Open Access

Synergistic impacts by an invasive amphipod and an invasive fish explain native gammarid extinction

Contributed equally
BMC EcologyBMC series – open, inclusive and trusted201616:32

https://doi.org/10.1186/s12898-016-0088-6

Received: 2 April 2016

Accepted: 7 July 2016

Published: 14 July 2016

Abstract

Background

Worldwide freshwater ecosystems are increasingly affected by invasive alien species. In particular, Ponto-Caspian gobiid fishes and amphipods are suspected to have pronounced effects on aquatic food webs. However, there is a lack of systematic studies mechanistically testing the potential synergistic effects of invasive species on native fauna. In this study we investigated the interrelations between the invasive amphipod Dikerogammarus villosus and the invasive fish species Neogobius melanostomus in their effects on the native amphipod Gammarus pulex. We hypothesized selective predation by the fish as a driver for displacement of native species resulting in potential extinction of G. pulex. The survival of G. pulex in the presence of N. melanostomus in relation to the presence of D. villosus and availability of shelter was analyzed in the context of behavioural differences between the amphipod species.

Results

Gammarus pulex had a significantly higher susceptibility to predation by N. melanostomus compared to D. villosus in all experiments, suggesting preferential predation by this fish on native gammarids. Furthermore, the presence of D. villosus significantly increased the vulnerability of G. pulex to fish predation. Habitat structure was an important factor for swimming activity of amphipods and their mortality, resulting in a threefold decrease in amphipods consumed with shelter habitat structures provided. Behavioral differences in swimming activity were additionally responsible for higher predation rates on G. pulex. Intraguild predation could be neglected within short experimental durations.

Conclusions

The results of this study provide evidence for synergistic effects of the two invasive Ponto-Caspian species on the native amphipod as an underlying process of species displacements during invasion processes. Prey behaviour and monotonous habitat structures additionally contribute to the decline of the native gammarid fauna in the upper Danube River and elsewhere.

Keywords

Dikerogammarus villosus Gammarus pulex Neogobius melanostomus Selective predation Anti-predator behaviour Species displacement

Background

Worldwide freshwater ecosystems are undergoing major changes in biodiversity, mainly caused by anthropogenic habitat modification and biological invasions [1, 2]. Alteration of habitat and ship traffic are known to be mainly responsible for the dispersal of invasive alien species (IAS) in aquatic ecosystems. Transportation vessels are vectors for introduction of IAS via ballast water or ship-hull transfer from their origins to new areas [3]. In case of successful introduction, the establishment of IAS can result in significant declines of native taxa. Channelized rivers thereby provide both suitable habitat structures [4] as well as migration corridors by interconnecting catchments and enabling dispersal of non-native species [5].

Over the last two decades, ongoing invasions, especially by Ponto-Caspian crustaceans, molluscs and fishes have been reported from the middle and upper sections of the Danube River [69], the Rhine River [10, 11] and other parts of the world [12, 13].

Among these successful invaders, the amphipod Dikerogammarus villosus (Sovinskij 1894) has frequently been proposed to affect native amphipod populations worldwide [5, 1418]. Corresponding to its first records in the German sections of the Danube River in 1992 and the Rhine River in 1995 [19], significant decreases in abundance and distribution of indigenous amphipods such as Gammarus pulex L. 1758 and Gammarus roeselii Gervais 1835 have been observed [16, 20, 21], yet these declines have not been mechanistically linked to the simultaneous increase in the abundance of invasive species. Several recent studies identified functional feeding responses and asymmetric mutual predation of D. villosus as important mechanisms probably facilitating competitive advantages over other amphipods [21, 22]. Previous studies focused on selective predation of native and invasive amphipods by fishes, e.g., rainbow trout Oncorhynchus mykiss, European perch Perca fluviatilis [23], and burbot Lota lota [24]. Since these fishes only occur in low abundances and in certain areas of the upper Danube River, they can be ruled out as an explanation for the massive declines of native amphipod populations in this river section.

About one decade after the introduction of D. villosus, the round goby Neogobius melanostomus (Pallas 1814) arrived in the German section of the Danube River in 2004 [25], rapidly spread and displayed high levels of population differentiation [26, 27]. Within few years, this goby species strongly increased in abundance, currently contributing more than three quarters of abundance and about two thirds of biomass of fish in artificial rip-rap bank habitats [7, 8]. Its range expansion and population growth is still ongoing. Extensive samplings during the years 2010 to 2011 [7, 8] were no longer able to detect the formerly abundant native amphipod species, G. pulex and Gammarus roeseli, in the upper Danube River. Instead they only found the non-native D. villosus as the currently most abundant and widely distributed amphipod. Consequently, declines to extinction of native amphipods in the upper Danube River could have been influenced by the invasion of N. melanostomus, particularly since amphipods were identified as most important prey for this fish in different field surveys [7, 28, 29].

Since synergistic effects of multiple invasive species can potentially accelerate biodiversity loss and may enforce further homogenization of biological communities, the term “invasional meltdown” has been proposed for such interaction on an ecosystem scale [30]. However, to our knowledge, no study systematically analysed mechanisms and potential  sympatric impact of N. melanostomus and D. villosus on indigenous amphipods in experimental trials, to validate species interaction in native species displacement processes.

Invasional processes in general cannot be captured within a short time-frame and the assessment of possible negative consequences for a respective system can be rather complex. However, investigations of species interactions under controlled conditions have the potential to reveal mechanistic relationships that can support the holistic understanding of complex invasion processes.

In this study, selective predation of N. melanostomus on native and non-native amphipods was investigated in controlled experiments. The main goals of the present study were (i) to estimate predation preference of N. melanostomus towards one of the respective amphipod species, (ii) to determine potential interaction between invasive D. villosus and native G. pulex, and (iii) to test how potential interactions affect predation by N. melanostomus, as synergistic impact in native species displacement.

We hypothesized that (i) there is a feeding preference in N. melanostomus towards indigenous amphipod species explaining their massive declines in nature, and that (ii) D. villosus can increase the relative predation risk of G. pulex by interference competition and a more effective predator avoidance due to the sympatric origin with N. melanostomus.

Methods

Test species

Neogobius melanostomus were collected by electrofishing in the upper Danube River (Bad Abbach, Germany: N48°57′11.56″, E11°59′12.53″) under license number 31-7562. The collection of gobies was additionally approved by the local owner of the fisheries rights (Kreisfischereiverein Kelheim e.V.) and the fisheries authority (Fischereifachberatung Niederbayern). Specimens with a mean total length (L T) of 10.0 ± 2.0 cm were used since this size class is known to preferentially feed on amphipods [7]. After transportation to the laboratory, fish were acclimatised in aerated aquaria (100 × 40 × 50 cm, density of 20 to 30 fish/m2, each equipped with 10 clay tubes as shelter) for 1 week prior to the experiments. During the acclimatisation period fish were fed with commercial fish-food (trout chow, Skretting, Norway). Invasive D. villosus were collected in the upper Danube River near Kelheim, Germany (N48°54′56.61″, E11°51′43.80″). Since no G. pulex were found in the main channel of the Danube River in recent years, we collected them from a small headwater tributary of the Danube River system, the Moosach River (Freising, Germany: N48°23′38.83‘‘, E11°43′26.15″). Accordingly, the native amphipods were completely “naïve” to the other test species. After kick-sampling, amphipods were sorted into two size classes (by sieving with mesh sizes of 2 and 1.5 mm) and transported to the laboratory in an aerated cooler. The smaller size class (size class 1) had a mean L T ± SD of 11.8 ± 1.5 mm for G. pulex and 12.0 ± 1.4 mm for D. villosus, (measured on random samples of n = 30 per species and size class).The larger size class (size class 2) in both species had a mean L T ± SD of 14.8 ± 1.3 mm for G. pulex and 18.3 ± 2.4 mm for D. villosus. Both amphipod species were acclimatised to laboratory conditions in separated aerated aquaria (60 × 30 × 30 cm with coarse pebbles as substratum, resembling their natural shelter) for 24–48 h. During the acclimatisation period, amphipods were fed ad libitum with pre-conditioned black alder-leaves (Alnus glutinosa) and fish-food (trout chow, Skretting, Norway). Test species collection and experimental testing was conducted in August and September 2012.

Test conditions

Experimental trials were conducted under constant physicochemical conditions (mean ± SD; temperature (T)12.3 °C ± 0.5 °C, dissolved oxygen (DO) 8.9 ± 1.2 mg L−1, electric conductivity (EC, at 25 °C) 1125 ± 5 µS cm−1 using local well water. Ionic composition of the water is given in Table 1. Light conditions were 12:12 h dark:light. The test setup consisted of 14 aquaria (40 × 25 × 25 cm) individually supplied with a constant water flow(-through) of 0.8 L min−1. Adjacent aquaria were shielded from each other to avoid learning effects between the test organisms of different treatments and to avoid mutual disturbance. Each aquarium was filled with 1.5 L of pebbles (16–32 mm) as substratum (“substratum present”) or kept without substratum (“substratum absent”) and equipped with a clay tube providing shelter for the goby (open side facing back wall). All experimental trials were conducted consequently under daylight conditions to minimize variation due to diel changes in the organisms’ behaviour [13, 31]. The experiments were conducted according to German Tierschutzgesetz (§11 TierSchG), approved by the local veterinary board (Landratsamt Freising, license number 32-568) and the animal welfare committee at TUM.
Table 1

Water chemistry parameters

Parameter

Concentration [mg L−1]

Sodium (Na+)

42.3

Potassium (K+)

9.6

Calcium (Ca2+)

123

Magnesium (Mg2+)

41.6

Iron (Fe3+)

<0.1

Manganese (Mn2+)

<0.05

Chloride (Cl)

130

Bromide (Br)

<0.05

Hydrogen carbonate (HCO3 )

311

Sulfate (SO4 2−)

58

Nitrate (NO3 )

10.4

Ortho-phosphate (PO4 3−)

0.09

Ammonium (NH4 +)

<0.02

Dissolved organic carbon (DOC)

0.94

Total inorganic carbon (TIC)

70.6

Preliminary experiments

The optimum duration of the feeding trial was determined by preliminary experiments that were not included in the statistical analyses. Duration of 3 h was identified to be most suitable for the quantification of consumption rates (20.7 % ± 9 SD), based on the criteria that 100 % of the fish were feeding and gut contents were not fully digested to allow prey identification.

Amphipod mortality without predator

Several experimental trials were conducted to test for gammarid mortality without the presence of N. melanostomus caused by random mortality or intraguild predation (IGP, [32]). Each experimental trial consisted of 15 (single species trial) or 30 (mixed species trial) replicates, respectively. Each replicate was stocked with a total of 40 individuals, either 40 specimens of the same species (single species trial), or 20 specimens from both species (mixed species trial). In mixed species trials, two different experimental approaches were conducted: On the one hand, G. pulex and D. villosus of the same size class (size class 1 as described above) were used. In addition, D. villosus individuals of size class 2 vs. G. pulex of size class 1 were exposed to account for the different maximum sizes of the two species. Test duration of each trial was 3 h.

Predation experiments

Experimental trials were conducted in absence and in presence of substratum. Predation experimental trials consisted of 11 (single species trial) or 22 (mixed species trial) replicates without substratum, respectively. In experiments with substratum added, a higher number of replicates was used in the single species trials (G. pulex: 49; D. villosus: 14) due to the greater expected behavioural variability in these treatments with more complex habitat structure. Aquaria were stocked with 40 individuals of the same amphipod species (single species trial) or 20 each from both species (mixed species trial). In mixed species trials, either G. pulex and D. villosus of the same size class (size class 1) were used or size class 2 D. villosus individuals vs. size class 1 G. pulex to account for naturally expected body size-dependent effects. Predation experiments for different size classes were conducted as mixed species trial.

Gobies were not fed 24 h prior to the experiments. Amphipods transferred to the test aquaria were allowed to settle and shelter for 1 h before fish were added. Test duration of each trial was 3 h. Within this timeframe, fish exhibited natural behaviour and consumed amphipods, as expected from preliminary experiments. These short-time experiments allowed for an easy assignment of amphipod species identities and numbers in subsequent gut-content analyses, since prey was nearly undigested. After each feeding trial, fish were caught with a dip-net, euthanized and immediately frozen at −20 °C. The remaining living and dead amphipods in each aquarium were counted. Prior to dissection for stomach content analysis, L T of the fish (to the nearest 1 mm) and total body mass (W T to the nearest 0.001 g) were measured. Sex of the fish was determined using the morphology of the urogenital papilla according to Kornis et al. [13] before the experiments and later double-checked during dissection. Intestinal tracts (from the pharyngeal teeth to the anus) were removed, full and empty wet-weight was measured (nearest 0.001 g) to calculate the gut content mass (W G).

Amphipod swimming behaviour

Experimental trials to quantify species-specific swimming behaviour and potential predator avoidance behaviour in both amphipod species were performed within 30 min intervals. For each amphipod species, 40 individuals were tested per trial with five replicates each. Gobies (n = 10) were not fed 24 h prior to the experiments. Amphipods transferred to the test aquaria were allowed to settle and shelter for 1 h before start of the trial. A single trial consisted of 15 min without plus 15 min with a single N. melanostomus present in the aquarium. A photograph was taken every minute to enable standardised counting of amphipods swimming freely in the water column. Experimental trials were conducted with substratum provided in the setup. Since amphipod species could not be differentiated exactly when observing them from outside the tanks, this experiment was not conducted for a mixed species setup.

Data analysis

To quantitatively test for differences in the overall feeding between N. melanostomus sexes within the experimental timeframe, the index of stomach fullness (ISF) = 100 W G *W T −1 was calculated, providing a standardized and representative estimate of prey consumption [7]. The general assumption was tested that females have a higher feed-uptake compared with males, due to a higher energy demand and potential differences in W T due to variable fish fecundity during the reproductive state in the respective time of the year. Since also single fish and amphipod biomasses were not identical in the experiments, the use of ISF (instead of amphipod counts) to quantify prey consumption enabled higher statistical power, since extremes could be included. For the comparison of differences in feeding on different species and size classes of amphipods, consumption ratios were used as dataset.

Selectivity of goby predation was evaluated using Manly’s selectivity index α (discussed by Chesson [33]), calculated according to Eq. 1:
$$\alpha_{i} = \frac{{\ln (\frac{{n_{i0 } - r_{i} }}{{n_{i0} }})}}{{\mathop { \Sigma }\nolimits_{j = 1}^{m} \ln (n_{j0} - r_{j} )/n_{j0} }} ,\quad i = 1,2, \ldots , m$$
(1)
where n i0 is the initial number of prey species i, r i is the number of prey species i consumed by the goby and m is the number of species, which is two in this study.
In some cases all individuals of the preferred prey species were consumed in the trial, so calculation of α was not possible. To account for that, the approach by Klecka and Boukal [34] was applied, modifying the equation in single cases. Therefore one additional individual was added to the respective n i0 and n j0 , assuming this additional individual would have survived. This was the case for two trials in which all G. pulex were consumed, resulting in a slightly conservative estimate of α. Dead individuals were excluded from this calculation. Statistical analysis were performed on Manly’s alpha (α). Differences in selective feeding of the fish were tested by comparing calculated Manly’s α values against a hypothetical value of 0.5 (no selectivity) using Wilcoxon’s one sample test. For graphical presentation, the alpha values were converted in electivity indices as described in Chesson [33].
$$\varepsilon_{i} \,= \frac{{m\alpha_{i} - 1}}{{\left( {m - 2} \right)\alpha + 1}},\quad i = 1, \ldots ,m,$$
(2)

Electivity index ε can reach values between −1 and 1 per prey species, where positive values indicate a preferred prey and a value of 0 corresponding to unselective feeding.

All datasets were tested for normal distribution and homogeneity of variance using Shapiro-Wilks and Levene’s test, respectively. For comparison of ISF and sex-specific predation, pairwise t tests were used. As assumptions for parametric tests were not met for comparison of mortality rates in control trials without fish, nonparametric Kruskal–Wallis-test was used for multiple comparisons. For post hoc comparisons, Bonferroni-corrected Mann–Whitney-U tests were used. For comparison of amphipod consumption rate of gobies a factorial ANOVA design was chosen, using the factors species (G. pulex, D. villosus), substratum (absence, presence) and trial (single, mixed). Since the data was skewed, rank transformation of the data was used to overcome violation of ANOVA assumptions. We applied an aligned rank transformation procedure by Wobbrock et al. [35], which is suitable for factorial designs including interactions, since the same main effect, and interaction structure as the original data is retained. Analyses were conducted on partial datasets since including single species and mixed species trials in the same model was not possible, due to the non-independent data. Species were analysed separately using a between subject design with the factors substratum (absence, presence) and trial (single, mixed). Gammarus pulex and D. villosus were compared with each other separately in single species trials, using a between-subject-design with the factors species (G. pulex, D. villosus) and substratum (absence, presence). Subsequently, one-way ANOVAs were conducted to compare differences in single vs. mixed trials with and without substratum separately. Mixed species trials for both same and different size-class comparisons were tested using a within-subject-design (repeated measures ANOVA) and are presented separately for each size class comparison.

To test for differences in swimming activity and potential predator avoidance of amphipods, Wilcoxon signed-rank test was applied to compare numbers of free-swimming gammarids with and without fish present (average of 15 observations in 5 tanks). Statistical analyses were performed using the software SPSS 22 (IBM, USA). Significance was accepted at p < 0.05.

Results

Amphipod mortality without predator

No significant predation effects (either asymmetric predation by IGP or symmetric by cannibalism) between the tested amphipod species were observed in the trials with the same size class (size class 1) of amphipods, within the test duration of 3 h. Control trials without predator presence revealed similar mortalities in both G. pulex and D. villosus which remained on average below 2 % (Table 2A). No significant differences were found neither between trials with and without substratum (single trials, Mann–Whitney: G. pulex, z = −0.83, p > 0.05; D. villosus, z = −0.27, p > 0.05; mixed trials, Mann–Whitney: G. pulex, z = −1.76, p > 0.05; D. villosus, z = −2.46, p > 0.05), nor between single- or mixed-species trials (without substratum, Mann–Whitney: G. pulex, z = −1.78, D. villosus, z = −2.93, p > 0.05; with substratum, Mann–Whitney: G. pulex, z = −0.94, p > 0.05; D. villosus, z = −0.67, p > 0.05). Control trials with different size classes of the two amphipod species (e.g. G. pulex size class 1, D. villosus size class 2 resulted in different mortality rates in comparison to trials with equal size distribution of both species. Higher mortality rates for both species were recorded in the mixed species trial without the presence of substratum (Table 2B), without being statistically significant. Mortality was highest for G. pulex (2.5 %) when both amphipod species were held together in the absence of substratum. As above, no significant differences could be found neither between trials with and without substratum (single trials, Mann–Whitney: D. villosus, z = −0.18, p > 0.05; mixed trials, Mann–Whitney: G. pulex, z = −1.29, p > 0.05; D. villosus, z = −1.19, p > 0.05) nor between mixed- or single-species trials (without substratum, Mann–Whitney: G. pulex, z = −0.83, D. villosus, z = −0.08, p > 0.05; with substratum, Mann–Whitney: G. pulex, z = −1.39, p > 0.05; D. villosus, z = −1.56, p > 0.05).
Table 2

Relative comparison of amphipod losses found after 3 h under experimental conditions without N. melanostomus presence

 

Trial

Substratum

Mortality [%] mean (± SD)

N

G. pulex

D. villosus

A

Single

Presence

1.8 (±3.0)

1.2 (±2.5)

17

Mixed

Presence

1.3 (±2.7)

1.3 (±3.2)

24

Single

Absence

0.9 (±2.0)

1.0 (±2.2)

17

Mixed

Absence

0.3 (±1.3)

0.0 (±0.0)

29

B

Single

Presence

1.8 (±3.0)

1.7 (±2.1)

6

Mixed

Presence

0.7 (±1.8)

0.7 (±1.8)

16

Single

Absence

0.9 (±2.0)

1.3 (±1.4)

6

Mixed

Absence

2.5 (±4.1)

1.9 (±3.1)

15

A G. pulex and D. villosus with the same size class, B D. villosus larger than G. pulex. Single only one species per aquarium, Mixed both species in aquarium

Predation experiments

Sex-specific predation

Comparing the ISF (Fig. 1), no significant differences in prey consumption between female (mean ± SD, 4.3 ± 1.5) and male (3.3 ± 1.1) fish were observed in experiments, neither without (T(20) = 1.552, p = 0.136) nor with substratum (T(20) = 0.532, p = 0.600). Mean ISF ± SD ranged between 1.50 ± 1.1 for female and 1.52 ± 1.7 for male N. melanostomus in presence of substratum. In case of generally reduced availability of amphipods in the open water, i.e. when substratum was provided, ISF was about two times lower than in tests without substratum (Fig. 1). In female N. melanostomus, the ISF was significantly (T(24) = 5.259, p < 0.001) lower in treatments with substratum than in treatments without substratum. This finding was also significant (T(16) = 3.538, p = 0.003) in tests with male N. melanostomus.
Fig. 1

Influence of fish sex on gammarid predation depending on absence or presence of substratum. Index of stomach fullness (ISF) confirms a generally higher predation rate in absence of substratum. No differences between female (white) and male (grey) N. melanostomus were observed. [Asterisks indicate significant differences at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***)]. Boxplots represent 25 to 75 % (boxes ) and 5 to 95 % percentiles (whiskers). Circles represent outliers (exceeding 1.5 interquartile range). n = 14: female/substratum absent; n = 8: male/substratum absent; n = 12: female/substratum present; n = 10: male/substratum present)

Between-subject comparison Gammarus pulex

The main factor substratum had significant effects on the amount of G. pulex consumed by N. melanostomus (F(1,100) = 12.26, p = 0.001, Table 3). No significant effects of the main factor trial were observed, but a significant interaction between the factors substratum and trial was evident (F(1,100) = 6.03, p = 0.03). Separate examination of the datasets with and without substratum showed that, without substratum, no statistically significant difference was evident between single and mixed trials (Fig. 2). Neogobius melanostomus consumed 37 ± 14.7 (mean ± SD) and 58.2 ± 24.9 % of G. pulex, respectively. In contrast, a significant difference could be observed in the presence of substratum between single and mixed species trials (F(1, 69) = 22.05, p < 0.001). Gobies consumed 8.6 ± 12.8 (mean ± SD) percent of G. pulex in single trials and 27.3 ± 18.2 percent of G. pulex when held together with D. villosus.
Table 3

Comparison of predation rates of N. melanostomus on Gammarus pulex based on factorial ANOVA with the factors trial (single species vs. mixed species) and substratum (absence vs. presence)

Factor

SS

df

MS

F

p

Trial

 

7284.869

1

7284.869

2.565

0.112

 Error

283960.694

100

2839.607

  

Substratum

 

29647.329

1

29647.329

12.256

0.001

 Error

241893.087

100

2418.931

  

Interaction trial x substratum

 

20296.829

1

20296.829

6.031

0.016

 Error

336567.269

100

3365.673

  

Italic values indicate significance of p value (p < 0.05)

SS sum of squares, df degrees of freedom, MS mean square, F F value, p p value of single factors and their interactions, respectively

Fig. 2

Predation of N. melanostomus on amphipods within a 3 h time-period where either single species (white, n = 11) or a combination of both species (grey, n = 22) were tested. a Substratum absent. b Substratum present. [Asterisks indicate significant differences at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***)]. Boxplots represent 25 to 75 % (boxes) and 5 to 95 % percentiles (whiskers). Circles represent outliers (exceeding 1.5 interquartile range)

Between-subject comparison Dikerogammarus villosus

Similar to the results of G. pulex, strongest effects on D. villosus predation were observed for the main factor substratum (F(1,65) = 22.97, p < 0.001; Table 4). Also the main factor trial showed to be significant (F(1,65) = 6.88, p = 0.01), but no significant interaction between the main factors was observed. Separate examination of the datasets without and with substratum showed that, without substratum there were no significant differences in consumed D. villosus from single and mixed trials (Fig. 2). N. melanostomus consumed 9.5 ± 9.3 (mean ± SD) and 16.4 ± 11.7 percent of D. villosus, respectively. In the presence of substratum as shelter single and mixed trials were significantly different (F(1,34) = 5.12, p = 0.03). N. melanostomus consumed 2.3 ± 3.5 percent of D. villosus in single and 6.8 ± 7.6 in mixed trials.
Table 4

Comparison of predation rates of N. melanostomus on Dikerogammarus villosus based on factorial ANOVA with the factors trial (single species vs. mixed species) and substratum (absence vs. presence)

Factor

SS

df

MS

F

p

Trial

 

20544.067

1

20544.067

6.881

0.011

 Error

194070.532

65

2985.7

  

Substratum

 

61621.819

1

61621.819

22.974

<0.001

 Error

174342.766

65

2418.931

  

Interaction trial x substratum

 

5549.039

1

5549.039

1.982

0.164

 Error

181939.994

65

2799.077

  

Italic values indicate significance of p value (p < 0.05)

SS sum of squares, df degrees of freedom, MS mean square, F F value, p p value of single factors and their interactions, respectively

Between-subject comparison—single species trials (G. pulex vs. D. villosus)

The comparison of predation rates showed that both main factors species and substratum had strong effects of amphipod consumption by N. melanostomus (species: F(1,81) = 59.16, p < 0.001; substratum: F(1,81) = 48.91, p < 0.001; Table 5). The observed feeding pattern in N. melanostomus revealed a clear preference for G. pulex in the experimental trials with amphipods of the same size-class and a higher susceptibility of this native amphipod compared to non-native D. villosus to predation in the absence of substratum (Fig. 2). Also a significant interaction of the factors species and substratum was observed (F(1,81) = 7.14, p = 0.009).
Table 5

Comparison of predation rates of N. melanostomus for single species predation trials based on factorial ANOVA with the factors species (G. pulex vs. D. villosus) and substratum (absence vs. presence)

 Factor

SS

df

MS

F

p

Species

 

59950.019

1

59950.019

59.165

p < 0.001

 Error

82074.539

81

1013.266

  

Substratum

 

94868.898

1

94868.898

48.913

p < 0.001

 Error

157103.081

81

1939.544

  

Interaction species x substratum

 

12026.53

1

12026.53

7.145

0.009

 Error

136349.47

81

1683.327

  

Italic values indicate significance of p value (p < 0.05)

SS sum of squares, df degrees of freedom, MS mean square, F F value, p p value of single factors and their interactions, respectively

Separate examination of the datasets with and without substratum showed significant differences between consumed G. pulex and D. villosus in the absence of substratum (F(1,20) = 12.36, p = 0.002). This was even more pronounced in the presence of substratum (F(1,61) = 60.31, p < 0.001). When G. pulex was the only food source available, each goby consumed 15 ± 6 (mean ± SD) amphipod specimens per trial (without substratum) and 4 ± 5 specimens (with substratum present), respectively. If D. villosus was the only prey species available, 4 ± 4 (mean ± SD) individuals were consumed without and 1 ± 1 with substratum present. In the latter case, several gobies (21 %) with empty guts were recorded.

Within-subject comparison—mixed species trials (G. pulex vs. D. villosus)

In the mixed species trials, i.e. when both amphipod species were available as prey, round gobies clearly preferred G. pulex as prey (Fig. 2). This effect was even more pronounced in the presence of substratum, likely due to biological interactions between the both amphipod species within the shelter (Fig. 2). Predation rates were significantly higher on G. pulex (percent consumed: median 47.5, mean 42.7, SD 26.6) compared to D. villosus (percent consumed: median 10, mean 11.6, SD 10.9), F(1,43) = 31.34, p < 0.001, with an effect size of 0.42. Separate examination of the datasets with and without substratum showed that without substratum, in 95.4 % cases more G. pulex were consumed and there was a significant higher predation on G. pulex (percent consumed: median 60, mean 58.2, SD 24.9) compared to D. villosus (percent consumed: median 15, mean 16.4, SD 11.7), F(1,21) = 9.5, p = 0.006, with an effect size 0.31). With substratum, in 90.9 % cases more G. pulex were consumed and in no case more D. villosus were consumed. In 9.1 % of the cases, no difference was found. With substratum, there was a significant higher predation on G. pulex (percent consumed: median 27.5, mean 27.3, SD 18.2) compared to D. villosus (percent consumed: median 5, mean 6.8, SD 7.6), F(1,21) = 35.24, p < 0.001, and the difference was large (effect size 0.63).

Size effects

In the mixed trials with larger D. villosus (size class 2) vs. smaller G. pulex (size class 1), similar trends were observed as with equally sized amphipods (Fig. 3). Significantly more G. pulex were consumed as compared to D. villosus, regardless if substratum was absent (F(1,10) = 8.44, p = 0.16, effect size 0.46) or present (F(1,11) = 24.68, p < 0.001, effect size 0.69). Again, the availability of substratum as shelter resulted in a stronger effect on predation. The fish consumed more G. pulex (percent consumed: median 40, mean 38.2, SD 16.9) than D. villosus (percent consumed: median 10, mean 16.8, SD 14.7) in 83.3 % of the cases in the absence of shelter. With substratum provided, more G. pulex (percent consumed: median 32.5, mean 26.2, SD 14.8) than D. villosus (percent consumed: median 5, mean 8.7, SD 8.8) were consumed in 90.9 % of the cases.
Fig. 3

Predation of N. melanostomus within a 3 h time-period on G. pulex (grey) and D. villosus (white). Gammarus pulex mean size (± SD) 11.8 ± 1.5 mm and D. villosus 18.3 ± 2.4 mm. [Asterisks indicate significant differences at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***)]. Boxplots represent 25 to 75 % (boxes) and 5 to 95 % percentiles (whiskers). Circles represent outliers (exceeding 1.5 interquartile range); n = 12 per substratum trial)

Selectivity index

Selectivity analyses using Manly’s alpha confirmed a clear preference for G. pulex as prey item in mixed species trials (Fig. 4). Alpha values above 0.5 indicate a preference for G. pulex and were highest in trials without substratum added and the same size-class of the two amphipod species (0.83 ± 0.18, mean ± SD) compared to trials with presence of substratum (0.81 ± 0.23, mean ± SD). Selectivity was significant for G. pulex as preferred prey item (without substratum: z = 3.94, p < 0.001; with substratum z = 3.49, p < 0.001). Interestingly, no significant preference was evident from the selectivity index observed for in the different size-class trials, but higher values in the presence of substratum (0.65 ± 0.37, mean ± SD) were observed compared to trials without substratum (0.50 ± 0.38, mean ± SD).
Fig. 4

Prey selectivity of N. melanostomus. Mean values ± SE of electivity index are presented. Positive values represent preferred prey species in presence (open circles and squares) or absence (black circles and squares) of substratum. Dashed line indicates no selectivity. a Same size class of G. pulex and D. villosus. b Larger size class of D. villosus. Asterisks indicate significant differences to zero at p < 0.05 (repeated measures ANOVA): (*), p < 0.01 (**), p < 0.001 (***)

Amphipod swimming behaviour

Generally, behaviour of the amphipods differed significantly between species: Gammarus pulex was about three to four times more active in terms of individuals swimming freely in the water-column during the observation period (Fig. 5). After adding the predator, there was a significant reduction in freely swimming G. pulex (mean 2.53, SD 1.67 per min) compared to controls without gobies (mean 0.46, SD 0.27 per min, n = 5), z = −2.02, p = 0.042, and the difference was large (effect size r = −0.90).
Fig. 5

Mean number (± SD) of individual amphipods swimming actively in the water column within a timeframe of 15 min in absence (grey) and 15 min in presence (black) of a single N. melanostomus specimen. (n = 5 per species)

Dikerogammarus villosus showed a generally low activity, even in the control group (mean 0.32, SD 0.05 per min, n = 5). After addition of the predator, this effect became more pronounced (mean 0.01, SD 0.03). Wilcoxon comparison revealed a significant difference in activity patterns caused by the presence of the fish (z = −2.03, p = 0.042, effect size r = −0.91). Pairwise comparison between species identified significant differences between the control groups (z = −2.63, p = 0.009, effect size r = −0.83) and in the presence of the fish (z = −2.12, p = 0.034, effect size r = −0.67).

Discussion

In several European waterbodies, the arrival of non-native D. villosus has simultaneous occurred as the disappearance of native amphipods [16, 36, 37] which is often explained by asymmetric mutual predation [24]. A sympatric and synergistic impact by invasive D. villosus and invasive N. melanostomus has not yet been considered, but appears likely based on the results of our study.

In our experiments, D. villosus faced a lower predation risk against N. melanostomus due to a relatively lower swimming activity as compared to G. pulex. These findings match obervations in other species pairs such as those by Kinzler and Maier [23] who detected similar results for rainbow trout and European perch. According to van Riel et al. [38], the presence of D. villosus can lead to an increased number of G. pulex swimming freely in the water column, indicating interference competition. Analogously, we observed a threefold higher predation of N. melanostomus on G. pulex in the presence of D. villosus when shelter was available compared to a single exposure of G. pulex. Since substrate structures are often heavily modified in natural habitats for bank erosion protection, this may directly affect predation risk of native gammarids, particularly in anthropogenically modified habitat structures.

The role of IGP

Dikerogammarus villosus is able to prey on other amphipod species without being restricted to vulnerable post-molting stages [36]. On the other hand, the mouthparts of D. villosus are not highly specialized just for carnivory and predation, and the species has been described to use a wide spectrum of different food sources [39, 40]. Also gut contents and stable isotope analyses by Koester and Gergs [41] did not provide evidence for D. villosus being a carnivorous “killer shrimp”. In our study, no significant asymmetric mutual predation between both amphipod species could be confirmed within the short experimental duration. Gammarids are also known to naturally occur at similar densities and the selected experimental setup allowed for isolated analysis of the effects caused by N. melanostomus predation. Based upon the results of our experiments, intraguild predation seemingly only plays a minor role and cannot explain the invasion success of D. villosus outside of its natural distribution range. In contrast, interactions with D. villosus seem to facilitate excessive consumption of the native gammarid by the newly introduced top-predator N. melanostomus and appears to play a more important role in the decline of autochthonous amphipods.

Structural diversity of invaded habitats can mitigate potential impact by spatial separation due to different habitat preferences, since it uncouples competition between native and invasive species [42]. Especially in heavily modified waterbodies such as shipping canals, structural diversity is much lower compared to natural aquatic systems. Technolithal and other monotonic structures within these waterbodies can thus force the respective species towards direct interaction and may consequently increase competitive stress in autochthonous, less generalized species.

Thus, main behavioural traits acting in combination can provide a plausible explanation for the observed eradication of G. pulex in natural habitats invaded by D. villosus and N. melanostomus. In case of N. melanostomus and G. pulex, preference for one species as a superior food item accompanied by higher susceptibility to predation by different activity patterns in swimming behaviour interacts synergistically and speeds up this process.

Comparing predation of N. melanostomus against G. fasciatus, an autochthonous amphipod species native to North America, and its Ponto-Caspian invasive counterpart E. ischnus, Kestrup and Ricciardi [43] found that competition with E. ischnus did not increase the vulnerability of G. fasciatus to goby predation. In their experiments E. ischnus and G. fasciatus appeared to be equally susceptible to goby predation and no preference in feeding of N. melanostomus was observed. Kestrup and Ricciardi [43] finally concluded that N. melanostomus did not influence the replacement of G. fasciatus by E. ischnus in the St. Lawrence River. Remarkably, outside laboratory conditions in Lake Erie and Lake Ontario, where N. melanostomus continues to spread, E. ischnus has replaced G. fasciatus as the dominant amphipod on substrates fouled by dreissenids [44]. These amphipods are mutual predators, too, and dominance of competing amphipods varies with conductivity [45]. Thus, IGP can also be influenced by environmental heterogeneity.

Since N. melanostomus used in Kestrup and Ricciardi [43] had a significantly lower (mean = 6.6 cm ± 0.1 SE) L T compared with our experiments, the experimental setup chosen might have led to misinterpretation: Due to a significantly smaller gape limitation, N. melanostomus of this size class preliminary feeds on smaller prey than adult amphipods, such as chelicorophiids, isopods, zooplankton and juvenile amphipods [7]. At a size of about 10–12 cm, depending on time since invasion, N. melanostomus is known to mainly feed on amphipods correlating with an ontogenetic diet switch [8]. Thus, in our experiments N. melanostomus and amphipods were size-selected with expected maximum impact.

Selective feeding by fish or different vulnerability

Among others, predation by fish is an important factor influencing the composition of amphipod communities. Neogobius melanostomus has a rather opportunistic feeding strategy, using the most available prey as compared to other more specialized goby species such as P. kessleri [7, 29]. However, N. melanostomus shows a clear preference for amphipods over molluscs at early stages of the invasion process when their availability is not limited [8]. Thus, according to our experiments, amphipods appear to be a superior and favourite prey for both sexes, likely facilitating the invasion success of invasive gobies [7, 28, 46]. The choice of amphipods as important prey item in our study appears legitimate since the proportion of amphipods in N. melanostomus prey is highest in all seasons: about two thirds to three quarters to the total food uptake and amphipods were consumed in higher proportions as compared to their availability in environment [7]. Other potentially preferred prey items such as Chironomid larvae have limited availability due to emergence of imagos in summer and autumn. Amphipods can percept predatory fish or injured conspecifics by chemical stimuli and are thus able to react by predator avoidance behaviour (co-evolutionary adaptation) such as reducing time in open water [47], or change of habitat use [48]. Behavioural responses can even be flexible when complex microhabitat structures are available [49]. Since amphipods are known to possess distinct substrate preferences, the presence of preferred microhabitats can mitigate their predation risk [24]. Dikerogammarus villosus preferred gravel with a low percentage of sand and stones, whereas G. pulex showed no clear substratum preference in a substratum choice experiment [38]. However, in line with our study (see Fig. 5), in presence of invasive D. villosus, native G. pulex can be found significantly less frequent in gravel [50, 51] independent of D. villosus abundance [38, 41]. Such an exclusion from shelter-providing habitats can lead to increased swimming activity in G. pulex and consequently to an increased exposure to predatory fish. It might therefore play an important role in the displacement of native amphipods by D. villosus [51]. Besides the feeding behaviour, high reproductive potential and rapid growth of D. villosus [16, 52, 53] make this species a strong competitor not only for food but also for habitat space [41]. In our study, both sexes of invasive N. melanostomus asymmetrically consumed significantly more native G. pulex than invasive D. villosus, regardless if there was substratum present or not. Calculations of a preferential prey choice index [34] corroborated these findings. On a first glance, these findings are in contrast to Kley et al. [24], who considered that differences in use of spatial niches can permit the co-existence of D. villosus and G. roeseli in the wild when substrates are diverse. However in their tests, it was autochthonous burbot (Lota lota) that did not affect substrate choice or predation risk of G. roeseli in presence of D. villosus. It is probably due to the smaller prey selectivity of L. lota compared to the invasive N. melanostomus which can explain these differences. For freshwater amphipods, a predator avoiding mechanism would be to respond to an open water predator by reducing activity and moving towards the sediment, whereas a benthic predator would generally be better avoided by an active escapement through swimming or increased drift escape when predator abundance is low [47, 49]. In case of high or very high numbers of predators, most likely after mass development of invasive alien species, hiding seems to be a better survival strategy. Proportion of invasive N. melanostomus to the total fish fauna can be extremely high, particularly in man-made rip rap habitats where they have been reported to contribute two thirds of all fish counts and about 50 % of total fish biomass [9]. Thus, drift might not work as a perfect escape mechanism. Instead, escaping by drift likely leads to increased predation when predator abundance is high and thus might favour extinction in invaded ecosystems. This theory matches actual field observations from the Danube River, where the breakdown of G. roeseli in sections invaded by non-native gobies is reported [7]. Considering these principles, unreduced activity seems to be mostly responsible for higher losses in native G. pulex, whereas D. villosus appears to be generally less active, consequently being less vulnerable to the benthic ambush predator N. melanostomus. According to foraging theories, it therefore appears to be more energy efficient for N. melanostomus to feed on native G. pulex as an “easy” prey.

The results from our experiments provide a mechanistic explanation for higher feeding rates on G. pulex by N. melanostomus in absence of suitable substratum which in pristine riverine environments could provide shelter against predation. The observed higher consumption of G. pulex by N. melanostomus indicates species-specific vulnerability due to gammarid acitivity. Both species similarly respond to a predator, but D. villosus is less vulnerable due to its overall lower activity. Consequently, this study provides evidence that a synergistic impact of invasive species with distinct foraging strategies can be greater than the independent effects of the single species. On an ecosystem scale such added effects can pose an important factor in the context of an “invasional meltdown” scenario as defined by Simberloff and Von Holle [30]. Such an invasional meltdown supposedly still seems to occur in the upper Danube River to date [8].

Conclusions

Behavioural traits and interactions between closely related species are important basic mechanisms in understanding species displacement by IAS. The results of this study suggest that both the foraging selectivity of invasive N. melanostomus, as well as the behavioural interaction between D. villosus and G. pulex together result in an increased predation risk for the native gammarid. Previous invasions of D. villosus thus likely facilitate the rapid range expansion of N. melanostomus, in turn increasing predation on autochthonous amphipods. The results of this study are thus in line with processes described as “invasional meltdown” on the scale of ecosystems. Not only competition for food and habitat resources as well as asymmetric mutual predation between the closely related (amphipod) species can lead to declines of less competitive autochthonous species, but also becoming an easy prey for a new predator enabling excessive resource consumption may lead to the extinction of autochthonous species. Sympatrically, both invasive Ponto-Caspian IAS possess the ability to effectively restructure food-web composition by mutually and synergistically facilitating their invasive potentials, thus promoting further range expansion of both species.

Notes

Abbreviations

ANOVA: 

analysis of variance

DO: 

dissolved oxygen

EC: 

electric conductivity

IAS: 

invasive alien species

IGP: 

intraguild predation

ISF: 

index of stomach fullness

L T

total length

SE: 

standard error

SD: 

standard deviation

WT

total body weight

W G

gut content weight

Declarations

Authors’ contributions

The study was conceived and designed by JG and JB. Field sampling of gobies and gammarids, as well as the laboratory work was conducted by SB, JB, and AFC. Data analyses was mostly conducted by SB, with continuous revision by JG and JB. SB, JG and JB jointly interpreted the data and drafted the manuscript which was finalized by SB and JG. All authors read and approved the final manuscript.

Acknowledgements

The authors like to express their special thanks to Lydia Schübel and Heidrun Kerber for their laboratory assistance. This project was partly funded by the Deutsche Forschungsgemeinschaft (DFG), project-numbers GE 2169/1-1(AOBJ: 569812) and DFG SCHL567/5-1.

Competing interests

The authors declare that they have no competing interests.

Animal ethics

This study was conducted in compliance with the international animal care guidelines of the Association for the Study of Animal Behaviour [55] and the ARRIVE guidelines [56]. Gobies were collected by electrofishing in the Danube under license number 31-7562. The collection was additionally approved by the local owner of the fisheries rights (Kreisfischereiverein Kelheim e.V.) and the state fisheries authority (Fischereifachberatung Niederbayern). Laboratory experiments were conducted according to German legislation (German Tierschutzgesetz, §11 TierSchG), license number 32-568), approved by the local veterinary board (Landratsamt Freising, license number 32-568) and the animal welfare committee at TUM.

Availability of supporting data

The data set supporting the results of this article is available in the Dryad Digital Repository [54], http://dx.doi.org/10.5061/dryad.c5m0r.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
Aquatic Systems Biology Unit, School of Life Sciences Weihenstephan, Technical University of Munich
(2)
Wasserwirtschaftsamt Regensburg

References

  1. Dudgeon D, Arthington AH, Gessner MO, Kawabata Z, Knowler DJ, Lévêque C, Naiman RJ, Prieur-Richard A, Soto D. Stiassny. freshwater biodiversity: importance, threats, status and conservation challenges. Biol Rev. 2006;8:163–82.View ArticleGoogle Scholar
  2. Keller RP, Geist J, Jeschke JM, Kühn I. Invasive species in Europe: ecology, status and policy. Environ Sci Eur. 2011;23(23):1–17.Google Scholar
  3. Wolter C, Röhr F. Distribution history of non-native freshwater fish species in Germany: how invasive are they? J Appl Ichthyol. 2010;26(Suppl. 2):19–27.View ArticleGoogle Scholar
  4. Sindilariu PD, Freyhof J, Wolter C. Habitat use of juvenile fish in the lower Danube and the Danube Delta: implications for ecotone connectivity. Hydrobiologia. 2006;571:51–61.View ArticleGoogle Scholar
  5. Bijdevaate A, Jazdzewski K, Ketelaars HAM, Gollasch S, Vander Velde G. Geographical patterns in range extension of Ponto-Caspian macroinvertebrate species in Europe. Can J Fish Aquat Sci. 2002;59:1159–74.View ArticleGoogle Scholar
  6. Haertl M, Cerwenka AF, Brandner J, Borcherding J, Geist J, Schliewen UK, First record of Babka gymnotrachelus from Germany (Teleostei, Gobiidae, Benthophilinae). Spixiana. 2012;1857(35):155–9.Google Scholar
  7. Brandner J, Auerswald K, Cerwenka AF, Schliewen U, Geist J. Comparative feeding ecology of invasive Ponto-Caspian gobies. Hydrobiologia. 2013;703:113–31.View ArticleGoogle Scholar
  8. Brandner J, Cerwenka AF, Schliewen UK, Geist J. Bigger is better: characteristics of round gobies forming an invasion front in the Danube River. PLoS ONE. 2013;8(9):e73036.View ArticlePubMedPubMed CentralGoogle Scholar
  9. Brandner J, Pander J, Mueller M, Cerwenka A, Geist J. Effects of sampling techniques on population assessment of invasive round goby. J Fish Biol. 2013;82:2063–79.View ArticlePubMedGoogle Scholar
  10. Van Riel MC, Vander Velde G, Rajagopal S, Marguillier S, Dehairs F, Bijdevaate A. Trophic relationships in the Rhine food web during invasion and after establishment of the Ponto-Caspian invader Dikerogammarus villosus. Hydrobiologia. 2006;565:39–58.View ArticleGoogle Scholar
  11. Borcherding J, Staas S, Krüger S, Ondračková M, Šlapanský L, Jurajda P. Non-native Gobiid species in the lower River Rhine (Germany): recent range extensions and densities. J Appl Ichthyol. 2011;27:1–3.Google Scholar
  12. Strayer DL. Twenty years of zebra mussels: lessons from the mollusk that made headlines. Front Ecol Environ. 2009;7:135–41.View ArticleGoogle Scholar
  13. Kornis MS, Mercado-Silva N, Vander Zanden MJ. Twenty years of invasion: a review of round goby Neogobius melanostomus biology, spread and ecological implications. J Fish Biol. 2012;80(2):235–85.View ArticlePubMedGoogle Scholar
  14. Ricciardi A, MacIsaac HJ. Recent mass invasion of the North American Great Lakes by Ponto-Caspian species. Trends Ecol Evol. 2000;15:62–5.View ArticlePubMedGoogle Scholar
  15. Dick JTA, Platvoet D, Kelly DW. Predatory impact of the freshwater invader Dikerogammarus villosus (Crustacea: Amphipoda). Can J Fish Aquat Sci. 2002;59:1078–84.View ArticleGoogle Scholar
  16. Kley A, Maier G. Reproductive characteristics of invasive gammarids in the Rhine-Main-Danube catchment, South Germany. Limnologica. 2006;36:79–90.View ArticleGoogle Scholar
  17. Bollache L, Dick JTA, Farnsworth KD, Montgomery WI. Comparison of the functional responses of invasive and native amphipods. Biol Lett. 2008;4:166–9.View ArticlePubMedGoogle Scholar
  18. Bacela-Spychalska K, Van der Velde G. There is more than one ‘killer shrimp’: trophic positions and predatory abilities of invasive amphipods of Ponto-Caspian origin. Freshw Biol. 2013;58:730–41.View ArticleGoogle Scholar
  19. Tittizer T, Schöll F, Banning M, Haybach A, Schleuter M. Aquatische Neozoen im Makrozoobenthos der Binnenwasserstraßen Deutschlands. Lauterbornia. 2000;39:1–172.Google Scholar
  20. Haas G, Brunke M, Streit B. Fast turnover in dominance of exotic species in the Rhine River determines biodiversity and ecosystem function: an affair between amphipods and mussels. In: Leppäkoski E, Gollasch S, Olenin S, editors. Invasive Aquatic species of Europe. Distribution, impacts and management. Dordrecht: Kluwer Academic Publishers; 2002. p. 426–32.View ArticleGoogle Scholar
  21. Poeckl M. Success of the invasive Ponto-Caspian amphipod Dikerogammarus villosus by life history traits and reproductive capacity. Biol Invasions. 2009;11:2021–41.View ArticleGoogle Scholar
  22. Dodd JA, Dick JTA, Alexander ME, Macneil C, Dunn AM, Aldridge DC. Predicting the ecological impacts of a new freshwater invader: functional responses and prey selectivity of the ‘killer shrimp’, Dikerogammarus villosus, compared to the native Gammarus pulex. Freshw Biol. 2014;59:337–52.View ArticleGoogle Scholar
  23. Kinzler W, Maier G. Selective predation by fish: a further reason for the decline of native gammarids in the presence of invasives? J Limnol. 2006;65:27–34.View ArticleGoogle Scholar
  24. Kley A, Kinzler W, Schank Y, Mayer G, Waloszek D, Maier G. Influence of substrate preference and complexity on co-existence of two non-native gammarideans (Crustacea: Amphipoda). Aquat Ecol. 2009;43:1047–59.View ArticleGoogle Scholar
  25. Paintner S, Seifert K. First record of the round goby, Neogobius melanostomus (Gobiidae), in the German Danube. Lauterbornia. 2006;58:101–7.Google Scholar
  26. Cerwenka AF, Alibert P, Brandner J, Geist J, Schliewen UK. Phenotypic differentiation of Ponto-Caspian gobies during a contemporary invasion of the upper Danube River. Hydrobiologia. 2014;721:269–84.View ArticleGoogle Scholar
  27. Cerwenka AF, Brandner J, Geist J, Schliewen UK. Strong versus weak population genetic differentiation after a recent invasion of gobiid fishes (Neogobius melanostomus and Ponticola kessleri) in the upper Danube. Aquat Invasions. 2014;9:71–86.View ArticleGoogle Scholar
  28. Borcherding J, Dolina M, Heermann L, Knutzen P, Krüger S, Matern S, van Treeck R, Gertzen S. Feeding and niche differentiation in three invasive gobies in the Lower Rhine. Germany. Limnologica. 2012;43:49–58.View ArticleGoogle Scholar
  29. Brandner J, Auerswald K, Schaeufele R, Cerwenka AF, Geist J. Isotope evidence for preferential dispersal of fast-spreading invasive gobies along man-made river bank structures. Isot Environ Healt S. 2015;51(1):80–92.View ArticleGoogle Scholar
  30. Simberloff D, Von Holle B. Positive interactions of nonindigenous species: invasional meltdown? Biol Invasions. 1999;1:21–32.View ArticleGoogle Scholar
  31. Elliott JM. Day-night changes in the spatial distribution and habitat preferences of freshwater shrimps, Gammarus pulex, in a stony stream. Freshw Biol. 2005;50:552–66.View ArticleGoogle Scholar
  32. Polis GA, Myers CA, Holt RD. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annu Rev Ecol Syst. 1089;20:297–330.View ArticleGoogle Scholar
  33. Chesson J. The estimation and analysis of preference and its relationship to foraging models. Ecology. 1983;64(4):1297–304.View ArticleGoogle Scholar
  34. Klecka J, Boukal DS. Who eats whom in a pool? A comparative study of prey selectivity by predatory aquatic insects. PLoS ONE. 2012;6(7):e37741.View ArticleGoogle Scholar
  35. Wobbrock JO, Findlater L, Gergle D, Higgins JJ. The aligned rank transform for nonparametric factorial analyses using only ANOVA procedures. Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI ‘11). Vancouver, British Columbia (May 7-12, 2011). New York: ACM Press; 2011: 143–46.Google Scholar
  36. Dick JTA, Platvoet D. Invading predatory crustacean Dikerogammarus villosus eliminates both native and exotic species. P Roy Soc B. 2000;267:977–83.View ArticleGoogle Scholar
  37. Kley A, Maier G. Life history characteristics of the invasive freshwater gammarids Dikerogammarus villosus and Echinogammarus ischnus in the river Main and the Main-Donau canal. Arch Hydrobiol. 2003;156:473–81.View ArticleGoogle Scholar
  38. Vanriel M, Healy EP, vander Velde G, Bijdevaate A. Interference competition among native and invader amphipods. Acta Oecol. 2007;31:282–9.View ArticleGoogle Scholar
  39. Mayer G, Maier G, Maas A, Waloszek D. Mouthparts of the Ponto-Caspian invader Dikerogammarus villosus (Amphipoda: Pontogammaridae). J Crustacean Biol. 2008;28:1–15.View ArticleGoogle Scholar
  40. Boeker C, Geist J. Effects of invasive and indigenous amphipods on physico-chemical and microbial properties in freshwater substrates. Aquat Ecol. 2015. doi:https://doi.org/10.1007/s10452-015-9539-y.Google Scholar
  41. Koester M, Gergs R. No evidence for intraguild predation of Dikerogammarus villosus (Sowinsky, 1894) at an invasion front in the Untere Lorze, Switzerland. Aquat Invasions. 2014;9:489–97.View ArticleGoogle Scholar
  42. Harrison SSC, Bradley DC, Harris IT. Uncoupling Strong predator-prey interactions in streams: the role of marginal macrophytes. Oikos. 2005;108:433–48.View ArticleGoogle Scholar
  43. Kestrup Å, Ricciardi R. Are interactions among Ponto-Caspian invaders driving amphipod species replacement in the St. Lawrence River? J Great Lakes Res. 2009;35:392–8.View ArticleGoogle Scholar
  44. Dermott R, Witt J, Um EM, Gonzalez M. Distribution of the Ponto-Caspian amphipod Echinogammarus ischnus in the Great Lakes and replacement of native Gammarus fasciatus. J Great Lakes Res. 1998;24:442–52.View ArticleGoogle Scholar
  45. Kestrup Å, Ricciardi R. Environmental heterogeneity limits the local dominance of an invasive freshwater crustacean. Biol Invasions. 2009;11:2095–105.View ArticleGoogle Scholar
  46. Polačik M, Janáč M, Jurajda P, Adámek Z, Ondračková M, Trichkova T, Vassilev M. Invasive gobies in the Danube: invasion success facilitated by availability and selection of superior food resources. Ecol Freshw Fish. 2009;18:640–9.View ArticleGoogle Scholar
  47. Wudkevich K, Wisenden BD, Chivers DP, Smith RJF. Reactions of Gammarus lacustris to chemical stimuli from natural predators and injured conspecifics. J Chem Ecol. 1997;23:1163–73.View ArticleGoogle Scholar
  48. Dahl J, Greenberg L. Effects of habitat structure on habitat use by Gammarus pulex in artificial streams. Freshw Biol. 1996;36:487–95.View ArticleGoogle Scholar
  49. Baumgärtner D, Koch U, Rothhaupt KO. Alteration of kairomone-induced antipredator response of the freshwater amphipod Gammarus roeseli by sediment type. J Chem Ecol. 2003;29(6):1391–401.View ArticlePubMedGoogle Scholar
  50. MacNeil C, Platvoet D. The predatory impact of the freshwater invader Dikerogammarus villosus on native Gammarus pulex (Crustacea: Amphipoda); influences of differential microdistribution and food resources. J Zool. 2005;267:31–8.View ArticleGoogle Scholar
  51. Boets P, Lock K, Messiaen M, Goethals PLM. Combining data-driven methods and lab studies to analyse the ecology of Dikerogammarus villosus. Ecol Inform. 2010;5:133–9.View ArticleGoogle Scholar
  52. Grabowski M, Bacela K, Konopacka A. How to be an invasive gammarid (Amphipoda: Gammaroidea)—comparison of life history traits. Hydrobiologia. 2007;590:75–84.View ArticleGoogle Scholar
  53. Poeckl M. Strategies of a successful new invader in European fresh waters: fecundity and reproductive potential of the Ponto-Caspian amphipod Dikerogammarus villosus in the Austrian Danube, compared with the indigenous Gammarus fossarum and G. roeseli. Freshw Biol. 2007;52:50–63.View ArticleGoogle Scholar
  54. Beggel S, Brandner J, Cerwenka A, Geist J. Data from: synergistic impacts by an invasive amphipod and an invasive fish explain native gammarid extinction. Dryad Digital Repository. http://dx.doi.org/10.5061/dryad.c5m0r.
  55. Guidelines for the treatment of animals in behavioural research and teaching. Anim Behav. 2012; 83:301–09. doi:https://doi.org/10.1016/j.anbehav.2011.10.031.
  56. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 2010;8:e1000412. doi:https://doi.org/10.1371/journal.pbio.1000412.View ArticlePubMedPubMed CentralGoogle Scholar

Copyright

© The Author(s) 2016