INTRODUCTION
A central topic in invasion ecology is why certain species are successful invaders. This involves both: (1) the intrinsic biological traits of the invasive species in question; and (2) extrinsic aspects of the invaded habitat, defined by an array of environmental and biotic parameters. The interaction between these intrinsic and extrinsic factors ultimately decides the success of a potentially introduced species. Traits that confer invasiveness have been identified in some taxonomic groups (e.g. marine macroalgae: Nyberg & Wallentinus, Reference Nyberg and Wallentinus2005; plants: Finnoff & Tschirhart, Reference Finnoff and Tschirhart2005; freshwater fish: Wonham et al., Reference Wonham, Carlton, Ruiz and Smith2000; Vila-Gispert et al., Reference Vila-Gispert, Alcaraz and a-Berthou2005). For example, the dispersal characteristics of many successful invaders, particularly in relation to anthropogenic vectors are similar and predictable (Nyberg & Wallentinus, Reference Nyberg and Wallentinus2005).
The biotic pressures of predation, competition, disease and parasitism are particularly important for determining invasion success (Torchin et al., Reference Torchin, Lafferty and Kuris2001; Wolfe, Reference Wolfe2002; Parker & Hay, Reference Parker and Hay2005; Parker et al., Reference Parker, Burkepile and Hay2006). If the overall biotic pressure on a newly introduced species is less than in the native range, this may facilitate the invasion (Elton, Reference Elton1958; Parker & Hay, Reference Parker and Hay2005; Parker et al., Reference Parker, Burkepile and Hay2006). The absence of coevolved specialist enemies and the preferential consumption of native species by native generalists could give introduced species a competitive advantage (Parker & Hay, Reference Parker and Hay2005). Alternatively, non-native prey might be unable to deter native predators in the introduced range due to the absence of co-evolved defences, as in the increased susceptibility/biotic resistance hypothesis (Elton, Reference Elton1958; Hokkanen & Pimentel, Reference Hokkanen and Pimentel1989; Colautti et al., Reference Colautti, Ricciardi, Grigorovich and MacIsaac2004; Parker et al., Reference Parker, Burkepile and Hay2006).
The relative contributions of intrinsic and extrinsic factors to the success of a particular invasion have rarely been studied. De Rivera et al. (Reference De Rivera, Ruiz, Hines and Jivoff2005) found that the distribution in North America of the introduced European green crab Carcinus maenas (Linnaeus) was partly determined by geographical variation in biotic pressures. Along a geographical gradient, the intrinsic invasive traits were eventually outweighed by predation from the native crab Callinectes sapidus (Rathbun). For invasive species, extrinsic factors can be compared in the native and introduced ranges (Wolfe, Reference Wolfe2002), but it is difficult to quantitatively compare ecological processes that are similar but involve completely different sets of interacting species. An alternative approach is to contrast the biotic pressures experienced by introduced species and sympatric native species in the same habitat, potentially allowing predictive assessments of likely success within other areas.
The brown seaweed Sargassum muticum (Yendo) Fensholt is native to Japan and China, and has become a conspicuous and successful invasive species along vast stretches of European and North Pacific shorelines (Wallentinus, Reference Wallentinus1999). Sargassum muticum has many of the intrinsic traits of an invasive species (Nyberg & Wallentinus, Reference Nyberg and Wallentinus2005; Pedersen et al., Reference Pedersen, Staehr, Wernberg and Thomsen2005; Sanchez & Fernandez, Reference Sanchez and Fernandez2005) such as: (1) very high growth rates of 2–4 cm per day (Critchley, Reference Critchley1981; Lewey & Farnham, Reference Lewey and Farnham1981; Nicholson et al., Reference Nicholson, Hosmer, Bird, Hart, Sandlin, Shoemaker and Sloan1981); (2) high fecundity, monoecious receptacles and a perennial life history (Fensholt, Reference Fensholt1955; Norton, Reference Norton1976; Norton & Deysher, Reference Norton, Deysher, Ryland and Tyler1989); and (3) multiple-range dispersal mechanisms including germling settlement, peripatetic plants and drifting fertile thalli (Norton, Reference Norton1976). In Strangford Lough (Northern Ireland), S. muticum has been a particularly successful invader (Strong et al., 2005). The habitat of S. muticum in Strangford Lough includes extensive soft sediment areas with only a small component of mobile hard substrata. This is a predominantly empty niche for macroalgae locally, which suggests that S. muticum might be benefiting from an element of release from inter-specific competition in these particular areas.
From the literature, we identified the two main elements of biotic pressure on native and introduced macroalgae in temperate marine habitats as: (1) mesoherbivore grazing, including detaching and consuming material (Duffy, Reference Duffy1990); and (2) epiphyte overgrowth, which can reduce photosynthesis below the compensation point and decrease gas exchange (Wahl, Reference Wahl1989; Steinberg & de Nys, Reference Steinberg and de Nys2002) but may also protect the alga from mesoherbivore grazing (Wahl & Hay, Reference Wahl and Hay1995; Karez et al., Reference Karez, Engelbert and Sommer2000) and excessive light (Dodds, Reference Dodds1991). Clearly the effects of mesograzers in combination with epiphytes may be different from either of these factors presented alone. For example, mesograzers could preferentially graze epiphytes and thereby reduce the impact of fouling, or conversely could consume the algal basiphyte and further impact the already fouled macrophyte (Karez et al., Reference Karez, Engelbert and Sommer2000).
The objective of this study was to compare the biotic pressures of mesoherbivore grazing and epiphyte fouling on S. muticum and three sympatric native brown macroalgae, Saccharina latissima (formerly known as Laminaria saccharina), Halidrys siliquosa and Fucus serratus. The native species were selected due to their locally high abundance, similar thallus size, and habitat overlap with S. muticum. These species are the most ecologically comparable to S. muticum although they differ morphologically and in phenology. Ultimately, this analysis should provide insights into the importance of intrinsic and extrinsic factors for S. muticum as an invasive species in Europe.
We tested the hypothesis that S. muticum will suffer less mesoherbivore grazing and algal fouling than native algal species because native grazers are not adapted to exploit it as a food source or as a substratum. This was tested by examining whether: (1) S. muticum has less fouling than native macroalgae; (2) fouling affects loss rate of thalli; (3) fewer mesograzers are found on S. muticum than on native macroalgae; (4) Dexamine spinosa (Montagu) (the main epiphytic mesograzer locally) prefers native macroalgae over S. muticum; and (5) the combined effects of Dexamine spinosa and epiphytic fouling differ from the effects of each alone.
MATERIALS AND METHODS
Study site
Strangford Lough, a large sea-lough with a narrow entrance, is situated on the north-east coast of Ireland (Figure 1). All collections were made in the Dorn (54° 26.105 N 005° 32.475 W), an inlet on the south-eastern shore of the Lough (Figure 1) which was colonized by Sargassum muticum in 1996 (Davison, Reference Davison1999). The substratum in the study area of the Dorn consists largely of small pebbles on mud, with some scattered loose rock fragments. In the subtidal fringe of the Dorn, S. muticum is found with three large native brown algae, Saccharina latissima (Linnaeus) C.E. Lane et al., Halidrys siliquosa (Linnaeus) Lyngbye and Fucus serratus Linnaeus. Sargassum muticum was studied at this site over a three-year period (2000–2003) and the presence of an ectocarpoid fouling epiphyte was observed in the spring to summer periods of each year. Monthly monitoring of epiphytic algae on S. muticum was carried out in the Dorn from February 2000 to November 2003.
Maps showing the locations of Strangford Lough in Northern Ireland (within box in top right map), the Dorn in the Lough (within box in left map) and the study site, marked with a cross, within the Dorn (bottom right map).
Quantification of epiphytic algae and fauna on Sargassum muticum and native macroalgae
The biomass of algal epiphytes on Sargassum muticum and the three selected native species (Saccharina latissima, Halidrys siliquosa and Fucus serratus) was determined during the peak of epiphytic fouling in June 2002. Eight thalli of each species (seven of H. siliquosa) were randomly taken from the same depth (−0.35 m below mean low water neap tides) and carefully bagged individually to prevent loss of epiphytes. At the laboratory, each thallus was carefully washed to remove sediment then the surface epiphytes were detached. Epiphytic algal fouling was not separated into species, as preliminary investigations revealed that the vast majority was ectocarpoids (probably Hincksia spp.). Basiphytic and epiphytic material was blotted dry, weighed fresh, and weighed again after drying for 24 hours at 60°C in an oven.
To obtain the mobile epifauna present at the peak of the epiphytic fouling, a separate collection of basiphyte thalli was made using the method described above, with additional precautions to prevent the loss of epiphytic fauna. Each basiphyte was weighed in the field with a spring balance and only thalli weighing 400 g ± 50 g were retained. At the laboratory freshwater washes were used to detach the fauna, which was subsequently collected on a 500 µm sieve. The collection bags were also washed into the sieve to gather any epiphytic fauna that detached during transport. All material retained on the sieve was sorted by species and enumerated with a dissecting microscope.
For data analysis, wet weight epiphytic algal biomass was standardized to a value per gram dry weight of basiphyte. Comparisons between epiphytes on each basiphyte species were made with a 1-way ANOVA and Tukey–Kramer post hoc tests using SPSS (SPSS for Windows, Version 11.0.1). A Levene test for homogeneity was performed on all un-transformed data. Data sets not showing homogeneity of variance were Log(n) transformed and the Levene test repeated to confirm homogeneity. The mobile epiphytic fauna was analysed with the multivariate community analysis program ANOSIM in PRIMER (Plymouth Routines in Marine Ecological Research; Primer for Windows, Version 5.2.9. 2002. PRIMER-E Ltd). SIMPER analysis in PRIMER identifies which species generate the most dissimilarity between ‘treatments’. For the SIMPER routine, the raw data were square root transformed and reporting was limited to species with more than 2.5% contribution to dissimilarity. The DIVERSE program in PRIMER was used to calculate univariate community descriptive statistics (Clarke & Gorley, Reference Clarke and Gorley2001). Analysis of these values was also undertaken with ANOVA in SPSS.
Effect of Dexamine spinosa on Sargassum muticum and native macroalgae
Dexamine spinosa was selected as the experimental mesograzer as it was the most abundant and ubiquitous species on all basiphytes. Clean epiphyte-free tips of Sargassum muticum, Saccharina latissima, Fucus serratus and Halidrys siliquosa were obtained from the Dorn in June 2002. Dexamine spinosa were collected from the same four species of macroalgae by vigorously washing thalli in buckets of seawater, after which the amphipods were sorted according to species and size.
The experimental units for the grazing preference consisted of sixteen 1 l plastic containers with numerous fine perforations too small to allow loss of amphipods or algal biomass, yet sufficient to allow rapid water exchange. A cutting of each of the four macroalgal species 150 mm in length (~5–7 g wet weight) was placed in each container. In half of the containers (amphipod treatments), six large (6–12 mm) and six small (<6 mm) individuals of D. spinosa were added; eight containers were controls without amphipods. This density of amphipods per gram basiphyte was comparable to field observations made earlier. It was necessary for all four species to be placed into each beaker to provide equal access and choice to the grazers added to them. The total biomass of algae and amphipods per container resulted in only a very low stocking level, hence maintaining high water quality over the 2-week treatments.
Four containers were randomly allocated into each of four vigorously aerated seawater baths containing a large volume of filtered and UV sterilized seawater which effectively made the beakers independent of each other. Seawater was maintained at ~15°C to match seasonal field values and illumination from above by two white fluorescent tubes provided ~75 µmol m−2 s−1 in a 12/12 h light/dark cycle. Regular water changes were made throughout the experiment to maintain a high water quality and prevent nutrient limitation. After two weeks the algal tips were carefully blotted dry of surface water and weighed.
It is recognized that by having all four species present in the same containers that the consumption of each is not strictly independent: had replication been greater, one randomly chosen piece of algae from each container would have been selected to overcome this limitation (Pavia et al., Reference Pavia, Carr and Åberg1999).
Paired t-tests were carried out on the weights of the macroalgal cuttings before and after the experiment to establish whether there were any significant changes in weights during the experiment. The values were converted to percentage change values and a nested ANOVA was used to test for differences between amphipod and control treatments within an algal species and to determine grazing preference between macroalgal species. Unequal variances (shown by the Levene statistic) were Log(n) transformed.
Interactions between Sargassum muticum, ectocarpoid epiphytes and Dexamine spinosa
Sixteen non-reproductive thalli of Sargassum muticum were collected from the Dorn (on 4 June 2002), as previously described, and branches (secondary laterals) were taken from the mid-section of each thallus. Epiphytic ectocarpoids and Dexamine spinosa were obtained by vigorously washing plants of S. muticum in a bucket of seawater; amphipods were sorted according to species and selected with a large diameter pipette. The loosely attached ectocarpoids were easily detached during the washing although other fouling species required extra effort to remove them from the thalli. The same experimental apparatus as used above was used for this series of treatments with twelve perforated containers split between six seawater baths. Each seawater bath was filled with 10 l seawater and vigorously aerated (irradiance and photoperiod were as above). In all containers, water changes were undertaken throughout the experiment to maintain high water quality conditions.
A factorial experimental design (Table 1) was used with S. muticum or plastic basiphytes (artificial aquarium plants of a similar architecture to S. muticum); each type of basiphyte was present by itself and with amphipods and/or ectocarpoid fouling. In treatments 1 and 2 the effect of amphipods on the basiphyte S. muticum is quantified (Table 1). Ectocarpoids on S. muticum in treatments 3 and 4 were used to determine the effect of amphipods when ectocarpoids were present. Treatments 5 and 6 used plastic basiphytes to isolate the interaction between epiphytic amphipods and ectocarpoids in the absence of S. muticum.
Treatments within the factorial experiment examining the interaction between basiphyte (Sargassum muticum), epiphytic alga (ectocarpoid species) and epiphytic fauna (Dexamine spinosa).
Treatments with amphipods received three large (6–12 mm) and two small (<6 mm) individuals of Dexamine spinosa (the density of amphipods differs from the experiment above as densities are standardized to the amount of basiphytic biomass). Treatments with algal epiphytes received approximately 1 g blotted ectocarpoid for each g of basiphyte (S. muticum or plastic basiphyte). Ectocarpoids were loosely attached to the basiphyte in a similar fashion to natural fouling observed in the Dorn. At the end of weeks three, five and seven, S. muticum and ectocarpoids were carefully separated, blotted and weighed. When returned to the containers, the ectocarpoids were again loosely wound round the laterals of S. muticum.
At the end of the experiment in week seven, changes in biomass of the basiphyte and the epiphyte were analysed separately with two-way ANOVA in SPSS. The design of this experiment had to make some compromises on the level of replication so that all of the variables could be examined. The factorial design partially helped to overcome some of these replication-related compromises. For the basiphyte, the two-way ANOVA included the presence of the amphipods and the epiphyte as individual factors, as well as the possible interaction between factors. The two-way ANOVA for the epiphyte had the presence of the basiphyte and amphipods as individual factors. Unequal variances (shown by the Levene statistic) were Log(n) transformed.
Effect of ectocarpoid epiphytes on loss rates in natural Sargassum muticum and Saccharina latissima stands
To assess the effect of ectocarpoid fouling on the loss rate of basiphytes in the Dorn, four separate rectangular 18 m2 areas were marked out in the subtidal fringe with similar substratum, depth and water flow and roughly equal proportions of Saccharina latissima and Sargassum muticum. The numbers of S. latissima (N = 167) and S. muticum (N = 114) were counted and each thallus was labelled with a small plastic tag inserted through the stipe of the former species and a primary lateral of the latter. Each area was split in half to produce 9 m2 boxes allocated to either ‘fouled’ or ‘cleaned’ treatments. For ‘cleaned’ treatments, thalli of S. latissima and S. muticum were cleaned of algal epiphytism by hand each week and thereby kept artificially clear of fouling. In the ‘fouled’ treatments, thalli were allowed to accumulate ectocarpoid fouling naturally. After 12 weeks, the numbers of labelled S. latissima and S. muticum were recorded in each box.
The density of thalli in the fouled stands and those in the cleaned treatments were compared after 12 weeks with paired t-tests. Two-way ANOVA was used to compare the percentage decline of individuals between fouled and cleaned treatments and between macroalgal species.
RESULTS
Epiphytic algae on Sargassum muticum and native macroalgae
In the Dorn, fouling by epiphytic algae on Sargassum muticum was greatest from March to June, concurrent with the period of greatest growth of S. muticum, and declined rapidly in July (Figure 2). Thalli of S. muticum became reproductive from July onwards, by which time the epiphytism had almost completely disappeared. Similar observations showed the timing of the epiphytic bloom to be similar for all species of basiphyte considered below.
Wet weight biomass of Sargassum muticum and attached epiphytic ectocarpoid epiphytism in 2002 in the Dorn, Strangford Lough; bars are Standard Deviation.
The amount of epiphytic algae on S. muticum was not significantly different from that on Fucus serratus and Halidrys siliquosa (Figure 3). There was a significantly lower epiphyte biomass on Saccharina latissima than that on the three other species of basiphyte examined. Laboratory observations indicated that the fouling biomass was almost exclusively ectocarpoids on all four basiphyte species.
Total epiphytic algal biomass on three species of native macroalgae and Sargassum muticum collected from hard substratum in the Dorn, Strangford Lough. ANOVA (log transformed data) was used to test epiphytic fouling between basiphytes (one-way ANOVA, df = 3.19, f = 9.831, fcrit = 3.072, P = 0.001). Bars are standard deviations. Letters A and B above bars indicate homogeneous sub-sets as identified by a post-hoc Tukey–Kramer tests.
Epiphytic fauna on Sargassum muticum and native macroalgae
All comparisons of the mobile epiphytic fauna between basiphyte species were significantly different (Table 2). Each basiphyte species had a characteristic epiphytic faunal community during the period of greatest epiphytic fouling (Figure 4). The assemblages on S. muticum and H. siliquosa were more strongly clustered with higher levels of community similarity when compared with those found on F. serratus and S. latissima.
Multidimensional scaling plot (PRIMER) of the epiphytic assemblages on Sargassum muticum and three native macroalgae collected from the Dorn, Strangford Lough.
Pairwise ANOSIM comparisons of epiphytic faunal composition on four macroalgal species collected from the Dorn, Strangford Lough. As there is an element of multiple hypothesis testing, a P value of 0.008 has been used for the global significance threshold of P < 0.05.
ANOSIM, one-way, sample statistic (Global R): 0.729, P < 0.001.
The mean number of mobile epiphytic species did not differ significantly between basiphytes (Table 3). The greatest variation was in the total community abundance (i.e. number of epiphytic individuals per thallus). On average, each S. muticum thallus hosted over 900 individuals, much greater than H. siliquosa, F. serratus and S. latissima, which each supported 50–100 individuals. The species richness values on S. muticum were lower than for the other basiphytes, on a per thallus basis (Table 3), due to uneven distribution of the high community abundance among the species (80% of the individuals were one species).
Univariate descriptive statistics for the mobile epiphytic fauna diversity per thallus on four macroalgal species collected from the Dorn, Strangford Lough. All values are per thallus (all thalli weighed 400 g±50 g).
Number of species: ANOVA one-way, df = 3,28, f = 0.317, P = 0.813; number of individuals (log transformed): ANOVA one-way, df = 3,28, f = 25.294, P < 0.000; species richness (log transformed): ANOVA one-way, df = 3,28, f = 6.789, P = 0.001; Pielou's evenness: ANOVA one-way, df = 3,28, f = 18.513, P < 0.000. Superscript letters are homogeneous sub-sets as identified by post-hoc Tukey–Kramer tests.
The SIMPER analysis identified eight important epiphytic faunal species that characterize similarity within assemblages on particular basiphyte species and hence differences between basiphyte species (Table 4). There were very high densities of Dexamine spinosa (over 700 individuals) on each thallus of S. muticum and variable numbers on the other basiphyte species (Table 4). Moderate numbers of Littorina mariae and Isopoda were characteristic of F. serratus (Table 4). Saccharina latissima supported a broad community with many species contributing to similarity, although Gibbula umbilicalis, Corophium volutator and Ischyrocarus anguipes were particularly characteristic. However, the overall abundance of fauna not found on S. muticum but present on the other species was very low and inconsistent between the native species.
SIMPER analysis (PRIMER) of the epiphytic faunal communities on four macroalgal species collected from the Dorn, Strangford Lough.
* All thalli weighed 400 g ± 50 g, N = 8.
Effect of Dexamine spinosa on Sargassum muticum and native macroalgae
The wet weight of control samples of F. serratus, S. latissima and H. siliquosa increased during the experiment (Figure 5). Control samples of S. muticum decreased in weight by 0.05% (Figure 5). In the amphipod treatments F. serratus and H. siliquosa still increased in weight—this was not statistically different to the controls. Conversely, S. latissima declined by approximately 10% when amphipods were present, not significantly different from the control treatment. Samples of S. muticum grown with amphipods declined by 62% after two weeks, significantly different from control values for this species: observations during the experiment indicated that the leaflets were consumed first and many of the bladders detached. The loss of basiphytic material, in the presence of Dexamine spinosa, follows this series: H. siliquosa < F. serratus < S. latissima < S. muticum.
Percentage change in wet weight of four macroalgal species after two weeks in control (no amphipods present) or amphipod (Dexamine spinosa) grazing treatments. Asterisks indicate significant differences (post-hoc: Tukey HSD) between amphipod and control treatments within a macroalgal species (two-way ANOVA; df = 1, f = 39.405, P = 0.000). Comparisons between seawater baths (nested ANOVA; df = 15, f = 1.136, P = 0.352).
Interactions between Sargassum muticum, ectocarpoid epiphytic algae and Dexamine spinosa
In the absence of ectocarpoids, Sargassum muticum grown without amphipods gained significantly more wet weight than thalli cultured with amphipods, which were consumed as quickly as the algal laterals grew over the experimental period (Figure 6). In the presence of ectocarpoid fouling, the biomass of S. muticum remained constant when there were no amphipods but when amphipods were present it decreased throughout the seven-week period (Figure 6), and finally appeared to be mostly consumed. Ectocarpoid biomass placed onto plastic basiphytes without amphipods increased in weight over the 7-week period (Figure 7). When ectocarpoids were present on the plastic basiphytes with amphipods, the biomass remained constant over the experimental period (Figure 7).
Effect of Dexamine spinosa and ectocarpoid epiphytes on the growth of Sargassum muticum in a factorial experimental. Mean (±standard deviation, N = 2) wet weight biomass of thalli of S. muticum shown. Letter indicates significant difference between treatments at week 7; two-way ANOVA (presence/absence of ectocarpoid), df = 1, f = 23.868, fcrit = 7.709, P = 0.008; two-way ANOVA (presence/absence of amphipods) df = 1, f = 41.070, fcrit = 7.709, P = 0.003; interaction df = 1, f = 0.971, fcrit = 7.709, P = 0.380. Power analysis states that replication used will detect differences of 5.00023 g with a power of 0.8.
Effect of Dexamine spinosa and basiphyte presence on the growth of ectocarpoid epiphytism. Mean (±standard deviation, N = 2) change in the wet weight biomass of ectocarpoid shown. A plastic basiphyte was used for treatments without Sargassum muticum. Letters indicate significant difference between treatments; ANOVA two-way (presence/absence of S. muticum), df = 1, f = 9.348, fcrit = 7.709, P = 0.038; ANOVA two-way (presence/absence of amphipods) df = 1, f = 5.233, fcrit = 7.709, P = 0.084; interaction df = 1, f = 0.236, fcrit = 7.709, P = 0.652. Power analysis states that replication used will detect differences of 5.00023 g with a power of 0.8.
Effect of ectocarpoid epiphytes on loss rates in natural Sargassum muticum and Saccharina latissima stands
The epiphytized S. muticum thalli suffered a mean loss of 70% of the plants, whereas the stand of fouled S. latissima thalli experienced a significantly lower (ANOVA) loss of 9% (Table 5). By contrast, in the ‘cleaned’ treatments, there was no significant change in the plant density of S. latissima or S. muticum over the 12 week period (Table 5).
Density of Sargassum muticum and Saccharina latissima stands before and after 12 weeks with and without cleaning of epiphytic ectocarpoid fouling (Dorn, Strangford Lough).
Before and after comparisons of actual thalli density: paired t-test two-tailed, critical t = 3.182 at 3 degrees of freedom. Superscript letters: ANOVA two-way (comparisons between fouled and unfouled treatments) df = 1, f = 129.084, fcrit = 4.747, P = <0.001 and (comparisons between macroalgal species) df = 1, f = 104.435, fcrit = 4.747, P = <0.001, interaction df = 1, f = 74.983, fcrit = 4.747, P = <0.001. Homogeneous sub-set as identified by post-hoc Tukey–Kramer tests.
DISCUSSION
Epiphytic algae and fauna on subtidal macroalgae in the Dorn, Strangford Lough
Sargassum muticum was heavily epiphytized and grazing by a common and generalist native herbivore was substantial. The peak in fouling algal biomass occurred in May and June within the Dorn. Unlike the findings of most other reports (e.g. Bjæke & Fredriksen, Reference Bjæke and Fredriksen2003), but as in the Danish study by Thomsen et al. (Reference Thomsen, Wernberg, Stæhr and Pedersen2006), the peak of epiphytism on S. muticum was during the period of greatest growth rather than later in the season after reproduction. During this peak in ectocarpoid fouling, the attached epiphytic biomass on S. muticum was similar to that seen on the three sympatric species. The biomass of algal fouling was greater on the branched and complex algae, i.e. S. muticum and Halidrys siliquosa, where their architecture retained loosely attached epiphytic biomass, especially the main bushy ectocarpoid species. It was also interesting that some silt was also associated with the ectocarpoid biomass that will have exacerbated smothering and shading of the basiphyte. Future work will have to account for the impact from both the epiphytic biomass and the silt bound within the fouling mass.
With regard to the epiphytic fauna, herbivorous amphipods associated with S. muticum were exceptionally numerous, which corresponds with previous studies on their habitat preferences and macroalgal structure (Gee & Warwick, Reference Gee and Warwick1994; Heckscher et al., Reference Heckscher, Hauxwell, Jimenez, Rietsma and Valiela1996; Schmidt & Scheibling, Reference Schmidt and Scheibling2007) and is probably related to the complex morphology of the basiphyte. Equally, the complex morphology of S. muticum is also likely to trap more suspended sediment and thereby modify the abundance and nature of epiphytic fauna with a detrital component to their diet.
The epiphytic community on S. muticum in Strangford Lough was very similar to fauna observed on S. muticum by Norton & Benson (Reference Norton and Benson1983) at Friday Harbor, Washington, and Viejo (Reference Viejo1999) in El Truhan Inlet, northern Spain. Both studies found assemblages characterized by abundant amphipods. However, Norton & Benson (Reference Norton and Benson1983) also reported a high abundance of the herbivorous gastropod Lacuna variegata, whereas very few epiphytic gastropods were seen on this alga in Strangford Lough. In contrast, thalli of S. muticum collected in the Solent from 1975 to 1978 by Gray (Reference Gray1978) had a substantially different epiphytic fauna characterized by sedentary and encrusting filter feeders confined to the older parts of the thallus (Withers et al., Reference Withers, Farnham, Lewey, Jephson, Haythorn and Gray1975). Gray (Reference Gray1978) reported that a few species of amphipods were numerous, with Gammarus locusta being the most abundant (up to 150 individuals per thallus), compared to single-species abundances in Strangford Lough that consistently exceeded 700 per thallus. Equally, the analysis of S. muticum in Limfjorden (Denmark) by Wernberg et al. (Reference Wernberg, Thomsen, Staehr and Morten2004) also found a rich epiphytic fauna which differed substantially from that found in this study. The differing composition of epiphytic faunal assemblages may be due to environmental and community differences in each area. It is a clear indication also of how biotic pressure from native species can vary throughout the introduced range.
Grazing preference experiments using Dexamine spinosa indicate that S. muticum was consumed in preference to the native macroalgae, hence thalli provide both habitat and grazing material for this amphipod. Norton & Benson (Reference Norton and Benson1983) also reported that several members of the epiphytic fauna at Friday Harbor were herbivorous—grazing experiments confirmed that the abundant gastropod Lacuna variegata and the amphipod Ampithoe mea consumed substantial quantities of S. muticum. Field observations of S. muticum by Norton & Benson (Reference Norton and Benson1983) revealed that 78% of the primary laterals lacked tips and secondary laterals exhibited progressively more damage away from the apices. Gray (Reference Gray1978), Norton & Benson (Reference Norton and Benson1983), Viejo (Reference Viejo1999) and Withers et al. (Reference Withers, Farnham, Lewey, Jephson, Haythorn and Gray1975) have all documented grazing of S. muticum by herbivorous amphipods.
Mesoherbivore feeding preferences vary among macroalgae for several reasons. Commonly, preferences are set by either the concentration or potency of chemical defences in the algal thallus (Hay & Fenical, Reference Hay and Fenical1988), local abundance or the nutritional requirements of the mesoherbivore (Cruz-Rivera & Hay, Reference Cruz-Rivera and Hay2001). The density, size and isolation of host basiphytes can also influence the grazing pressure placed upon them. Certain mesograzers will change algal preference depending on the relative abundance of each macroalgal species, hence poorer hosts are often selected over more nutritious ones when they are more abundant (Poore, Reference Poore2004). The levels of phlorotannin compounds were not measured in this study, but all of the macroalgal species examined are known to contain tannin-like substances (S. muticum; Conover & Sieburth, Reference Conover and Sieburth1964; Glombitza et al., Reference Glombitza, Eckhardt and Farnham1982; Hay & Fenical, Reference Hay and Fenical1988; Plouguerne et al., Reference Plouguerne, Le Lann, Connan, Jechoux, Deslandes and Stiger2006). Within single and mixed species treatments, S. muticum was preferentially grazed. Cruz-Rivera & Hay (Reference Cruz-Rivera and Hay2001) found that the amphipod Ampithoe longimana did not rely on mixed macroalgal diets but underwent compensatory feeding (increased consumption of material of a lower nutritional value) to obtain the required nutritional balance/intake within its diet. The range of herbivore responses to macroalgal foodstuffs is highly varied. Without knowing more about the nutritional and defensive properties of the algae used in this study it is difficult to understand the grazing response. What was apparent from this study is that S. muticum was heavily populated with D. spinosa and that native biotic resistance from grazing is strong.
Interactions between basiphyte and epiphytic algae and fauna
The cumulative interactive impact of epiphytic algae and mesoherbivore grazing resulted in the rapid decline of young and non-reproductive S. muticum thalli. Dexamine spinosa was able to consume both basiphyte and epiphyte, like other epiphytic mesograzers including amphipods (Cruz-Rivera & Hay, Reference Cruz-Rivera and Hay2000; Karez et al., Reference Karez, Engelbert and Sommer2000). Although most mesoherbivores consume the epiphyte in preference to the basiphyte, many species will consume both types of vegetation (Bulthuis & Woelkerling, Reference Bulthuis and Woelkerling1983; Cattaneo, Reference Cattaneo1983; Duffy & Hay, Reference Duffy and Hay2000). Grazers can often switch diet depending on the relative abundance of epiphyte and basiphyte vegetation (Brawley & Fei, Reference Brawley and Fei1987) or on the nutritional requirements needed to sustain optimum fitness. The amphipod, Ampithoe marcuzzii Ruffo, strongly preferred Sargassum filipendula Agardh over the epiphyte, Ectocarpus siliculosus (Dillwyn) Lyngbye (Duffy, Reference Duffy1990). Fitness of this amphipod was no greater in a mixed algal treatment (hypothesized to generate greatest fecundity and hence fitness) than when fed on either the basiphyte or epiphyte alone (Cruz-Rivera & Hay, Reference Cruz-Rivera and Hay2000).
Karez et al. (Reference Karez, Engelbert and Sommer2000) documented enhanced negative effects on basiphytes during the interaction between epiphytism and mesograzing, which they termed ‘co-consumption’. The epiphytic isopod Idotea granulosa consumed significantly more Fucus vesiculosus when it was fouled by an epiphyte than when presented as an unfouled basiphyte. Co-consumption may also explain why basiphyte biomass declined so quickly in our experiment when both amphipods and epiphytic algae were in the same culture.
The presence of fouling itself can have many disadvantages for the basiphyte. In culture, when the epiphyte was present on S. muticum (no amphipods), there was no net growth of basiphyte suggesting that ectocarpoid fouling had a detrimental effect, possibly through competition for nutrients, dissolved gases and light. Photosynthetic studies of S. muticum have revealed that this species requires high irradiance levels (Hales & Fletcher, Reference Hales and Fletcher1989; Rico & Fernandez, Reference Rico and Fernandez1997) (approximate saturation at 700 µmol photons m−2 s−1; Strong unpublished), hence it was presumed that fouling always represented an impact, most likely from shading, and no mutualism would be evident.
Comparison between ‘fouled’ and ‘cleaned’ thalli of S. muticum and S. latissima in the field showed that smothering by ectocarpoids caused a significant loss of thalli of the former species. Field observations also found that fouling was much greater on S. muticum, probably due to the complex architecture of the thalli. Individuals of both species were attached to loose rock. The loss of S. muticum thalli was probably related to excessive shading and grazing, as well as the high levels of fouling generating more drag, which will have pulled individuals out of the monitored area. The fate of the individuals lost from the experimental site is not known but field observations suggest some were cast high on the shore and died. The occupation of loose substrata by S. muticum is characteristic of large areas of Strangford Lough and leads to very high rates of peripatetic movement that may not be seen in other areas of this species distribution (Strong et al., Reference Strong, Maggs and Dring2005).
The impact of epiphytism in field experiments has been best studied in seagrass habitats. The decline of seagrass beds is particularly likely in nutrient-enriched water, where it is largely caused by the excessive growth of algae on the seagrass blades (McGlathery, Reference McGlathery2001). Phytoplankton biomass and total suspended particles also increase in nutrient-enriched water and further reduce light penetration to benthic communities (McGlathery, Reference McGlathery2001). Algal fouling of submerged aquatic vegetation has been found to reduce light by 80%, and thereby reduce photosynthesis of some basiphytes below their compensation point (Bulthuis & Woelkerling, Reference Bulthuis and Woelkerling1983; Bronmark, Reference Bronmark1985). In Chesapeake Bay Zostera marina transplants into areas previously occupied by this species failed to survive, which has been related to seasonal variation in light attenuation and the accumulation of a dense epiphyte layer during late spring (Moore et al., Reference Moore, Neckles and Orth1996).
In response to the original hypothesis, our results suggest that:
1) Ectocarpoid fouling on S. muticum was comparable to that on native macroalgae.
2) Herbivorous amphipod abundance was much greater on S. muticum in comparison with other native macroalgae.
3) Basiphytes of S. muticum were preferentially consumed by D. spinosa over native macroalgae.
4) Dexamine spinosa consumed both S. muticum and ectocarpoid biomass (even when given the option of either vegetation).
5) The combination of amphipod grazing and ectocarpoid fouling led to a rapid decline of both S. muticum and ectocarpoid biomass.
6) Heavy fouling by epiphytic ectocarpoids resulted in thalli loss from natural stands of S. muticum.
Both the periodic ectocarpoid fouling and the continuous mesograzer presence represent significant biotic pressures on the population of S. muticum, with sizable declines in the density of this species being observed during the bloom period; based on these observations, hypotheses 1, 3 and 4 should be rejected. It is apparent that S. muticum is thoroughly exploited as a substratum for epiphytic algae and a food source for epiphytic mesoherbivores. This exploitation represents a serious biotic pressure that generates a thinning mechanism on the invasive population.
Despite being a recent introduction to Strangford Lough, S. muticum is under substantial pressure from epiphytic algae and mesoherbivores, probably resulting in a high biotic resistance. On balance of the available data, it is highly likely that extrinsic factors, in the form of ecological release, are only of minor importance in determining the success of S. muticum as an invasive species, although interspecific competition between macroalgae cannot be ruled out. It is highly likely that the fundamental intrinsic traits of this species, e.g. fast growth, multiple dispersal mechanisms and high fecundity, ultimately underpin its ability to disperse and invade such large areas of new habitat. Only through the balanced investigation of intrinsic and extrinsic factors of particular invasions can the concepts of invasive traits and habitat invasibility be understood. These studies may also yield interesting insights into the control of established invasion and the mechanisms behind permanent control and naturalization.
ACKNOWLEDGEMENTS
We wish to thank Professor M.J. Dring for advice and assistance on aspects of experimental design. Substantial amounts of help with the fieldwork were kindly provided by Emma McLaughlin, Laura Knipe and Niamh Small. Support for this research came as a CAST studentship award from the Environment and Heritage Service Northern Ireland and the Department of Agriculture and Rural Development Northern Ireland.
