior to ecdysis (Zhou, Shirley & Kruse, 1998; Reynolds & O'Keefe, 2005).

Mesocosm experiment

To compare the community impacts of E. sinensis and P. leniusculus, both independently and in combination, an outdoor mesocosm experiment of 4 weeks duration was conducted in summer 2012 using circular plastic pools (0.78 m2, 0.65 m depth) (n = 16) sunken into a meadow at the University of Leeds Field Research Unit, UK. Biosecurity measures such as new fencing around the ponds were implemented following consultation with Cefas. Pools were tightly lined with polyester netting (0.9 mm mesh), to aid recovery of invertebrates at the end of the experiment. The bases were covered with a mix of pure sand and dried loam soil (60:40) with a sporadic covering of stones (~20-30 mm, long axis) and the pools filled with groundwa-ter to a depth of 0.5 m (0.39 m3). An aliquot (1 L) of lake water was added to each pool to seed the zooplankton community.

Macrophytes (Ceratophyllum demersum, 55 g wet mass; Callitriche stagnalis, 15 g; Potamogeton perfoliatus, 25 g) were planted in the pools 9 days before the experiment. Filamentous algae (Cladophora sp., 8 g), terrestrial leaf

litter (100 g, 80:20 beech and alder, soaked for over 2 months) and periphytic algae colonised in a large outdoor pond for 6 weeks on two ceramic tiles (16 cm2 surface area each) in identical conditions were also added to each pool to test the effects of P. leniusculus and E. sinensis on basal resources. All macrophytes, algae, detritus and ceramic tiles were hand cleaned of invertebrates before they were added to the pools.

Macroinvertebrates representing a range of functional feeding groups were added to each pool in abundance ratios approximating those observed during collection. The community added to each pool comprised: 27 Mol-lusca (8 Radix peregra: 7 of size ~11 mm, longest axis, and 1 of size ~16 mm, longest axis; 2 Lymnaea stagnalis (Lymnaeidae): ~23 mm, longest axis; 17 Physa fontinalis (Physidae): 12 of size ~6 mm, longest axis, and 5 of size ~3.5 mm longest axis); 5 Trichoptera (Limnephilidae); 56 Isopoda (Asellus aquaticus); 135 Amphipoda (Gammarus pulex), and 500 chironomid larvae. Macroinvertebrates were added 4 days prior to the start of the experiment to allow acclimation in the absence of decapod predators.

Decapods were added to the pools within three treatments: E. sinensis, P. leniusculus, and both E. sinensis and P. leniusculus, in addition to a no decapod control. Pools were assigned using a randomised block design with four replicates in each treatment group and control. Four juvenile decapods were assigned to each treatment pool, with two individuals of each species in the mixed treatment. Eriocheir sinensis ranged from 13 to 22 mm in carapace width (19.14 ± 1.84; mean ± SD) and P. leniusculus ranged from 19 to 26 mm in carapace length (23.11 ± 1.20; mean ± SD). Sex ratios were 50 : 50 in all pools. Total decapod biomass ranged from 19.5 to 22.9 g across all treatment pools and did not vary between treatments (F2,9 = 0.75, P = 0.50). Eight sections of PVC pipe (2 x 50 mm diameter, 120 mm length; 6 x 25 mm diameter, 80 mm length) were added as refugia. After addition of the decapods, the lining nets were closed using cable ties and pools were covered with Envi-romeshR (Agralan, Swindon, U.K.) secured with shock cord to prevent animals escaping and disturbance by birds.

Pools were checked after 2 weeks for decapod mortalities and evidence of moult; moults were removed if found. Midday water temperature ranged from 17.2 to 18.9 °C and did not vary between treatments (ANOVA: F3,12 = 0.73, P = 0.55). Sub-surface water samples collected at the end of the experiment for chemical analysis showed no difference between treatments for the main parameters (ANOVA: Nitrate F3,12 = 1.47, P = 0.27;

Phosphate F^ = 0.56, P = 0.65; Sulphate F^ = 1.49, P = 0.27 and Calcium F3,12 = 1.33, P = 0.31).

At the end of the experiment, the decapods were collected, the ceramic tiles were frozen and the net linings transported to the laboratory in plastic bags for processing. Macroinvertebrates, algae, terrestrial leaf and macrophyte fragments were carefully recovered using a net (1 mm mesh size) and sorted. Macroinvertebrates were counted and macrophytes, algae and leaf fragments (exceeding ~4 mm) were hand cleaned of macroinvertebrates, blotted dry and weighed. Total chlorophyll was used as a proxy measure for the remaining biomass of periphytic algae on the ceramic tiles. Each tile was soaked overnight in 90% ethanol, then, extractants were centrifuged at 4 x 104 rpm for 20 min and analysed using a spectrophotometer (Biochrom WPA Biowave II) to measure absorbance at 750 nm, 664 nm, 647 nm and 630 nm wavelengths (1 cm path length). Total chlorophyll (ig) per tile was calculated as the sum of chlorophyll-a and b (Huang & Cong, 2007).

Stable-isotope and gut-content analysis of wild-caught specimens

Wild E. sinensis and P. leniusculus were collected from two sites on the River Stour, Suffolk, U.K., during October and November 2012. The two species have been sympatric in this lowland watercourse for at least 10 years, and have been observed at locations within 22 km (Adam Piper, Environment Agency, pers. comm.). Eriocheir sinensis were collected immediately upstream of the tidal limit (51°57'17.59"N, 1°1'32.31"E) and P. lenius-culus were collected 62 km further upstream (52°3'31.55" N, 0°29'32.58"E). Sites exhibited similar channel mor-phometry (9 m to 15 m width) with macrophyte communities dominated by Sparganium spp., Phalaris arundinacea and Nuphar lutea, with overhanging Salix and Alnus spp.

Baited fladen traps were deployed at both sites and checked daily. The bait (detrital leaves, chironomid larvae and sardine in oil) was encased within a nylon mesh (1 mm) and metal mesh box (5 cm2) to ensure that animals could not consume it. Captured decapods were immediately frozen. Collections of potential diet items were made at both sites during the same period using a combination of kick sampling, dredge trawling and hand collection. All macroinvertebrates (Lymnaea sp., Theodoxus fluviatilis, chironomid larvae, Ephemeroptera, Gammarus pulex, Asellus aquaticus, Limnephilidae, Notonectidae and Odonata) were maintained live in

distilled water for 24 h to clear their gut contents before being frozen. Plant material (Elodea canadensis, Nuphar lutea, Cladophora sp., Phalaris arundinacea, decaying Spar-ganium erectum, Rorippa nasturtium-aquaticum, Myosotis scorpioides and assorted terrestrial detritus) was carefully rinsed in distilled water prior to freezing. Three small fish (Perca fluviatilis, Rutilus rutilus and Gasterosteus aculeatus) found dead in the trap netting and a juvenile Gobio gobio accidentally killed during a dredge trawl were filleted to isolate the muscle tissue before freezing.

Only adult decapods were captured during the sampling. Claw muscle tissue from E. sinensis (n = 5) (54 to : 87 mm carapace width) and P. leniusculus (n = 4) (3746 mm carapace length) was extracted, freeze dried, weighed and analysed for stable isotope ratios (13C:12C and 15N:14N) expressed as 8 values (&). In addition to : the decapods, samples of 14 and 17 potential diet items were analysed from the E. sinensis and P. leniusculus collection sites respectively. All macroinvertebrates were separated into genera, freeze dried and combusted whole, with the exception of gastropods for which only the muscle tissue of the foot was used. Fish muscle was freeze dried and weighed. All plant material was freeze dried, then immersed in liquid nitrogen and ground to a fine homogenous powder using a pestle and mortar before weighing. Due to restrictions on the number of samples that could be analysed and to ensure sufficient mass of material, composite samples were used for the following animal groups: chironomids (10-23 individuals), Asellus aquaticus (2 individuals), Notonectidae (2-3 individuals), Theodoxus fluviatilis (4 individuals) and Odonata (2 individuals). All plant samples comprised a minimum of three leaves/stems.

Samples were analysed at the University of Leeds using an Isoprime continuous flow mass spectrometer coupled to an Elementar Pyrocube elemental analyser. ! Standards of ammonium sulphate USGS-25 (—30.1&) : and USGS-26 (+53.7&) for Nitrogen; and ANU-sucrose (—10.47&) and IAEA-CH-7 (polyethylene film, —31.83&) for carbon, were interspersed every 8-12 samples to calibrate the system and compensate for drift. Stable isotope ratios are expressed in conventional notation as parts per thousand (&) using delta notation (8), relative to international standards (Pee Dee Belemnite for carbon and atmospheric nitrogen). Analytical precision on both i isotope measurements was 0.2& or better.

Foregut contents of E. sinensis (n = 5) (37-46 mm CW) and P. leniusculus (n = 10) (37-52 mm CL) were examined under a dissecting microscope using a gridded Petri dish with 24 squares (25 mm2) sub-divided into smaller squares (1 mm2). First, the number of small

squares (1 mm2) with material present was recorded as a percentage of each larger square (25 mm2). Second, the material in each small square was assigned to one of seven categories: inorganic; algae; macrophyte; moss; leaves (terrestrial); unidentifiable plant matter and macroinvertebrates. Where possible, the macroinverte-brate fragments were identified to order, and occasionally genera.

Data analysis

All statistical analyses were run in R (version 3.0.0, R Core Team 2013) and all mean values are quoted ± standard error (SE). The numbers of prey items remaining in the treatment groups in both the prey preference and egg predation experiments were corrected for the mean reduction recorded during control trials, then the mean total number of prey items (all prey types combined) consumed compared between treatments using a t-test. Selection indices wi were subsequently quantified for each prey type (Manly, McDonald & Thomas, 1993):

wi = —

where ci is the proportion of prey i consumed (corrected for reduction during controls) and ai is proportion of prey i available (corrected for reduction during controls). Indices were standardised by dividing each index by the sum of the four indices then arcsine square root transformed (Rehage, Barnett & Sih, 2005). T-tests were used to compare the mean indices for each prey type between treatments for both experiments, and compare indices between prey types in the egg predation experiment. Mann-Whitney U-tests were used where data could not be normalised. Kruskal-Wallis with post hoc Nemenyi-Damico-Wolfe-Dunn tests were used to compare prey type indices within each treatment for the prey preference experiment.

Differences in decapod mass across species groups in the functional response experiment were tested using linear mixed-effects models (LMEs) with identification number as a random factor because individuals were used multiple times across densities. A chi-square test was used to test for a significant difference in log likelihoods between models with and without species as an explanatory variable. Logistic regression of the proportion of prey consumed against initial prey density indicated that all three decapods exhibited a Type II functional response whereby consumption rate decelerates with increasing prey density (Murdoch, 1973). Therefore, functional response data for each species

were modelled using Rogers random predator equation (Rogers, 1972), modified with the Lambert W function, to obtain coefficients of a (attack rate) and h (handling time):

N = N0 —

where N is the number of prey eaten, N0 is the number of prey supplied, a is attack rate, h is handling time and W is the Lambert W function (Bolker, 2008). This model accounts for decreasing prey density during the trial as prey were not replaced. Data were bootstrapped (n = 2000) and 95% confidence intervals for a and h calculated within the 'frair' package (Pritchard, 2014).

Mesocosm data on the remaining biomass of macro-phytes, chlorophyll concentration (averaged from 2 tiles) and absolute abundances of macroinvertebrate taxa were tested for normality using Shapiro-Wilk test and were log 10 + 1 transformed where necessary. Levene's test was used to determine compliance with the assumption of homogeneity of variance between groups. One-way ANOVA with treatment as a factor and Tukey's HSD post hoc test was used to detect and identify differences between treatment groups. Kruskal-Wallis with post hoc Nemenyi-Damico-Wolfe-Dunn tests were used where data could not be normalised. A between groups test was not conducted for Lymnaea snails as there were too few individuals. Shannon diversity and evenness indices were calculated for each pool and compared among treatments using one-way ANOVA. Percentage change in the biomass/abundance of each taxon was calculated using the final value minus the initial value, as a percentage of the initial value.

Four E. sinensis individuals moulted during the course of the study. Two P. leniusculus individuals in separate pools in the P. leniusculus treatment were missing at the end of the experiment. There was no evidence that the animals had climbed out of the tanks or broken through the netting, so it was assumed that they had died (perhaps during moult) and been consumed by the other decapods and detritivores. It was decided not to exclude these pools from the dataset because checks on day 14 revealed that all decapods were still present, hence pools had their intended decapod biomass for at least half the duration of the experiment. Further, preliminary analyses of the data revealed that macroinvertebrate abundance (all species) and macro-phyte biomass (all species) of the two pools in question did not differ significantly from other pools within the same treatment.

Volumetric proportions of food types from gutcontents analysis were arcsine root transformed and compared between decapod species using independent samples t-tests. Delta values for stable isotopes 13C and 15N measured in field samples were also compared between decapod species and between functional groups across the two collection sites using independent samples t-tests. Bayesian stable-isotope mixing models (SIAR; Parnell et al., 2008) were used to estimate the relative contributions of the potential food sources sampled to the diet of E. sinensis and P. leniusculus. Assumed fractionation factors of 2.4 ± 0.18 & for 815N and 0.5 ± 0.17 & for S13C, based on a meta-analysis of studies using non-acidified samples (Mccutchan et al., 2003), were used to adjust the isotopic values of food sources. Elemental concentrations of C and N within each of the food sources were also incorporated in the model to account for concentration-dependent variation in frac-tionation (Phillips & Koch, 2002).

Results

Prey preference and egg predation

Overall, E. sinensis consumed more prey items per trial than P. leniusculus (mean 26.04 ± 2.86 and 16.48 ± 2.61 respectively) (t18 = 2.5, P = 0.02) and all four prey types were consumed by the decapods to some extent.

Consumption in the control (by G. pulex and/or A. aquaticus) was less than half the consumption in the : presence of the decapods (t23 = 4.4, P < 0.01), although there was a considerable reduction in chronomid larvae in the control (mean 14 prey items).

Comparison of selection indices between the decapod predators indicated a greater preference for G. pulex among E. sinensis compared to P. leniusculus (t18 = 3.22, P < 0.01) (Fig. 1). There was no difference between predators for the other three prey types (Mann-Whitney U-tests, U = 0.33, P = 0.35; U = 0.53, P = 0.63; U = 0.69, P = 0.53 for A. aquaticus, chironomid larvae and R. pere-gra respectively). Chironomid larvae were the most preferred prey type of both E. sinensis (H3,40 = 28.4, P < 0.01) and P. leniusculus (H3,40 = 15.2, P < 0.01). Both invasive decapods preyed heavily on the fish eggs relative to the control (Mann-Whitney U-test, U = 3.33, P < 0.01), eating the majority offered (60 to 100% across all trials). There was no difference in overall consumption between decapod species (t12 = 0.30, P = 0.77), but there was a preference for R. rutilus eggs among both E. sinensis and P. leniusculus (t12 = 20.63, P < 0.01 and t12 = 4.17, P < 0.01 respectively).

Predatory functional response

The maximum predatory functional response of E. sinen-sis (44 prey items) was 57% higher than that of the

Fig. 1 Prey selection indices (mean ± SE) during trials (n = 10) in which predators and Eriocheir sinensis and Pacifastacus leniusculus were offered four prey items simultaneously: Gam-marus pulex (amphipod), Asellus aquaticus (isopod), chironomid larvae and Radix peregra (gastropod). Data corrected for mean prey reduction during control trials (n = 5) when no decapod was present. *denotes significant difference at 0.05 level.

native A. pallipes (28 prey items). Pacifastacus leniusculus had an intermediate maximum between E. sinensis and A. pallipes of 35 prey items. The functional response curve of E. sinensis was significantly higher than A. pallipes, whereas the upper 95% confidence interval of the P. leniusculus curve overlapped with the lower 95% confidence interval of the E. sinensis curve indicating no significant difference between the two species (Fig. 2). The lower 95% confidence interval also overlapped with the upper A. pallipes confidence interval, similarly indicating no difference in consumption between the species

(Fig. 2). Attack rate did not vary between decapod species, but handling time did (P < 0.05). Eriocheir sinensis exhibited a faster handling time than A. pallipes, but handling time by P. leniusculus did not differ from the other two decapod species (Fig. 3).

Outdoor mesocosm experiment

Basal resources. The mass of terrestrial leaf litter was reduced by on average 43% (± 3.9) in the decapod treatments and 35% (± 2.6) in the control; there was no significant difference in detrital mass among treatments at the end of the experiment (Table 1). Change in biomass of filamentous algae (Cladophora sp.) varied greatly between individual pools, with no significant difference between treatments (Table 1). Periphyton chlorophyll at the end of the experiment was on average 48% higher in the combined E. sinensis and P. leniusculus treatment relative to control, but did not differ between other treatments (Fig. 4, Table 1). There was a general reduction in the biomass of all three macrophyte species during the experiment. Callitriche stagnalis was absent from all except four pools (which were within a range of treatments), and so was omitted from further analyses. Remaining macrophyte biomass did not vary between treatments (Table 1).

Shredders. The abundance of G. pulex was significantly reduced in both treatments containing E. sinensis, relative to the control, but not in the P. leniusculus only

Fig. 2 Fitted functional response curves (dashed lines) for three decapod predators Eriocheir sinensis, Pacifastacus leniusculus and Austropotamobius pallipes preying upon Gammarus pulex. Shaded polygons indicate empirical 95% confidence intervals generated by bootstrapping (n = 2000).

Fig. 3 Functional response attack rates (a) and handling times (h) for three decapod predators Eriocheir sinensis, Pacifastacus leniusculus and Austropotamobius pallipes preying upon Gammarus pulex. Error bars indicate empirical 95% confidence intervals generated by bootstrapping (n = 2000). *denotes difference at 0.05 significance level.

Table 1 Biomass of macrophytes, mass of detritus and abundances of invertebrate taxa remaining in pools (n = 16) after four weeks compared across treatments: Eriocheir sinensis, Pacifastacus leniusculus, both Eriocheir sinensis and Pacifastacus leniusculus, and no decapod control using one-way ANOVA and Tukey's HSD post-hoc tests (or Kruskal-Wallis and Nemenyi-Damico-Wolfe-Dunn post hoc tests where data could not be normalised). Bold values indicate difference at 0.05 significance level.

Treatment effect test statistic

F/H342 P

Pairwise comparisons

E. sinensis versus control P

P. leniusculus versus control P

Both versus

control

E. sinensis versus P. leniusculus P

E. sinensis versus both P

P. leniusculus versus both P

Basal resources

Potamogeton perfoliatus 0.845 0.495 -

Ceratophyllum demersum 2.424 0.116 -

Cladophora sp. 1.675 0.225

Terrestrial leaf litter 0.857 0.49 -

Periphyton 4.251 0.017 0.075 Grazers

All 8.571

Radix peregra 12.91(H)

Physa fontinalis 4.370

Lymnaea stagnalis n/a Shredders

Gammarus pulex 13.560 <0.001 <0.001

Asellus aquaticus 28.09 <0.001 <0.001

Trichoptera sp. 11.76(H) 0.008 0.009 Filterers/ collectors

Chironomid larvae 8.99(H) 0.029 0.043

0.003 0.002 0.005 <0.001 0.027 0.022

0.042 <0.001

0.075 <0.001

0.017 0.014

0.006 <0.001 0.004

0.317 1.000 0.471

0.733 0.571

0.603 0.973 0.873

0.220 0.583 0.997

0.946 0.973 0.880

0.518 0.993 0.401

'e Periphytic aigae

" 7-1 ■o

о " ---

H Control E. sinensis P. leniusculus Both

Fig. 4 Periphyton biomass (mean ± SE) on tiles after four weeks in mesocosm pools (n = 16) within the treatments: Eriocheir sinensis, Pacifastacus leniusculus, both Eriocheir sinensis and Pacifastacus leniusculus, and no decapod control.

treatment relative to control (Table 1). Conversely, the other shredders, Asellus aquaticus and larvae of the order Trichoptera (Limnephilid sp.) were significantly reduced in all the decapod treatments relative to the control (Fig. 5, Table 1).

Grazers and collectors. The abundance of Gastropoda increased in the controls during the experiment, but was

significantly reduced in all the decapod treatments with no differences among them (Table 1, Fig. 5). The abundance of chironomid larvae was similarly significantly reduced in all decapod treatments relative to the control, but did not vary among the three decapod treatments (Table 1).

The Shannon diversity index in each pool at the end of the experiment ranged from 0.44 to 1.44 and did not vary between treatments (F3,i2 = 1.43, P = 0.283). Shannon evenness ranged from 0.28 to 0.86 and also did not vary between treatments (F3,12 = 0.28, P = 0.839).

Stable-isotope analysis

Eriocheir sinensis had a significantly lower 813C value than P. leniusculus with a mean of —29.90 ± 0.21 & compared to —28.9 ± 0.17 & (t6.3 = 3.85, P < 0.01). Mean 815N values were 17.04 ± 0.41 & and 17.40 ± 0.25& for E. sinensis and P. leniusculus, respectively, and did not vary between species (t6.3 = 0.76, P = 0.48). With regard to potential food sources, some species sampled varied between the two collection sites but isotope signatures of functional groups were similar, with the exception of Gastropoda for which 813C of the single composite sample collected at Flatford was shifted (—38.9&) compared to the three samples collected at Wixoe (—30.0 to 31.8 &) (Fig. 6).

Fig. 5 Percentage change (mean ± SE) in abundance of Gammarus pulex, Asellus aquaticus, Gastropoda and Trichoptera larvae after four weeks in mesocosm pools (n = 16) within the treatments: Eriocheir sinensis, Pacifastacus leniusculus, both Eriocheir sinensis and Pacifastacus leniusculus, and no decapod control.

Eriocheir sinensis were more 15N-enriched than chi-ronomids, amphipods and isopods (t4.4 = 2.85, P = 0.04) and Gastropoda, and 15N-depleted relative to fish, thereby occupying the same trophic level as the Ephe-meroptera, Odonata and Notonecta, which were similarly 15N-enriched (t4.9 = 1.05, P = 0.34) (Fig. 6). Iso-topic signatures indicated a similar trophic position for P. leniusculus at the second collection site (Fig. 6). Equivalent basal resources did not differ between the two collection sites in their 813C values (macrophytes: t = 1.31, P = 0.24, 5.4 d.f.; terrestrial leaf litter: t = 1.4, d.f. = 1.86, P = 0.31). For filamentous algae, the one composite sample differed marginally between sites, though the difference (0.02& for S13C) was less than analytical precision. In both sites, 813C values of the decapods were intermediate between those of macrophytes and terrestrial leaf litter; however, E. sinensis was shifted towards macro-phytes and P. leniusculus towards terrestrial detritus and filamentous algae, indicating differences in the basal energy sources used by these invaders.

Concentration-dependent mixing models estimated that basal resources comprised the majority of the

diet of both decapods. Eriocheir sinensis relied most on macrophytes followed by terrestrial leaf litter, whereas P. leniusculus was most dependent on terrestrial leaf litter and filamentous algae (Fig. 7). Estimated contributions of the remaining potential diet items sampled were broadly similar between the two decapod species, though chironomids appeared slightly more important for P. leniusculus than E. si-nensis (Fig. 7).

Gut-content analysis

Invertebrate material comprised the largest proportion of E. sinensis gut contents (n = 5) 21.1% (± 1.53), followed by macrophytes (16.7 ± 2.7%) and algae (10.7 ± 1.7%). Terrestrial detritus was the least detected category, comprising on average 6.5 ± 1.04%. The invertebrate taxa detected in E. sinensis gut contents included Trichoptera, Coleoptera, Ephemeroptera and Gastropoda, with Trichoptera the most commonly encountered. In contrast to E. sinensis, the gut contents of P. leniusculus (n = 10) was significantly more

(a) 24 -,

22 ■ 20 ■ 18

£ 16 ■

мэ~ i4 12 10 8 ^

(b) 24

14 12 10

Е. sinensis collection site

Predatory responses and interactions of invasive decapods 437 (a) 1.0-

б ■ -40

-36 -32 -28

5„C (%o)

P. leniusculus collection site

■ Gastropoda П Ephemeroptera • Fish

A Macrophytes

-36 -32

onC (%o)

+ Amphipods and isopods Odonata, Notonecta Л Terrestrial leaf litter О Filamentous algae X Chironomids

l-1-1-1-1-1-1-г

Macrophytes Gastropods Chironomids Fish Filamentous Terrestrial Isopods & Ephemeroptera, 1— leaf litter amphipods Odonata &Notonecta

Source

Fig. 6 Stable-isotope biplots for (a) Eriocheir sinensis (triangles) (n = 5) and (b) Pacifastacus leniusculus (open circles) (n = 4), and potential food sources (n = 1 to 5) in the lower River Stour, U.K. Where more than one source or composite sample was analysed, signatures denote mean values (±1 SE) adjusted for trophic enrichment factors.

Fig. 7 Results of SIAR (95, 75 and 50% credibility intervals) showing estimated contribution of each potential food source to the diet of a) Eriocheir sinensis and b) Pacifastacus leniusculus.

dominated by terrestrial leaf litter (38.2% ± 2.7) (t13 = 9.04, P < 0.01). Algae was second most common (14.3% ± 1.3) and did not differ between decapods (t13 = 1.85, P = 0.087). Invertebrates were found in lower proportion in P. leniusculus (12.9% ± 1.4) than E. sinensis (t13 = 3.39, P< 0.01), and included the taxa Gastropoda, Isopoda and Trichoptera. Fish scales were also detected in three P. leniusculus individuals.

Discussion

Using multiple approaches at a range of scales, our study identified differences in the trophic interactions and potential impacts of two omnivorous decapods which are rapidly expanding their invasive freshwater ranges. Quantitative comparison of predatory functional response showed a higher per capita impact of E. sinen-sis than native A. pallipes on a keystone shredder. Preference of E. sinensis for this prey item was supported by prey choice trials. Community impacts of the decapods

investigated through a mesocosm experiment were broadly similar with marked declines in all animal taxa; however, there was again greater impact on the shredder G. pulex by E. sinensis compared to the invasive crayfish. Presence of the decapods in sympatry caused a trophic cascade resulting in elevated periphyton levels.

Eriocheir sinensis and P. leniusculus both appeared to operate as 'generalist omnivores' consuming a wide variety of food items, which accords with previous studies (Guan & Wiles, 1998; Rudnick & Resh, 2005; Stenroth et al., 2006; Czerniejewski et al., 2010), although clear preferences for certain prey types were detected. The selectivity hierarchy of both decapods generally reflected a decline in preference that may reflect ease of capture and handling, with chironomids most preferred and the gastropod least preferred. Prey used in laboratory experiments and field mesocosms were from sites where crayfish and E. sinensis had not yet been detected and were therefore presumed naive to these predators. Prey may exhibit diverse and often complex predator-avoidance behaviours (Covich et al., 1994; Cotton, Rundle & Smith, 2004; Sih et al., 2010), which may be evolved over time and passed to successive generations genetically, or reflect localised behavioural plasticity (Alvarez & Nicieza, 2003); therefore, use of naive prey may have led to overestimation of predation rates in this study.

Both E. sinensis and P. leniusculus preyed heavily upon the eggs of coarse fish, with apparent preference for R. rutilus, perhaps reflecting greater ease of handling of these larger eggs. In several trials, decapods consumed all the eggs that were accessible to them, i.e. not deeply embedded in the spawning medium, suggesting that they would have consumed more than they had been provided. Only fish eggs were available in this trial so we cannot assess the preference for fish eggs relative to other items, however our results do support previous suggestions that eggs are likely to be present in the diet of both E. sinensis (Culver, 2005, unpubl. data; Morritt et al., 2013) and P. leniusculus (Edmonds et al., 2011). These data suggest that invasion by P. leniusculus and E. sinensis may impact recruitment of these common fish species which spawn on vegetation in the mid to lower reaches of rivers, where both E. sinensis and P. leniuscu-lus reach their highest densities in freshwaters (Rudnick et al., 2003; Weinlaender & Fuereder, 2009).

Invasive E. sinensis displayed a 57% higher per capita consumption rate on a keystone freshwater shredder, G. pulex, compared to the native crayfish A. pallipes. There was also a non-significant trend suggesting that E. sinensis may also be a stronger predator than P. le-niusculus for this prey type, supported by prey choice

trials in which E. sinensis consumed more prey items overall and displayed a greater preference for G. pulex than did P. leniusculus. Further, in the mesocosm experiment, the abundance of this prey was reduced more in the presence of E. sinensis than P. leniusculus. A stronger predatory functional response among invaders compared to natives has been demonstrated previously for crayfish (Haddaway et al., 2012); amphipods (Bollache et al., 2008), and gambusias (Rehage et al., 2005). Haddaway et al. (2012) showed that P. leniusculus preyed at a 10% higher rate than A. pallipes, and although this trend was also observed in this study with the same species, no clear species difference was apparent due to high intraspecific variability. From the higher predation rate of E. sinensis relative to the native decapod, mediated through a faster ability to handle this prey item, we infer that E. sinensis is likely to negatively impact native prey species as it invades.

Our ability to control for predator density afforded by the mesocosm and laboratory approaches enabled measurement of per capita effects on native prey. This is one of the three key elements considered important for prediction of invader impact, along with area invaded and abundance (Parker et al., 1999; Dick et al., 2014). There is evidence that P. leniusculus reaches higher densities (1015 individuals m~2 in a U.K. lowland river, Guan & Wiles, 1996; 26-39 individuals m~2 in a U.K. stream, Peay et al., 2014) than the native A. pallipes (2-4 individuals m~2 in a stream in France, Grandjean et al., 2000; 5.3 m~2 in a U.K. river, Pearson, 2011, unpubl. data), although direct comparison of densities in similar habitat are understandably lacking due to the spread of crayfish plague and generally rapid replacement of A. pallipes by P. leniusculus where they co-occur. There is generally a dearth of knowledge on population densities and the long-term dynamics of E. sinensis in invaded areas, particularly in its freshwater range. In one comprehensive study over 6 years, mean crab abundance increased to an estimated 30 individuals m~2 in 1999 in the southerly freshwater tributaries of San Francisco Bay, before a decline to 21.1 individuals m~2 in 2000 (Rudnick et al., 2003). In the upper tidal zone of the River Thames, abundances ranged from 0.6 to 2.25 juvenile crabs m~2 of boulder habitat (Gilbey et al., 2008). The higher per capita consumption by E. sinensis observed in our functional response experiment, combined with higher densities of both invasive species in the wild, suggest that impacts on native prey abundances are likely to be higher in invaded areas than in the presence of the native decapod alone. Further data concerning the population densities of both decapods in

their invasive range is necessary to build on our quantitative per capita consumption rates to inform predictions of invader impacts in the wild.

In the simplified mesocosm communities, the effects of decapod presence were generally similar for E. sinen-sis and P. leniusculus, with a strong decline in the abundances of all animal taxa. The stronger impact of E. sinensis than P. leniusculus on G. pulex is consistent with its demonstrated preference and tendency towards a higher functional response for this prey item, providing strong evidence that invasion by this decapod is likely to reduce shredder abundance in freshwater systems to a greater extent than would be the case where only crayfish are present. This could have cascading effects whereby reduced shredder abundance results in a dramatic decline in detrital processing. Woodward et al. (2008) found the predatory impacts of bullhead on G. pulex in a chalkstream caused a dramatic decline in detrital processing. Conversely, macroconsumers of detritus such as crayfish may decouple such a cascade, functionally replacing the more specialised shredders and thereby still creating availability of nutrients to pass to higher levels (Usio & Townsend, 2000; Vanni, 2002; Moore et al., 2012), although crayfish effects are likely to be species- (Dunoyer et al., 2014) and size-dependent (Mancinelli, Sangiorgio & Scalzo, 2013). Our combined evidence from the mesocosms, gut contents and stable isotopes indicated that E. sinensis is also a significant consumer of terrestrial leaf litter, so, depending on comparative processing rates and the degree to which shredder populations are reduced, this invader could similarly decouple such a cascade.

There was evidence of additive community impacts in the combined presence of both decapods. Periphyton biomass increased significantly in the joint presence of both invasive decapods, but did not differ in single species and control treatments. The observed increase in periphyton is likely to be a consequence of a top-down cascade created by the decapods feeding on gastropods, and thereby reducing grazing pressure on algae. A similar cascade has been reported caused by reduction of gastropods in the presence of single species invasion by the crayfish Orconectes rusticus (Weber & Lodge, 1990; Charlebois & Lamberti, 1996). One explanation for this cascade in the sympatric treatment may be a synergistic effect of the invasive decapod species. Eriocheir sinensis had a stronger impact than P. leniusculus on the abundance of the gastropod P. fontinalis in mesocosms; however, impacts on the other grazer R. peregra were similar, so it is unclear why E. sinensis in isolation did not also cause a trophic cascade. An alternative explanation may

be that, along with grazing pressure by the gastropods, periphyton was directly consumed by E. sinensis to a greater extent than by P. leniusculus. The importance of periphyton in the diet of mitten crabs has been reported previously (e.g. Czerniejewski et al., 2010 for Chinese mitten crab, Kobayashi, 2009 for Japanese mitten crab), and although crayfish may graze directly on periphyton, they are inefficient consumers compared to gastropod : grazers (Nystrom, 1999). Gastropods in the sympatric treatment were reduced by an intermediate amount relative to the allopatric treatments; hence, the observed increase in periphyton may reflect reduced grazing pressure. However, compared to the E. sinensis allopatric treatment with four crabs per pool, there was likely to be half as much direct consumption by the decapods (two crabs per pool), potentially causing the outcome of higher periphyton than the allopatric treatments and control.

Stable-isotope analysis of wild specimens supports the indication from laboratory and mesocosm results that the invasive decapods share a varied diet, with potential for overlap and therefore competition for resource use. Both decapods were similarly enriched in 15N and therefore occupy the same trophic level. The alignment of P. leniusculus towards algal basal resources, compared to the closer alignment of E. sinensis with macrophytes was supported by the gut-content analysis and may reflect dietary preference or variation in their availability between the two study reaches. Mixing models suggest that algal and plant materials constitute the majority of the diet for both species. Due to limitations in sample collection and analysis, only small samples of the wild decapods were captured and analysed, and only adult specimens, so the stable-isotope and gut-content analyses presented here provide only an initial indication of dietary patterns and interpretation must be cautious. For example, there may be undetected temporal variation or ontogenetic differences in diet, though there is little evidence for ontogenic diet shift among P. leniuscu-lus (Bondar & Richardson, 2009; Stenroth et al., 2006; Usio et al., 2009; but see Bondar & Richardson, 2013) or E. sinensis (Rudnick, Halat & Resh, 2000, unpubl. data).

Comparisons of invader impacts are crucial for managers to assess where best to target limited resources for invasive species control. Our study provides the first quantitative comparison of potential impacts of P. lenius-culus and E. sinensis on the communities they invade. The most widely reported impacts of E. sinensis are for estuarine environments where it causes substantial bank erosion (Dittel & Epifanio, 2009). However, our results suggest that the spread of E. sinensis into freshwaters is

also cause for concern due to structural ecosystem effects including reduction in the abundances of native prey and altered community composition; particularly as it is likely that this invader will have an equal, if not higher, per capita impact on prey species than P. lenius-culus. The community impacts of an invader in the wild will depend on an array of interlinked factors including population density, the availability of prey, habitat complexity and other biotic interactions (Parker et al., 1999). Our use of several approaches, ranging from fully controlled laboratory experiments to analysis of wild-caught specimens, highlights the benefits of supplementing quantitative per capita measurements with community experiments to better understand the mechanisms of potential community impacts and facilitate prediction of invader impacts.

Acknowledgments

The authors thank Environment Agency staff at Calver-ton fish farm for their assistance in sourcing fish eggs, David Morritt, Paul Clark, Dan Hayter and Ben Norring-ton for assistance with sourcing E. sinensis samples and Cefas for advice on biosecurity measures for the meso-cosm experiment. We also thank Jonathan Grey for useful input during the initial stages of this manuscript and two anonymous reviewers whose comments led us to greatly improve it. PR was funded by a NERC Case Studentship with partners Lafarge-Tarmac. PR formulated the idea. PR, AD & RM designed the study. PR, CW & CG collected the data. PR analysed the data. RN ran the stable isotope analysis and assisted with data analysis of this section. AD & RM supervised study design and research. PR, AD, RN & RM wrote the manuscript.

References

Alderman D.J., Holdich D. & Reeve I. (1990) Signal crayfish as vectors in crayfish plague in Britain. Aquaculture, 86, 3-6.

Alvarez D. & Nicieza A. (2003) Predator avoidance behaviour in wild and hatchery-reared brown trout: the role of experience and domestication. Journal of Fish Biology, 63, 1565-1577.

Axelsson E., Nystrom P., Sidenmark J. & Bronmark C. (1997) Crayfish predation on amphibian eggs and larvae. Amphibia-Reptilia, 18, 217-228. Bolker B.M. (2008) Ecological Models and Data in R. Princeton

University Press, Princeton. Bollache L., Dick J.T., Farnsworth K.D. & Montgomery W.I. (2008) Comparison of the functional responses of invasive and native amphipods. Biology Letters, 4, 166-169.

Bondar C.A. & Richardson J.S. (2009) Effects of ontogenetic stage and density on the ecological role of the signal crayfish (Pacifastacus leniusculus) in a coastal Pacific stream. Journal of the North American Benthological Society, 28, 294-304.

Bondar C.A. & Richardson J.S. (2013) Stage-specific interactions between dominant consumers within a small stream ecosystem: direct and indirect consequences. Freshwater Science, 32, 183-192.

Charlebois P.M. & Lamberti G.A. (1996) Invading crayfish in a Michigan stream: direct and indirect effects on peri-phyton and macroinvertebrates. Journal of the North American Benthological Society, 15, 551-563.

Cohen A.N. & Carlton J.T. (1997) Transoceanic transport mechanisms: introduction of the Chinese mitten crab, Eriocheir sinensis, to California. Pacific Science, 51, 1-11.

Cotton P.A., Rundle S.D. & Smith K.E. (2004) Trait compensation in marine gastropods: shell shape, avoidance behavior, and susceptibility to predation. Ecology, 85, 1581-1584.

Covich A.P., Crowl T.A., Alexander J.E. Jr & Vaughn C.C. (1994) Predator-avoidance responses in freshwater decapod-gastropod interactions mediated by chemical stimuli. Journal of the North American Benthological Society, 13, 283290.

Crawford L., Yeomans W.E. & Adams C.E. (2006) The impact of introduced signal crayfish Pacifastacus leniuscu-lus on stream invertebrate communities. Aquatic Conservation: Marine and Freshwater Ecosystems, 16, 611-621.

Czerniejewski P., Rybczyk A. & Wawrzyniak W. (2010) Diet of the Chinese mitten crab, Eriocheir sinensis H. Milne Edwards, 1853, and potential effects of the crab on the aquatic community in the River Odra/Oder estuary (N.W. Poland). Crustaceana, 83, 195-205.

Diamond J.M., Ashmole N. & Purves P. (1989) The present, past and future of human-caused extinctions. Philosophical Transactions of the Royal Society of London B, Biological Sciences, 325, 469-477.

Dick J.T. (1995) The cannibalistic behaviour of two Gam-marus species (Crustacea: Amphipoda). Journal of Zoology, 236, 697-706.

Dick J.T., Alexander M.E., Jeschke J.M., Ricciardi A., Maci-saac H.J., Robinson T.B. et al. (2014) Advancing impact prediction and hypothesis testing in invasion ecology using a comparative functional response approach. Biological Invasions, 16, 735-753.

Dick J.T., Gallagher K., Avlijas S., Clarke H.C., Lewis S.E., Leung S. et al. (2013) Ecological impacts of an invasive predator explained and predicted by comparative functional responses. Biological Invasions, 15, 837-846.

Dittel A.I. & Epifanio C.E. (2009) Invasion biology of the Chinese mitten crab Eriochier sinensis: a brief review. Journal of Experimental Marine Biology and Ecology, 374, 79-92.

Dudgeon D., Arthington A.H., Gessner M.O., Kawabata Z.I., Knowler D.J., Leveque C. et al. (2006) Freshwater

biodiversity: importance, threats, status and conservation challenges. Biological Reviews, 81, 163-182.

Dunoyer L., Dijoux L., Bollache L. & Lagrue C. (2014) Effects of crayfish on leaf litter breakdown and shredder prey: are native and introduced species functionally redundant? Biological Invasions, 16, 1545-1555.

Edmonds N., Riley W. & Maxwell D. (2011) Predation by Pacifastacus leniusculus on the intra-gravel embryos and emerging fry of Salmo salar. Fisheries Management and Ecology, 18, 521-524.

Gilbey V., Attrill M.J. & Coleman R.A. (2008) Juvenile Chinese mitten crabs (Eriocheir sinensis) in the Thames estuary: distribution, movement and possible interactions with the native crab Carcinus maenas. Biological Invasions, 10, 67-77.

Gollasch S. (2006) NOBANIS-invasive alien species fact sheet-Eriocheir sinensis. From: Online Database of the North European and Baltic Network on Invasive Alien Species-NOBANIS URL www.nobanis.org.

Grandjean F., Cornuault B., Archambault S., Bramard M. & Otrebsky G. (2000) Life history and population biology of the white-clawed crayfish, Austropotamobius pallipes pal-lipes, in a brook from the Poitou-Charentes region (France). Bulletin Francais de la Pêche et de la Pisciculture, 356, 55-70.

Guan R.Z. & Wiles P.R. (1996) Growth, density and biomass of crayfish, Pacifastacus leniusculus, in a British lowland river. Aquatic Living Resources, 9, 265-272.

Guan R.Z. & Wiles P.R. (1998) Feeding ecology of the signal crayfish Pacifastacus leniusculus in a British lowland river. Aquaculture, 169, 177-193.

Guan R.Z. & Wiles P.R. (1999) Growth and reproduction of the introduced crayfish Pacifastacus leniusculus in a British lowland river. Fisheries Research, 42, 245-259.

Haddaway N.R., Wilcox R.H., Heptonstall R.E., Griffiths H.M., Mortimer R.J., Christmas M. et al. (2012) Predatory functional response and prey choice identify predation differences between native/invasive and parasitised/un-parasitised crayfish. PLoS ONE, 7, e32229.

Harvey G.L., Moorhouse T.P., Clifford N.J., Henshaw A.J., Johnson M.F., Macdonald D.W. et al. (2011) Evaluating the role of invasive aquatic species as drivers of fine sediment-related river management problems: the case of the signal crayfish (Pacifastacus leniusculus). Progress in Physical Geography, 35, 517-533.

Holling C.S. (1959) Some characteristics of simple types of predation and parasitism. The Canadian Entomologist, 91, 385-398.

Huang T.-L. & Cong H.-B. (2007) A new method for determination of chlorophylls in freshwater algae. Environmental Monitoring and Assessment, 129, 1-7.

Jackson M. & Britton J.R. (2014) Divergence in the trophic niche of sympatric freshwater invaders. Biological Invasions, 16, 1095-1103.

Jackson M.C. & Grey J. (2013) Accelerating rates of freshwater invasions in the catchment of the River Thames. Biological Invasions, 15, 945-951.

Jackson M.C., Jones T., Milligan M., Sheath D., Taylor J., Ellis A. et al. (2014) Niche differentiation among invasive crayfish and their impacts on ecosystem structure and functioning. Freshwater Biology, 59, 1680-1692.

Jin G., Xie P., Li Z. & Kuwabara R. (2001) Food habits and feeding rhythm of the early juvenile Chinese mitten crab (Eriocheir sinensis) in experimental tanks. Journal of Freshwater Ecology, 16, 647-648.

Johnson M.F., Rice S.P. & Reid I. (2011) Increase in coarse sediment transport associated with disturbance of gravel river beds by signal crayfish (Pacifastacus leniusculus). Earth Surface Processes and Landforms, 36, 1680-1692.

Karatayev A.Y., Burlakova L.E., Padilla D.K., Mastitsky S.E. & Olenin S. (2009) Invaders are not a random selection of species. Biological Invasions, 11, 2009-2019.

Keller R.P., Geist J., Jeschke J.M. & Kuhn I. (2011) Invasive species in Europe: ecology, status, and policy. Environmental Sciences Europe, 23, 1-17.

Kobayashi S. (2009) Dietary preferences of the Japanese mitten crab Eriocheir japonica in a river and adjacent seacoast in north Kyushu, Japan. Plankton and Benthos Research, 4, 77-87.

Leprieur F., Brosse S., Garcia-Berthou E., Oberdorff T., Olden J. & Townsend C. (2009) Scientific uncertainty and the assessment of risks posed by non-native freshwater fishes. Fish and Fisheries, 10, 88-97.

Lowe S., Browne M., Boudjelas S. & De Poorter M. (2000) 100 of the World's Worst Invasive Alien Species: A Selection From the Global Invasive Species Database, Invasive Species Specialist Group Species Survival Commission, World Conservation Union (IUCN), Auckland, New Zealand.

Mancinelli G., Sangiorgio F. & Scalzo A. (2013) The effects of decapod crustacean macroconsumers on leaf detritus processing and colonization by invertebrates in stream habitats: a meta-analysis. International Review of Hydrobiol-ogy, 98, 206-216.

Manly B., Mcdonald L. & Thomas D. (1993) Resource Selection by Animals: Statistical Design and Analysis of Field Studies. Chapman & Hall, London.

Mccutchan J.H., Lewis W.M., Kendall C. & Mcgrath C.C. (2003) Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos, 102, 378-390.

Moore J.W., Carlson S.M., Twardochleb L.A., Hwan J.L., Fox J.M. & Hayes S.A. (2012) Trophic tangles through time? Opposing direct and indirect effects of an invasive omnivore on stream ecosystem processes. PLoS ONE, 7, e50687.

Morritt D., Mills H., Hind K., Clifton-Dey D. & Clark P.F. (2013) Monitoring downstream migrations of Eriocheir sinensis H. Milne Edwards, 1853 (Crustacea: Brachyura: Grapsoidea: Varunidae) in the River Thames using cap-

ture data from a water abstraction intake. Management of Biological Invasions, 4, 139-147.

Murdoch W. (1973) The functional response of predators. Journal of Applied Ecology, 15, 237-240.

Noonburg E.G. & Byers J.E. (2005) More harm than good: when invader vulnerability to predators enhances impact on native species. Ecology, 86, 2555-2560.

Nystrom P. (1999) Ecological impact of introduced and native crayfish on freshwater communities: European perspectives. Crayfish in Europe as Alien Species, 11, 63-85.

Nystroom P., Bronmark C. & Graneli W. (1996) Patterns in benthic food webs: a role for omnivorous crayfish? Freshwater Biology, 36, 631-646.

Ojaveer H., Gollasch S., Jaanus A., Kotta J., Laine A.O., Minde A. et al. (2007) Chinese mitten crab Eriocheir sinen-sis in the Baltic Sea—a supply-side invader? Biological Invasions, 9, 409-418.

Panning A. (1939) The Chinese mitten crab. Smithsonian Annual Report 1938, 3491, 361-376.

Parnell A., Inger R., Bearhop S. & Jackson A.L. (2008) SIAR: Stable isotope analysis in R. (http://cran.r-project.org/ web/packages/siar/index.html).

Parker I.M., Simberloff D., Lonsdale W., Goodell K., Won-ham M., Kareiva P. et al. (1999) Impact: toward a framework for understanding the ecological effects of invaders. Biological Invasions, 1, 3-19.

Parker J.D., Torchin M.E., Hufbauer R.A., Lemoine N.P., Alba C., Blumenthal D.M. et al. (2013) Do invasive species perform better in their new ranges? Ecology, 94, 985-994.

Peay S., Dunn A.M., Kunin W.E., Mckimm R. & Harrod C. (2014) A method test of the use of electric shock treatment to control invasive signal crayfish in streams. Aquatic Conservation: Marine and Freshwater Ecosystems, 25, 874-880.

Phillips D.L. & Koch P.L. (2002) Incorporating concentration dependence in stable isotope mixing models. Oecologia, 130, 114-125.

Polis G.A. & Strong D.R. (1996) Food web complexity and community dynamics. American Naturalist, 147, 813-846.

Pratten D.J. (1980) Growth in the crayfish Austropotamobius-pallipes (Crustacea, Astacidae). Freshwater Biology, 10, 401412.

Preston D.L., Henderson J.S. & Johnson P.T. (2012) Community ecology of invasions: direct and indirect effects of multiple invasive species on aquatic communities. Ecology, 93, 1254-1261.

Pritchard D. (2014) Frair: Functional Response Analysis in R. R Package Version 0.4. http://CRAN.R-project.org/packa-ge=frair

R Core Team (2013) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: http://www.R-project.org/.

Rehage J., Barnett B. & Sih A. (2005) Foraging behaviour and invasiveness: do invasive Gambusia exhibit higher

feeding rates and broader diets than their noninvasive relatives? Ecology of Freshwater Fish, 14, 352-360.

Reynolds J.D. & O'Keefe C. (2005) Dietary patterns in stream- and lake-dwelling populations of Austropotamo-bius pallipes. Bulletin Francais De La Peche Et De La Pisciculture, 367, 715-730.

Ricciardi A. (2001) Facilitative interactions among aquatic invaders: is an" invasional meltdown" occurring in the Great Lakes? Canadian Journal of Fisheries and Aquatic Sciences, 58, 2513-2525.

Ricciardi A. & Macisaac H.J. (2011) Impacts of biological invasions on freshwater ecosystems. In Fifty Years of Invasion Ecology: The Legacy of Charles Elton. (Ed. Richardson D.M.), 211-224. Wiley-Blackwell, Chichester.

Rogers D. (1972) Random search and insect population models. The Journal of Animal Ecology, 41, 369-383.

Rogers D. & Watson E. (2011) Distribution database for crayfish in England and Wales. In: Species Survival: Securing White-Clawed Crayfish in a Changing Environment (Eds M. Rees, J. Nightingale & D. Holdich), pp.14-22, Crayfish Conservation South West, Bristol. 16-17 November 2010.

Rothhaupt K.-O., Hanselmann A.J. & Yohannes E. (2014) Niche differentiation between sympatric alien aquatic crustaceans: an isotopic evidence. Basic and Applied Ecology, 15, 453-463.

Rudnick D. & Resh V. (2005) Stable isotopes, mesocosms and gut content analysis demonstrate trophic differences in two invasive decapod crustacea. Freshwater Biology, 50, 1323-1336.

Rudnick D.A., Hieb K., Grimmer K.F. & Resh V.H. (2003) Patterns and processes of biological invasion: the Chinese mitten crab in San Francisco Bay. Basic and Applied Ecology, 4, 249-262.

Ruiz G.M., Fofonoff P., Hines A.H. & Grosholz E.D. (1999) Non-indigenous species as stressors in estuarine and marine communities: assessing invasion impacts and interactions. Limnology and Oceanography, 44, 950972.

Ruokonen T., Kiljunen M., Karjalainen J. & Haomaoloainen H. (2012) Invasive crayfish increase habitat connectivity: a case study in a large boreal lake. Knowledge and Management of Aquatic Ecosystems, 407, 08.

Salo P., Korpimaoki E., Banks P.B., Nordstroom M. & Dick-man C.R. (2007) Alien predators are more dangerous than native predators to prey populations. Proceedings of the Royal Society B: Biological Sciences, 274, 1237-1243.

Sih A., Bolnick D.I., Luttbeg B., Orrock J.L., Peacor S.D., Pintor L.M. et al. (2010) Predator-prey naivete, antipreda-tor behavior, and the ecology of predator invasions. Oikos, 119, 610-621.

Simberloff D., Martin J.-L., Genovesi P., Maris V., Wardle D.A., Aronson J. et al. (2013) Impacts of biological invasions: what's what and the way forward. Trends in Ecology & Evolution, 28, 58-66.

Simberloff D. & Von Holle B. (1999) Positive interactions of nonindigenous species: invasional meltdown? Biological Invasions, 1, 21-32.

Stenroth P., Holmqvist N., Nystrom P., Berglund O., Lars-son P. & Graneli W. (2006) Stable isotopes as an indicator of diet in omnivorous crayfish (Pacifastacus leniusculus): the influence of tissue, sample treatment, and season. Canadian Journal of Fisheries and Aquatic Sciences, 63, 821831.

Stenroth P. & Nystroom P. (2003) Exotic crayfish in a brown water stream: effects on juvenile trout, invertebrates and algae. Freshwater Biology, 48, 466-475.

Strayer D.L. (2010) Alien species in fresh waters: ecological effects, interactions with other stressors, and prospects for the future. Freshwater Biology, 55, 152-174.

Usio N., Kamiyama R., Saji A. & Takamura N. (2009) Size-dependent impacts of invasive alien crayfish on a littoral marsh community. Biological Conservation, 142, 1480-1490.

Usio N. & Townsend C.R. (2000) Distribution of the New Zealand crayfish Paranephrops zealandicus in relation to stream physico-chemistry, predatory fish, and invertebrate prey. New Zealand Journal of Marine and Freshwater Research, 34, 557-567.

Van Der Velde G., Rajagopal S., Kelleher B., Musko I. & De Vaate A.B. (2000) Ecological impact of crustacean invaders: general considerations and examples from the Rhine River. Crustacean Issues, 12, 3-34.

Vanni M.J. (2002) Nutrient cycling by animals in freshwater ecosystems. Annual Review of Ecology and Systematics, 33, 341-370.

Weber L.M. & Lodge D.M. (1990) Periphytic food and predatory crayfish - relative roles in determining snail distribution. Oecologia, 82, 33-39.

Weinlaender M. & Fuereder L. (2009) The continuing spread of Pacifastacus leniusculus in Carinthia (Austria). Knowledge and Management of Aquatic Ecosystems, 394-395,

Williamson M.H. & Fitter A. (1996) The characters of successful invaders. Biological Conservation, 78, 163-170.

Woodward G., Papantoniou G., Edwards F. & Lauridsen R.B. (2008) Trophic trickles and cascades in a complex food web: impacts of a keystone predator on stream community structure and ecosystem processes. Oikos, 117, 683-692.

Woodward G., Speirs D.C., Hildrew A.G. & Hal C. (2005) Quantification and resolution of a complex, size-structured food web. Advances in Ecological Research, 36, 85135.

Zhou S., Shirley T.C. & Kruse G.H. (1998) Feeding and growth of the red king crab Paralithodes camtschaticus under laboratory conditions. Journal of Crustacean Biology,

18, 337-345.

(Manuscript accepted 19 December 2015)