Direct and indirect effects of Gracilaria vermiculophylla on native Fucus vesiculosus

The perennial red macroalga Gracilaria vermiculophylla (Ohmi) Papenfuss has recently been introduced to the Baltic Sea and is a potential competitor to Fucus vesiculosus, the most common native perennial alga in large parts of the Baltic Sea. Gracilaria might interfere with Fucus through direct competition for resources. In addition, Gracilaria is a favoured refuge for mesograzers, which prefer to feed on Fucus. Mesocosm-experiments were conducted over one year in the Kiel Fjord in order to test the direct and indirect effects of Gracilaria on Fucus. Fucus was incubated with Gracilaria at three different densities and grazers in high or low abundances. High densities of Gracilaria inhibited the growth of Fucus adults and also reduced the half-life-time of Fucus germlings. Associated grazers also had a negative effect on Fucus adults. Our results suggest that Gracilaria is able to influence Fucus in the Baltic Sea through direct competition for resources and by exposing it to higher grazer pressure.


Introduction
Seaweeds are frequent marine invaders (Schaffelke et al. 2006).The most common effect of invasive macroalgae in their new habitat is space monopolization that results in a reduced abundance of native species (reviewed in Schaffelke and Hewitt 2007).Loss of biodiversity and effects on fish and invertebrates were also reported (e.g.Boudouresque et al. 1992;Levin et al. 2002;Relini et al. 1998).In the case of Caulerpa, toxic effects on other biota were documented (e.g.Paul and Fenical 1986, reviewed in Schaffelke and Hewitt).However, most information originates from field surveys or observational studies, and there are only very few experimental studies about the impacts of invasive algae on their recipient ecosystems (Schaffelke and Hewitt 2007).As a consequence, the mechanisms underlying ecological effects of invasive species are mostly unknown (Schaffelke and Hewitt 2007).Manipulative communitylevel field studies, combined with modeling, are necessary to identify the impacts of introduced seaweeds on native communities (Williams and Smith 2007).
The perennial red macroalga Gracilaria vermiculophylla (Ohmi) Papenfuss (hereafter: Gracilaria) has invaded many coastal areas worldwide.From its native distribution in East Asia it has spread to other temperate areas of the northern hemisphere, primarily along both North American coasts (Bellorin et al. 2004;Freshwater et al. 2006;Thomsen et al. 2005) and in the East Atlantic between Morocco and the Western Baltic Sea (Guillemin et al. 2008;Nyberg 2007;Rueness 2005;Thomsen et al. 2007;Weinberger et al. 2008).Accounting for 74% of the entire algal biomass, Gracilaria dominated the shallow water habitat in Hog Island Bay, Virginia (Thomsen et al. 2006).Two recent studies show that Gracilaria is able to negatively impact metabolism and survival of Zostera marina by creating an anoxic layer (Martínez-Lüscher and Holmer 2010; Höffle et al. 2011).Possible ecological effects of Gracilaria on macroalgae have not been studied.
Gracilaria tolerates high irradiances and can grow in shallow water.Therefore, it is predicted to be a strong competitor for the native Fucus vesiculosus L. (Nyberg 2007, hereafter: Fucus).Fucus is one of the most important phytobenthic species on hard bottom habitat in the Baltic coastal zone as it provides a habitat for speciesrich epiphytic and epibenthic communities (Torn et al. 2006).During the last few decades, populations of Fucus have declined over large areas in the Baltic Sea, most likely as a consequence of eutrophication (Torn et al. 2006).Nowadays, Fucus is mainly restricted to shallow water with depths between one and two meters in the western Baltic Sea (Torn et al. 2006).The presence of Gracilaria could increase the pressure on Fucus further, and observations from the Kiel Fjord indicate that Fucus may have declined in habitats where Gracilaria has spread (Weinberger et al. 2008).In the newly invaded habitats in the Baltic Sea, Gracilaria appears as drifting mats, sometimes partly buried in the sediment or entangled in other seaweeds or mussels.These drifting mats of Gracilaria overgrow both soft bottom and hard bottom substrates (Weinberger et al. 2008) and may interfere with the settlement of Fucus germlings and reduce the growth of adult Fucus.Orrock et al. (2010) developed the idea of refuge-mediated apparent competition, i.e. an indirect interaction whereby plants provide a refuge for a shared consumer, subsequently increasing consumer pressure on another plant species.In the Baltic Sea, Gracilaria provides an attractive habitat for herbivores (Thomsen 2010), especially for isopods (Idotea spp.), which are often present at higher densities on Gracilaria than on Fucus when both algal species live in the same habitat (Weinberger et al. 2008).Noticeably, the common periwinkle Littorina littorea and the isopod Idotea baltica show a preference for consuming Fucus rather than Gracilaria when both algae are offered in two-way-choice experiments (Weinberger et al. 2008).Herbivores can exert a strong influence upon Fucus species (Dethier et al. 2005).In the Baltic Sea, grazing by isopods (Idotea spp.) and periwinkles (Littorina spp.) leads to biomass loss of Fucus (Engkvist et al. 2000;Råberg and Kautsky 2008;Torn et al. 2006).Thus Gracilaria may potentially influence the decline of Fucus through two different mechanisms, 1) directly through overgrowth and competition for resources, 2) indirectly as it provides a habitat for grazers that feed on adjacent Fucus.
The main purpose of our study was to investigate experimentally whether Fucus was affected by the presence of Gracilaria or by mesograzers associated with it.As life stages can be influenced differently by invaders, we focused not only on the effect of Gracilaria on the growth of adult Fucus, but also investigated the effect upon the survival of Fucus germlings.

Material and methods
To investigate the impact of Gracilaria upon Fucus two (for Fucus germlings) or seven (for Fucus adults) sequential experiments were conducted as seasonal repetitions.In these experiments Fucus was incubated with Gracilaria in mesocosms floating on the Baltic Sea surface.Gracilaria was applied at three different densities, in order to test whether its effects on Fucus were density dependent.The applied Gracilaria contained grazers at either high or low density, which allowed determining the influence of grazers.The effect of Gracilaria and/or associated grazers was measured as growth of adult Fucus specimens and as half-life time of Fucus germlings.

Set-up, stocking and incubation of mesocosms
The experiments were carried out in the Kiel Fjord in a sheltered bay (N54°21.965', E10°8.908') between March 2008 and March 2009.The mesocosms consisted of white polyethylene boxes (60  40  40 cm) and had a bottom made of gauze (mesh size: 1 mm) to allow for an exchange of water with the Baltic Sea but to inhibit migration of grazers.They were floating on the water surface and contained a water column of ca. 30 cm and a water volume of approximately 70 l.
Light and temperature data were provided through constant measurements of the research division "marine meteorology", GEOMAR.
Each mesocosm was equipped with one stone with a medium-sized adult Fucus -devoid of grazers -and one tile with Fucus germlings.The chronology of the seven subsequent experiments with adult Fucus (hereafter: adults' experiments), the two incubation periods of Fucus germlings (hereafter: juveniles' experiments) and the environmental conditions during our studies are illustrated in Figure 1.
Adult Fucus individuals (mean 4.6 g WW ± 0.2 SE) growing on small stones were collected from the coastline of the Kiel Fjord.They were replaced with newly collected specimens in the beginning of each adults' experiment.In order to obtain juveniles, tiles (15  15 cm) were inoculated with Fucus zygotes in the laboratory.One series of tiles was inoculated in January 2008 and one series in June 2008, in order to cover both reproduction periods of Fucus in the Baltic Sea (Berger et al. 2001).The winter germlings had an age of approximately 50d when they were transferred to the mesocosms for the first juveniles' experiment.The summer germlings were only 10 days old when the second juveniles' experiment started, but due to lower growth rates in winter the initial size of germlings was similar in both experiments (approximately 10 cells).
There were three different Gracilaria density treatments with ten replicates each (Figure 2): No Gracilaria, Gracilaria at a low density (5 -10% of the 0.24 m² bottom of the mesocosm covered with Gracilaria, equalling 100 g of biomass FW) and Gracilaria at a high density (100 % covered, equalling 1000 g FW).These densities were considered relevant since Gracilaria ground coverage of up to 100% was previously observed in certain areas of the Kiel Fjord (Weinberger et al. 2008).For stocking, Gracilaria from drifting mats was collected together with its associated grazers with a dip net (mesh size 0.1 mm) in shallow parts of the Kiel Fjord.
Each of the three Gracilaria density treatments was conducted in five mesocosms with low grazer density and in five mesocosms with high grazer density (Figure 2).In mesocosms containing no Gracilaria the natural density of grazers associated with Fucus collected in nature was regarded as high, while absence of grazers was regarded as low.The natural grazer density on Fucus was determined prior to each experiment by counting all grazers that were present on five Fucus specimens collected with a gauze net (mesh size 0.1 mm) and comparable densities of grazers were added into mesocosms without Gracilaria, but with high grazer density (Figure 2).
In mesocosms containing Gracilaria the natural density of grazers in Gracilaria mats was considered as high, while an artificially reduced density was considered as low.For reduction of grazer densities half of the Gracilaria stocking material was treated with freshwater for two hours.This procedure resulted in the removal of a part of the associated grazers and thus generated low grazer densities.Preliminary tests had confirmed the efficiency of this method.Moreover, a control experiment was conducted after each freshwater treatment, in order to ascertain that Gracilaria was not affected.For this purpose, six samples of freshwater treated and six samples of untreated Gracilaria (10 g each) -in both cases grazers had been manually removed -were incubated in single mesh bags (mesh size: 1  1 cm) at 15°C and under artificial light in the same aquarium, provided with sea water from the Baltic Sea.After seven days the growth of treated and untreated Gracilaria was determined and the effect of the two treatments was compared.No significant differences between treated and untreated Gracilaria were found in any of the experiments (Mann-Whitney-U-test, p > 0.05).Thus, the freshwater-treatment had no detectable effect upon the growth of Gracilaria.
Prior to the start of each experiment the initial density of grazers associated with Gracilaria stocking material was determined by taking samples of untreated and freshwater treated Gracilaria with a dip net.The samples were fixed with formaldehyde in seawater (4%) and the grazers were picked under a stereo microscope, classified and counted.
In order to reduce the impact of fouling organisms attached to the mesocosm walls, they were replaced with clean mesocosms in the beginning of each adults' experiment.At these occasions Gracilaria was also removed and replaced with either 100 g or 1000 g of newly collected Gracilaria from the Kiel Fjord, either treated with freshwater or not, as described above.In contrast, tiles bearing Fucus juveniles were kept over several adults' experiments (see Figure 1), and transferred into the newly stocked mesocosms.

Sampling collection
For growth measurements, adult Fucus individuals were cleaned and spread under a plexiglass panel and photographed beside a benchmark in the beginning and at the end of each adults' experiment.The photos were evaluated by using a computer imaging program (Sigma Scan Pro, STATCON, Witzenhausen, Germany) which counted pixels representing the benchmark area and Fucus, thus allowing for calculation of the algal surface area.
Four times during each juveniles' experiment the germlings attached to tiles were counted by microscopy -in the beginning, the end and two times during the experimental run.
Animals present in the mesocosms were fixed with formaldehyde (4%) at the end of each adults' experiment.They were identified and counted under a stereo microscope.In mesocosms without Gracilaria and in mesocosms with Gracilaria at low density all grazers were quantified, while subsamples (90 ± 36 g FW) were analyzed from mesocosms containing Gracilaria at high density.
Comparison of the grazer densities in single mesocosms at the beginning and at the end of adults' experiments revealed that the intended adjustment of grazer densities was largely effective.However, small isopods proved able to migrate into the mesocosms in summer and Littorina densities sometimes declined due to escape, but mesocosms intended to contain grazers at high and low densities generally still showed this difference at the end of each adults' experiment.However, in adults' experiment 3, 4, 5 and 7 real and intended grazer densities did not fully correspond.Thus mean values of effective grazer densities at the beginning and at the end of each experiment were calculated for each mesocosm and for each of the main grazer genera.Afterwards, the five mesocosms with relatively high and the five mesocosms with relatively low mean grazer density had to be chosen post hoc for each Gracilaria density.The real mean grazer densities in the two juveniles' experiments were calculated based on grazer densities determined in the beginnings and ends of adults' experiments 1 to 3 (which together covered the time window of juveniles' experiment 1, see Figure 1) and adults' experiments 4 and 5 (covering juveniles' experiment 2, see Figure 1), respectively.Real and intended grazer densities fully corresponded in juveniles' experiment 1, while a post hoc distinction of mesocosms with relatively low and relatively high grazer density was necessary in juveniles' experiment 2. Figure 3 shows average grazer densities (i.e.mean values of grazer densities at the beginning and at the end of each experiment in five replicate mesocosms) in the course of the different experiments and in the different Gracilaria treatments for both high and low grazer treatments as assigned by us post hoc.

Data analysis
Computation of half-life times of Fucus germlings Using the "Prism 4.03"-software (GraphPad Software, La Jolla, Ca., USA.), an exponential decay function was fitted by adaptive iteration to the densities of Fucus germlings that were observed at successive countings.The function used was Y=Be (-KX) + A, with Y = germling density at a given time X, A = final germling density, B = maximal decrease of germling density, K = rate constant.In this way the halflife time Z of the germlings on each tile could be estimated, using the function Z = 0.69/K.

ANOVA and post-hoc tests
The effect of Gracilaria density, grazer density and the respective experimental period upon the half-life time of Fucus germlings and the growth of Fucus adults was tested for significant differences, using factorial ANOVA and Tukey's post-hoc test.Prior to statistical tests, all data were rank-transformed, as this transformation best reduced heteroscedasticity.In the adults' data set, homogeneity of variances could not be achieved for the factor "time" (Levine's test, p < 0.05) and results were taken as statistically significant if the probability of type I error was less than 1% (Underwood 1997), while other results were taken as statistically significant at a p-level of 5%.

Seasonal variation in environmental conditions
The environmental conditions followed a typical seasonal pattern (see Figue 1).Low temperatures of < 5°C were characteristic for the winter months, while peak temperatures of 20°C were reached in July and August.Light intensity was already relatively high in spring and slowly increased until June.Lowest light intensities were measured in December and January.In December and January there are only 7 h of daylight compared with 17 h in June and July.

Growth of adult Fucus individuals
Not surprisingly, the time when an experiment was carried out had a significant effect on the growth of Fucus adults (p<0.001;see Table 1 and Figure 4).Growth rates were highest in May/June (experiment 3), whereas the weight of Fucus declined in November/December (experiment 6) and moderate growth rates were achieved in spring and autumn (experiments 1, 2, 4, 5, 7).Also the presence of Gracilaria had a significant effect upon the growth of Fucus adults (Table 1 and Figure 5).The post-hoc test detected that presence of Gracilaria at high density resulted in significantly (p<0.001) less Fucus growth than presence at low density or absence.Grazers alone had no significant effect on the growth of Fucus at the chosen alpha level (p=0.053,Table 1), but ANOVA detected a significant interactive effect of Gracilaria density and grazer density (p<0.01,see Tab. 1 and Figure 5).At low Gracilaria density the presence of high grazer densities significantly reduced Fucus growth, while grazer treatments did not affect growth at other Gracilaria densities.This interactive effect was particularly obvious in April/May (experiment 2; see Figure 4) and from July to December (experiments 4, 5, 6; see Figure 4).

Development of germlings
During the first experimental period from March to June, the average initial density of Fucus germlings on tiles was 231.1 cm -2 , while their average half-life was 1.40 ± 0.07 weeks (mean ± S.E.).The average initial Fucus density during the second experimental period from June to September was lower than in the first period (36.9 germlings cm -2 ).The average half-life time (mean ± S.E.: 0.45 ± 0.06 weeks) was also significantly shorter than in the first experiment (p<0.001;Table 2 and Figure 6).Based on ANOVA results the factor "Gracilaria" also significantly influenced the survival of germlings (p<0.001;Table 2).Half-life time was longest in treatments with low Gracilaria density and shortest in treatments with high Gracilaria density (Figure 7).Grazers had no significant effect upon the half-life time of Fucus germlings (Table 2 and Figure 7).

Discussion
The main purpose of the experiments was to test the two hypotheses (1) that Gracilaria affects Fucus directly and (2) that mesograzers associated with Gracilaria have an impact on Fucus.Both hypotheses cannot be rejected.Gracilaria had an influence on adult Fucus individuals and Fucus germlings.High Gracilaria densities clearly reduced the half-life time of Fucus germlings and growth of Fucus adults as compared to low Gracilaria densities.Grazers only had an effect on Fucus adults at low Gracilaria densities.

Growth of adult Fucus
There was a significant seasonal variation in Fucus growth, as would be expected.Growth rates were highest in May/June when water temperature and light availability were elevated.With low water temperatures and reduced light availability in November/December, Fucus biomass declined.
Fucus adults grew least when Gracilaria was present at high density, but there was no difference in growth when Gracilaria was absent or present at low density.At high density the ground coverage by Gracilaria was 100 %, and one can assume that such conditions result in a reduced availability of light, oxygen and nutrients for Fucus, despite its capacity to protrude Gracilaria mats due to the buoyancy provided by gas filled bladders.Also, an allelopathic effect cannot be excluded.
Grazers reduced the growth of Fucus as well, but the overall effect was relatively small in size (a reduction of growth by approximately 8%) and statistically not significant.It is possible that epiphytes and other fouling organisms may have interfered with the interactions of Gracilaria, Fucus and grazers.Especially in spring and early summer ephemeral algae -first Ulothrix spp.and later tube forming diatoms -were present on mesocosm walls and entangled in Gracilaria and  on Fucus.These algae were also observed in spring and early summer on Gracilaria and on Fucus in nature.Several experiments have shown that many grazers (including Idotea) preferably feed on epiphytes and ephemeral algae, also those growing on Fucus (Karez et al. 2000;Orav-Kotta and Kotta 2004;Worm and Sommer 2000).Preferential consumption of ephemeric algae might thus reduce the negative effect of mesograzers upon Fucus.
The effect of grazers on Fucus differed with Gracilaria density.Treatments with high grazer density contained for both Gracilaria densities (low and high) roughly the double amount of grazers as treatments with low grazer density (see Figure 2 and Figure 3).High Gracilaria densities had the strongest negative impact on Fucus growth rates, irrespective of grazer density.Thus Gracilaria biomass clearly is of higher importance for Fucus growth than the grazers associated with it.Any effects of grazers are probably superimposed by effects of Gracilaria in high density.In contrast, at low Gracilaria density an increased density of grazers resulted in decreased Fucus growth.This  was not the case in absence of Gracilaria, where Fucus growth was generally high and again unaffected by grazers.Fucus in Gracilaria free environments is occasionally heavily grazed (Dethier et al. 2005;Engkvist et al. 2000), but the natural densities of grazers associated with Fucus during the experimental periods were too low to cause a significant decline in Fucus growth.Grazers thus affected growth only at low Gracilaria density.This indicates that the grazer population accommodated by Gracilaria poses an additional indirect threat to Fucus.However, the effect only gets visible when the direct effects of Gracilaria are relatively small.

Development of germlings
The early live stages of Fucus are generally especially vulnerable (Berger et al. 2003).In our experiments, germlings were found to have the shortest half-life time when Gracilaria was present at high density.This effect was especially pronounced in the second experimental period where none of the Fucus germlings survived at high Gracilaria treatments -irrespective of grazer density.Buried under thick Gracilaria mats, the germlings presumably suffered from a reduced availability of light and oxygen.However, both generations of Fucus germlings also tended to have a slightly longer half-life time in presence of Gracilaria at low density than in absence of Gracilaria, suggesting that Gracilaria may have both positive and negative effects upon the survival of Fucus germlings.In other experiments, low concentrations of Pilayella littoralis-exudates had a positive effect on the germination of Fucus zygotes (Råberg et al. 2005).This effect was attributed to the enhanced nutrient concentration in the exudates, which probably supported the germlings.Possibly, Fucus germlings in mesocosms with low Gracilaria density also profited from a slightly raised nutrient concentration, but other explanations can also not be excluded.

Ecological consequences and outlook
With some restrictions, the results of our mesocosm experiments can be compared to natural processes.The densities of both Gracilaria and grazers were based on those found in nature.Also the water depth was similar to that in nature, as both species occur mainly in shallow water in the Kiel Fjord.
Growth of Gracilaria in the southwestern Baltic is limited by light and largely restricted to summer and water depths above 3m (Weinberger et al. 2008).With Fucus appearing in similar depths of water, there could be a strong competition for space and light between the two species (Nyberg 2007).Based on our studies, Gracilaria appears to be able to restrict growth of Fucus in these shallow waters.Especially in sheltered bays with little water fluctuation and where high densities of Gracilaria can amass, the effect of Gracilaria on Fucus could be detrimental, even though it will presumably not lead to the extinction of single populations.The further effects, however, depend very strongly upon how Gracilaria influences the settlement and development of Fucus germlings.Our experiments show that in dense Gracilaria mats with numerous grazers, the survival of Fucus germlings approaches zero.The local density of Gracilaria in nature therefore is a crucial factor that determines which effect the presence of Gracilaria actually has.Other factors such as deposited matter trapped in algal mats have also to be taken into consideration in order to estimate the overall impact of Gracilaria upon Fucus.Furthermore, it has to be held that the results of the experiments conducted cannot be applied to the entire Baltic Sea without reservations, but first and foremost to the western Baltic Sea with its special conditions of salinity, light and temperature.
Currently, Gracilaria appears in the German Baltic Sea mainly in its unattached form.However, the species seems to be capable of sexual reproduction at relatively low salinities (own observations).Should Gracilaria recruits settle on hard substrate, the competition for space with Fucus germlings might increase even further.Space occupation by other algae has been shown to have a strong negative effect on the settling success of spores and zygotes (Vadas et al. 1992), also those of Fucus ssp.(Berger et al. 2003;Råberg et al. 2005;Steen 2004;Worm and Chapman 1996).In the long run, a rise of water temperatures as a consequence of global warming might intensify the competitive situation in the Baltic Sea (Leppäkoski et al. 2002;Stachowitz et al. 2002), because Gracilaria, unlike Fucus, grows even better at higher temperatures (Yokoya et al. 1999).
Gracilaria appears to show an exceedingly strong potential of changing the habitat in which it grows.It alters the habitat architecture by forming thick mats.Thereby it works as a trap for sediment and other particles and shades other algae (Nyberg 2007).
On the other hand, Gracilaria in small amounts adds structural complexity to a relatively homogeneous system.It sustains a large taxonomic richness and provides a new habitat for many taxa (Nyberg 2007;Thomsen 2010).Our experiments were conducted in an environment of relatively low diversity (Rönnbäck et al. 2007), but 18 different animal taxa were nonetheless found in Gracilaria (data not shown).Thomsen et al. (2012) found out that Gracilaria comosa had ambiguous effects in its new environment in Australia, as it inhibited the native seagrass Halophila ovalis itself, but facilitated most seagrass-associated invertebrates.Accordingly, the question should be asked, whether Gracilaria, which in fact harbours a high biodiversity, could be an adequate replacement for Fucus in the Western Baltic Sea if the decline of the latter continues or whether consequences of a further spread of the invader would rather be detrimental for the system.Kraufvelin and Salovius (2004) stated that substitution of Fucus by the filamentous green alga Cladophora glomerata does not immediately affect macroinvertebrate diversity negatively.Other authors also found out that following a decline of Fucus the majority of I. balthica switched to an alternate habitat (Orav-Kotta and Kotta 2004).Weinberger et al. (2008) postulated a possible functional replacement of Fucus by Gracilaria and a gain of habitat for seaweed associated organisms on soft bottom substrates.However, while animal abundances and biomass are higher in Gracilaria as compared to Fucus -probably due to the finer morphology of Gracilaria -epiphytic algae and sessile epiphytic animals are generally more abundant on fucoids and may lose substrates if Fucus is replaced by Gracilaria.Thereby, the habitat complexity may be reduced and energy fluxes may be affected (Kraufvelin and Salovius 2004).A decline of Fucus might for similar reasons be harmful to the stability of the whole community (Torn et al. 2006).Moreover, high densities of algal mats stress benthic communities directly, especially by creating oxygen deficiency (Arroyo et al. 2012;Berezina and Golubkov 2008;Norkko and Bonsdorff 1996;Norkko et al. 2000) and even night-time anoxia caused by rapid algal growth in the summer (e.g.Shen et al. 2008).In conclusion, opportunistic and mobile taxa may benefit from algal mats consisting of Gracilaria, but stenoecious key species adapted to Fucus may be lost, and fundamental changes to community function and food web might be induced.

Figure 1 .
Figure 1.Development of mean water temperature in the Kiel Fjord and of mean and maximal photosynthetically active radiation (PAR) during the experiments.Also shown are the different time periods for the subsequent experiments.

Figure 2 .
Figure 2. Experimental set-up.Three Gracilaria vermiculophylla densities (without, high and low) were combined with two grazer densities (low and high).

Figure 4 .
Figure 4. Fucus vesiculosus, adults.Growth in mesocosms with high or low Gracilaria vermiculophylla density or without Gracilaria, and with low or high grazer densities.Single results of seven successive experiments, n = 5.Different letters in the graph titles indicate the experiments that were significantly different in a Tukey-test (p < 0.01).

Figure 5 .
Figure 5. Fucus vesiculosus, adults.Growth in mesocosms with high or low Gracilaria vermiculophylla density or without Gracilaria, and with low or high grazer densities.Mean of seven successive experiments ± 95% CI, n = 35.Different lower-case letters indicate Gracilaria treatments that were significantly different in a Tukey-test (p < 0.05), while upper-case letters indicate differences in the interactive effect of Gracilaria and grazers (p < 0.05).

Figure 6 .
Figure 6.Fucus vesiculosus, germlings.Mean half-life time in mesocosms with high or low Gracilaria vermiculophylla density or without Gracilaria and with low or high grazer densities.Single results of two successive experiments, n = 5.Different letters in the graph title indicate that the experiments were significantly different in a Tukey-test (p < 0.05).

Figure 7 .
Figure 7. Fucus vesiculosus, germlings.Mean half-life time in mesocosms with high or low Gracilaria vermiculophylla density or without Gracilaria and with low or high grazer densities.Mean of two successive experiments ± 95% CI, n = 10.Different letters indicate treatments that were significantly different in a Tukeytest (p < 0.05).

Table 1 .
Growth of Fucus vesiculosus adults.Effect of time when an experiment was conducted, Gracilaria vermiculophylla density and grazer density in a fully factorial ANOVA.

Table 2 .
Half-life time of Fucus vesiculosus germlings.Effect of time when an experiment was conducted, Gracilaria vermiculophylla density and grazer density in a fully factorial ANOVA.