Eutrophication-induced changes in benthic algae

affect the behaviour and fitness of the marine

amphipod Gammarus locusta

Patrik Kraufvelin a,*, Sonja Salovius a, Hartvig Christie b,Frithjof E. Moy b, Rolf Karez c, Morten F. Pedersen d

a Abo Akademi University, Environmental and Marine Biology, Akademigatan 1, FIN-20500 Turku/Abo, FinlandbNorwegian Institute for Water Research (NIVA), P.O. Box 173 Kjelsaas, N-0411 Oslo, Norway

c Institut fur Meereskunde, Experimentelle Okologie, Dusternbrooker Weg 20, D-24105 Kiel, GermanydDepartment of Life Sciences and Chemistry, Roskilde University, P.O. Box 260, DK-4000 Roskilde, Denmark

Received 11 November 2004; received in revised form 12 August 2005; accepted 31 August 2005

Abstract

This study, conducted in mesocosms, natural field sites, and in laboratory aquaria, showed that eutrophication altered the nutrient status and

dominance patterns among marine macroalgae, which in turn, stimulated gammaridean density. Gammaridean abundance correlated positively

with both nutrient addition and the amount of green algae (also stimulated by nutrient enrichment). Path analysis indicated that the direct effect of

nutrients on gammaridean density was of less importance than the indirect effect through increased production of green algae. In cage colonisation

experiments, either in the field or in a control mesocosm kept under ambient nutrient conditions, more gammarids colonised nutrient enriched algae

(E-algae) than algae with ambient nutrient levels (A-algae). Gammarus locusta generally grew faster on nutrient enriched algal specimens and

when reared on green rather than on brown algae (fucoids). The nutrient status of periphytic algae did not affect gammaridean growth significantly,

but the number of egg-carrying females (and thus egg production) was significantly higher among gammarids reared on E-periphyton. The

gammaridean habitat preference order (red > green > brown > periphyton) was almost the reverse of their growth rate in feeding assays

(periphyton > green > brown). This implies that macroalgae may be more important as a habitat than as a food source for these animals, which

then have to become mobile in search of optimal food items. In this process, algal nutrient content was important as the gammarids in our study

actively chose high quality nutrient-rich food, which, in addition, increased their fitness. Stimulated growth rates and egg production may

ultimately lead to population increase, which, combined with the preference for high nutrient food items may dampen the initial effect of nutrient

enrichment (i.e. blooms of green macroalgae) in shallow coastal waters.

# 2005 Elsevier B.V. All rights reserved.

Keywords: Coastal eutrophication; Nutrient enrichment; Rocky shore; Mesocosm; Grazing; Plant–animal interaction

www.elsevier.com/locate/aquabot

Aquatic Botany 84 (2006) 199–209

1. Introduction

Temperate rocky shore ecosystems are typically dominated

by perennial macroalgae, which provide suitable habitats and

food resources for a large number of benthic animals. Among

these animals, gammaridean amphipods are often very

abundant and conspicuous and are capable of imposing

structuring effects on the algal substrate (Edgar, 1983; Pavia

et al., 1999). These effects are either beneficial when the

* Corresponding author. Tel.: +358 405727409; fax: +358 2 2153428.

E-mail address: patrik.kraufvelin@abo.fi (P. Kraufvelin).

0304-3770/$ – see front matter # 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.aquabot.2005.08.008

mesograzers are mainly feeding on epiphytic algae (while

releasing nutrients that may stimulate further plant growth), or

detrimental when considerable amounts of the perennial host

algae are also consumed (Shacklock and Doyle, 1983; Duffy,

1990; Karez et al., 2000; Worm et al., 2002). Plant–animal

interactions in marine ecosystems are rather poorly understood

compared to those in freshwater habitats (e.g. Jones et al., 1999;

Persson et al., 2001). It is, for example, unclear why most

mesograzers only consume a fraction of the seaweeds available

in their habitats as potential food (Duffy and Hay, 1991;

Granado and Caballero, 2001). Furthermore, experimental data

on the distribution, abundance, and feeding of mesograzers

inhabiting seaweeds of various chemical, structural, or

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209200

nutritional traits are available for only a few species and bio-

geographical areas (Pavia et al., 1999; Hemmi and Jormalainen,

2002; Cruz-Rivera and Hay, 2003). In general, the habitats of

crustacean mesograzers are closely tied to their selection of

diets, but the algal species that provides the best protection is

perhaps not the one with the highest nutritive values (Nicotri,

1980; Buschmann, 1990). Laboratory experiments have shown

how diets of different quality affect consumer fitness (Costa

et al., 1996; Hemmi and Jormalainen, 2002), and how consumer

fitness among amphipods in particular may increase on mixed

diets including additions of animal matter (Costa et al., 1996;

Cruz-Rivera and Hay, 2000a,b).

Coastal ecosystem deterioration is one of the most serious

threats to the marine environment (Gray et al., 2002) and

functional shifts caused by structural changes are to be

expected, for example in connection with eutrophication

(Cloern, 2001). The effects of eutrophication have, however,

been investigated far less in marine environments than in

freshwater systems (Smith, 2003). On rocky shores, coastal

eutrophication increases the growth of annual filamentous

algae, which may retard perennial species (Pedersen and

Borum, 1996; Schramm, 1999), known to be very important

habitats for macrofauna (Benedetti-Cecchi et al., 2001; Cloern,

2001), but see also Edgar et al. (2004) and Kraufvelin and

Salovius (2004). Although primary changes in the abundance

and nutrient content of dominating algae due to nutrient

enrichment may cause direct secondary nutritional and habitat

changes for the fauna in a bottom-up manner (Edgar, 1990;

Eggleston et al., 1999; Kraufvelin et al., 2002), the responses to

nutrient enrichment on rocky shores do not always occur

linearly (e.g. Dye, 1998; Lopez-Rodriguez et al., 1999; Karez

et al., 2004). Several authors have focussed instead on top-down

processes, by recognising the possible role of grazers in

buffering the primary effects of eutrophication (Geertz-Hansen

et al., 1993; Hillebrand et al., 2000;Worm et al., 2000; Balducci

et al., 2001), but see also Russell and Connell (2005) for

contrasting results. Many other factors such as the degree of

wave exposure, water exchange rates, water currents, weather

conditions, shore profiles, the successional stage of the plant

community, and the amount of predators contribute to structure

the community (e.g. Worm and Sommer, 2000; Duggins et al.,

2001; Bokn et al., 2003), making the full pattern of

eutrophication effects difficult to predict.

A long-term study on eutrophication effects on experimental

rocky shore communities, the EULIT-project, did not reveal

many negative effects on either the algal (Bokn et al., 2002,

2003; Karez et al., 2004) or animal assemblages (Kraufvelin

et al., 2002; Christie and Kraufvelin, 2004). These findings led

to several interesting questions. Firstly, in the absence of vast

community effects, do single animal populations respond

positively, e.g. with increased abundances, to the improved

nutritive conditions (increased N and P contents) and increased

amounts of green algae that has been observed previously in

nutrient enriched mesocosms? Secondly, would these animals

be able to find food sources of higher nutrient content when

offered both nutrient enriched plant turfs and plant turfs of

ambient nutrient levels? Finally, is it possible to get

experimental evidence of direct improvements in animal

fitness due to certain food preferences?

To answer these questions about plant–animal interactions

both descriptive and experimental techniques were used. The

cover of green macroalgae and the population development of

the dominant mesograzer group Gammarus spp., mainly

Gammarus locusta L., were monitored over time in meso-

cosms, but G. locusta was also used for a series of colonisation

and growth experiments. To clarify if G. locusta actively chose

certain algal species or preferred algae of a higher nutritive

quality when offered algal specimens of either ambient nutrient

concentration (A-algae) or nutrient enriched algae (E-algae),

colonisation experiments were set up under both natural field

conditions and in outdoor mesocosms. Finally, nutrient

enriched food (E-macroalgae and E-periphyton) was tested

to investigate its potential to stimulate individual and

population growth of gammarids in laboratory aquaria. The

primary hypotheses tested were: (1) With a possible shift to

more nutrient-rich and digestible green algae, the density of

Gammarus spp. may change. (2) More gammarids will choose

(colonise) E- than A-algae of the same algal species at ambient

nutrient concentrations (but not at nutrient enriched condi-

tions). (3) When offered green and brown macroalgae and

periphyton (E- and A-algae) under controlled aquarium

conditions, G. locusta will grow faster on E- than on A-algae.

(4) G. locusta will also potentially grow faster on green algae

and periphyton compared to brown algae due to higher nutritive

quality of green algae and/or the presence of higher levels of

secondary metabolites (grazing deterrents) in brown algae.

2. Materials and methods

All experiments were conducted at the Marine Research

Station Solbergstrand by the Oslofjord (598370N, 108390E) inSE Norway. Table 1 provides an overview of the measurements

and experimental manipulations including the statistical

techniques used for data analysis. Experimental algae (for

field, mesocosm and aquarium experiments) were always

sampled from mesocosms, either from control mesocosms

receiving un-manipulated fjord water (i.e. ambient nutrient

levels, A) or from nutrient enriched (E) mesocosms treated with

32 mM inorganic nitrogen (N) and 2 mM inorganic phosphorus

(P) above background levels. In the EULIT project, nutrients

were added to the six treatment mesocosms during the 1998–

2001 period along a geometrical gradient corresponding to 1, 2,

4, 8, 16 and 32 mmol l�1 N, including 0.06, 0.12, 0.25, 0.5, 1.0

and 2.0 mmol l�1 P above background levels (two basins served

as controls without extra nutrient addition). Nutrients were

added as a mixture, which for the highest nutrient addition level

(32 mmol N l�1/2 mmol P l�1) consisted of 14.3 mol N as

NH4NO3 and 0.9 mol P as H3PO4 with a N/P mol ratio of

16/1.

The Solbergstrand mesocosms had a water volume of 6–

12 m3 depending on tidal level and they received water at a rate

of 5 m3 h�1 (mean water residence time of about 2 h) from 1 m

depth in the Oslofjord. A tidal regime simulated natural

changes in water level reflecting the local tidal amplitude of

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209 201

Table 1

Description of the Gammarus locusta related data sets at Solbergstrand 1997–2001

Description Algal species Measured variables Design, statistics Date, duration

Mesocosm monitoring:

green algal cover and

Gammarus population

dynamics over time

Green algae,

artificial substrates

Algal cover in %,

gammaridean abundance

Regression analysis, path

analysis (N loading, green

algal cover, gammaridean

density = path)

September 1997–

August 2001,

4 years

Mesocosm experiment:

gammaridean

colonisation to

caged macroalgae

(ambient, A- or

enriched E-algae)

Ulva lactuca,

Fucus serratus,

Rhodomela

confervoides

/Ceramium spp.

G. locusta abundance, N

and P content of algae

Two-way factorial ANOVA:

design for G. locusta = 3 algal

groups � 2 origin � 3 replicates,

residual d.f. = 12, total d.f. = 17;

design for nutrients = 2

mesocosms � 2 origins

� 3 replicates, residual

d.f. = 8, total d.f. = 11

August 2000, 60 h

Field experiment:

gammaridean

colonisation to

caged macroalgae

(A- or E-algae)

U. lactuca G. locusta abundance, N

and P content of algae

Two-way factorial

ANOVA: design = 2 locations �2 origins � 4 replicates,

residual d.f. = 12, total d.f. = 15

August 2003, 60 h

Aquarium experiment:

gammaridean growth

when fed with A-

or E-macroalgae

U. lactuca,

F. serratus

G. locusta length growth of

juveniles, N and P content of algae

Two-way factorial ANOVA:

design = 2 algal groups � 2

nutrient levels � 3 replicates,

residual d.f. = 8, total d.f. = 11

September 1999,

4 weeks

Aquarium experiment:

gammaridean growth

and reproduction

when fed with A- or

E-periphyton algae

Periphyton

communities

pre-exposed

in mesocosms

on ceramic tiles

G. locusta length growth

of juveniles and number of

egg carrying females

One-way ANOVA August 2000,

4 weeks

36 cm, and comprising two high tides and two low tides

diurnally, while a wave machine generated constant wave

action (17 strokes per minute). Using a data logger, oxygen

concentration, water temperature, and salinity were continu-

ously monitored, whereas water samples for nutrient analysis

were collected on an average weekly basis from both water

intakes and outlets of each mesocosm (Bokn et al., 2001, 2002,

2003). Living communities were introduced in 1996 by

transplanting small boulders from the Oslofjord, with macro-

algae and associated animals attached, onto concrete steps in

each mesocosm. These steps represented different water depths

on a shoreline and consisted of two intertidal and two subtidal

steps per mesocosm. After the initiation phase, natural

community development was allowed to contribute to the

flora and fauna assemblages. Over the sample period, over 40

species of macroalgae and 80 species of macrofauna were

identified in the mesocosms. Brown algae Fucus serratus L.,

Fucus vesiculosus L. and Ascophyllum nodosum (L.) Le Jol, the

green algae Ulva lactuca L. and some seasonal red algae

dominated the macroalgal assemblages. Amphipods and

isopods were the most common animal groups.

2.1. Mesocosm monitoring of green algae and Gammarus

spp.

Green algal cover and Gammarus spp. abundance were

measured two to five times each year, but for comparative

purposes, only measurements from late summer/autumn each

year were used. Algal cover was measured in 16 quadrats

(42 cm � 42 cm, each divided into 25 sub-quadrats) forming a

fixed grid system in each basin. The amount of Gammarus spp.

and othermobile animals inmesocosmswere estimated using the

colonisation of artificial substrates. Gammarus spp. was chosen

as study object for this paper, since the group was very abundant

and at times constituted up to 80% of the total biomass of mobile

macrofauna (Christie and Kraufvelin, 2004). The artificial

substrates for animal sampling consisted of three 80 cm long

ropes, one stone, and one Petri dish, tied tightly together. These

substrates were able to collect an intermediate fraction of the

animals normally present in the most dominant algal types, i.e.

green, red, and brown algae (Kraufvelin et al., 2002). Three

replicate substrates (placed behind algae covered boulders) at

two different depths/steps (one intertidal and one subtidal step

just below the tidal zone) were exposed for 2 days in each basin.

Since the main interest lay in making comparisons between

mesocosms, the total number of six pooled artificial substrates

was used as one mesocosm measurement.

The green algal cover andGammarus spp. abundance versus

nutrient addition 1998–2001, as well as Gammarus density

versus green algal cover was analysed by regression analysis.

These analyses were extended into a path analysis (Bryman and

Cramer, 1990; Everitt and Dunn, 1991) to reveal the relative

importance of the direct effects of nutrients on Gammarus

density as well as the indirect effects through changes in the

cover of green algae.

2.2. Gammaridean colonisation of mesocosm macroalgae

To test the colonisation/preference hypothesis in mesocosms

green (U. lactuca), brown (F. serratus), and red algae

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209202

(a combination of Ceramium spp. and Rhodomela confervoides

(Huds.) P.C. Silva) were collected from one control mesocosm

and from one nutrient enriched mesocosm (the highest nutrient

addition level) in August 2000. Animals and visible epiphytes

were removed from the algae by carefulwashing andbyhand and

the algal turfs (each consisting of 5–20 g DW) were placed in

translucent PVC tubes enclosed with coarse nets (5 mm mesh

size) at each end. Three replicates of green, brown, and red algae

for both nutrient levels were re- and cross transplanted into the

two mesocosms to allow recolonisation of animals to take place

over a 60 h period (experiment start in the evening, sampling in

the morning after two full days). After termination of the

experiment, the animals were collected and preserved in 70%

ethanol for qualitative and quantitative analysis. The dry weight

biomass of each algal samplewas registered after drying for 24 h

at 100 8C. Total C and N content of the algae were determined

using a Carlo-Erba NA-1500 elemental analyser, while the P

content was measured after wet oxidation with boiling H2SO4

followed by spectrophotometric analysis (Strickland and

Parsons, 1968). The animal community structures of the caged

algae were also compared to the community structures of

adjacent algal-animal assemblages in the twomesocosms. These

samples of attached algae were obtained by random sampling

where the algaewere carefully enclosed in a plastic bag and cut at

the algal base. Three attached algal samples of each type were

taken from each mesocosm.

After checking for normality with Kolmogorov–Smirnov’s

test and homogeneity of variances with Levene’s test, the data

were log-transformed ln(x + 1), and abundance values of G.

locusta analysed by two-way factorial ANOVAs with algal

group (green, brown, and red) and algal origin (control or

nutrient enrichment) as fixed orthogonal factors (design = 3

algal groups � 2 origins � 3 replicates, total d.f. = 17). When

appropriate, a Bonferroni test was applied for a posteriori

comparisons of means (Hochberg, 1988; Day and Quinn,

1989). The ANOVAs were run separately for each mesocosm.

Similar experimental designs with both control (A-algae) and

treatment (E-algae) cages were thus applied to both meso-

cosms, such that each mesocosm became a separate case study

(one situation at ambient nutrient levels and one at a nutrient

enriched situation). Differences in algal N and P content were

analysed using two-way factorial ANOVAs. Here, both the

mesocosms (to which we transplanted the algae) and algal

origin were considered as fixed orthogonal factors with two

levels (design = 2 mesocosms � 2 origins � 3 replicates, total

d.f. = 11). Four separate statistical analyses were run on algal

nutrient content, one for each algal species (U. lactuca, F.

serratus, R. confervoides, Ceramium spp.).

2.3. Gammaridean colonisation of U. lactuca in the field

The field experiment was carried out in August 2003, more

than 2 years after the termination of the EULIT-project, and 1

year after the termination of nutrient addition (dosing ended in

September 2002). Thus, 2 weeks of renewed nutrient dosage to

the mesocosms were required in order to get algae (method only

applicable forU. lactuca) saturated by nutrients for experimental

transplantation. Eight U. lactuca turfs from one control

mesocosm and eight turfs from one nutrient enriched mesocosm

were sampled. The algae were carefully washed and the last

animals removed by hand. Algae (1–2 g DWA- or E-algae per

cage)were placed in cagesmade by folded and closed nets with a

mesh size of 7 mm. Two field sites in close vicinity of the

Solbergstrand station were used for the transplantation. At each

site, four cages containing A-algae and four cages containing E-

algae were tied to a rope and all samples anchored to the bottom

in a randomorder (samples placed at awater depth of ca. 0.5 m at

low tide). After 60 h, the cages (experimental start during

daytime and samplingduringdark night hours)were sampled and

the colonising animals separated from the algae. The animals

were preserved in 70% ethanol for qualitative and quantitative

analysis and the algae cleaned and dried before weighing and

analysing tissue C/N/P content. The number of G. locusta was

recorded and analysed together with the algal nutrient content by

two-way factorial ANOVAswith location (south and north of the

station) and algal origin (control or nutrient enrichedmesocosm)

as fixed orthogonal factors (design = 2 locations � 2 origins � 4

replicates, total d.f. = 15).

2.4. Aquarium experiment—gammaridean growth on

macroalgae

To study the growth of juvenile gammarids onU. lactuca and

F. serratus taken from different nutrient regimes (A-algae pre-

exposed at ambient nutrient levels in control mesocosms and E-

algae from a nutrient enriched mesocosm), a no-choice

aquarium experiment was run in late summer 1999 using

three replicates of each treatment. To each aquarium were

added 55 randomly chosen newly born G. locusta (mean length

1.3 mm) and the algae were introduced (three replicate aquaria

of each treatment). Each week the algaewere replaced with new

fresh specimens from the respective mother mesocosms. After

4 weeks, the body length of each individual was measured

under a microscope. The growth of G. locusta was analysed

using a two-way factorial ANOVAwith algal class (green and

brown) and algal origin (control or nutrient enriched

mesocosm) as two fixed orthogonal factors (design = 2 algal

groups � 2 nutrient levels � 3 replicates, total d.f. = 11).

2.5. Aquarium experiment—gammaridean growth on

periphyton

In August 2000, the growth of juvenileG. locusta offered A-

or E-periphyton algae during 4 weeks in laboratory aquaria was

also studied. Periphyton was obtained by exposing small

ceramic tiles (5 cm � 5 cm) vertically on holding devices that

precluded the access of crawling mollusc grazers but allowed

access to swimming crustaceans. To obtain A- and E-

periphyton, tiles were exposed in mesocosms with 0 and

32 mmol N l�1 addition, respectively, for 2 weeks before their

use as food for gammarids. Five replicate aquaria with A-

periphyton and five replicates with E-periphyton (pre-exposed

in the respective mesocosms) were used. To each aquarium

were added 25 juvenile G. locusta (mean length 1.3 mm). To

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209 203

Fig. 1. Direct effects in path analysis: (a) percent cover of green algae and (b)

density of Gammarus locusta on artificial substrates in relation to N addition

levels (mM) during years 1997–2001. Note that all samples from 1997 (pre-

dosing, open circles) are on x = 0.

maintain the nutritive differences between the aquaria the

periphyton tiles were replaced every week with new ones from

the respective mesocosms. Four periphyton tiles were placed

per aquarium on each occasion. During the course of the

experiment, the animals grew so fast that they started to

reproduce. Therefore the number of egg-carrying females was

also included as a study variable and this dataset was analysed

by one-way ANOVA with d.f. 1,8.

3. Results

3.1. Mesocosm monitoring of green algae and Gammarus

spp.

The cover of green algae, consisting mainly of U. lactuca

and filamentous green macroalgae on rocks and as epiphytes on

brown macroalgae, increased significantly along the nutrient

addition gradient both when analysed separately for each year

(Table 2) and when all years were pooled together (Fig. 1a).

Control mesocosms had consistently higher degrees of green

algal cover during 1998–2001 compared to the background

year 1997 (Fig. 1a). The abundance of Gammarus spp., on the

other hand, did not increase significantly with increasing

nutrient enrichment when all samples were pooled and analysed

together (Fig. 1b). When data from different years were

analysed separately, however, regressions were significant for

1998 and 2001 (Table 2). Gammarus abundance was further

significantly related to the cover of green algae (Fig. 2). Finally,

path analysis showed that the direct effect of nutrients on

Gammarus density (0.31) was smaller than the indirect effects

of nutrients via green algae: 0.73 � 0.62 = 0.45, where the

former factor stands for nutrients to green algae and the latter

factor for green algae to Gammarus. The total effect

(direct + indirect) of nutrient enrichment onGammarus density

was thus 0.31 + 0.45 = 0.76.

3.2. Gammaridean colonisation of mesocosm macroalgae

In the control mesocosm, significantly more G. locusta

colonised E-algae thanA-algae (Table 3a). For individual algal

Table 2

Regression analysis of green algal cover andG. locusta abundance in relation to

nutrient addition 1998–2001 (d.f. always = 1,6)

Factor %TSS p

Green algae 1998 84 0.001

Green algae 1999 76 0.005

Green algae 2000 55 0.035

Green algae 2001 52 0.044

Gammarus 1998 57 0.031

Gammarus 1999 42 0.081

Gammarus 2000 3 0.697

Gammarus 2001 65 0.016

Given are the percent of the total sums of squares contributed by a factor

(%TSS) as an indicator of the variance explained by that factor and the level of

significance ( p-value). For the %TSS to be meaningful, Type I or additive sums

of squares were used.

species, however, these differences were significant only

for the brown algae in post hoc tests with Bonferroni cor-

rection. The origin of algae had, in contrast, no effect on

gammaridean colonisation in the nutrient enriched mesocosm

Fig. 2. Indirect effect in path analysis: the relationship between percent cover of

green algae in mesocosms and the density of G. locusta on artificial substrates

during 1997–2001.

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209204

Table 3

Two-way factorial ANOVA of differences in abundance of G. locusta among

three algal groups (caged green, brown and red algae) of two different nutrient

origins (ambient and nutrient enriched)

Source d.f. %TSS p

(a) Control mesocosm (ambient, background)

Algal class 2,12 78 <0.001

Origin 1,12 9 0.006

Algae � origin 2,12 2 0.333

(b) Nutrient enriched mesocosm (addition of 32 mMDIN and 2 mMDIP above

background)

Algal class 2,12 80 <0.001

Origin 1,12 4 0.067

Algae � origin 2,12 4 0.181

Data was transformed by ln(x + 1) to homogenise variances.

Fig. 3. Abundance (�S.E.) of G. locusta per g DW for three species/groups of

algae, caged or attached in the mesocosms Caged A-algae (from control A-

mesocosm of ambient nutrient levels), caged E-algae (enriched in high nutrient

E-mesocosm) and attached algae were investigated in both an A-mesocosm

(control) and E-mesocosm (nutrient enriched).

(Table 3b). The amount ofG. locusta also differed significantly

among algal typewith a clear preference for red algae followed

by green and then brown (Fig. 3, Table 3a and b). These

differences were significant in both types of mesocosms in post

hoc tests with Bonferroni correction. The abundance of G.

locustawas always lower on attached mesocosm algae than on

caged algae of the same species (Fig. 3), and the abundance of

G. locusta was higher in the nutrient enriched mesocosm

compared to the control mesocosm (both for caged and

attached algae). The differences betweenmesocosms could not

be tested formally, however, due to the lack of true independent

mesocosm replicates.

Nutrient analyses of caged mesocosm algae revealed that

N and P levels in F. serratus and R. confervoides originating

from the nutrient enriched mesocosm were significantly

higher than in algae originating from the control mesocosm

and, that only small changes occurred in these transplanted

specimens during the 60 h experimental period (Fig. 4,

Table 4). By comparison, the N and P content of U. lactuca,

increased significantly in A-algae transplanted to the nutrient

enriched mesocosm and decreased in E-algae transplanted to

the control mesocosm, although the nutrient content

remained significantly different between the two nutrient

origins (Fig. 4, Table 4). The red algal species, Ceramium

spp. exhibited no differences in nutrient content due to algal

origin for either N or P. The carbon (C) content did not vary

systematically among treatments for any of the four algal

species examined.

3.3. Gammaridean colonisation of U. lactuca in the field

G. locusta clearly preferred E-Ulva above A-Ulva in the

field (Fig. 5, Table 5). In addition, there were significant

differences in the colonisation byG. locusta between locations,

but no interaction effect among location and algal nutrient

content was evident. The N and P content in E-Ulva were also

significantly higher than in A-Ulva, but a significant interaction

between the location and origin for both N and P showed that

the final nutrient content of the algae differed upon location

(Table 5).

3.4. Aquarium experiment—gammaridean growth on

macroalgae

The aquarium experiment on growth ofG. locusta juveniles

that were fed U. lactuca and F. serratus of different nutrient

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209 205

Fig. 4. Ulva lactuca, Fucus serratus, Ceramium spp. and Rhodomela con-

fervoides: N and P content of caged A- and E-algae (in percentage of algal DW,

means � S.E.) after 60 h exposure in an A-mesocosm (control) and E-meso-

cosm (nutrient enriched).

Fig. 5. G. locusta: number of individuals per g DW of U. lactuca in the field

experiment (mean � S.E.).

Table 5

Two-way factorial ANOVA on differences in abundance ofG. locusta and N and

P content (% of algal DW) in U. lactuca of two different nutrient origins

(ambient mesocosm, A and enriched mesocosm, E) transplanted to two field

sites, d.f. always = 1,12

Source Gam. N % P %

%TSS p %TSS p %TSS p

Location 85 <0.001 5 <0.001 <1 0.326

Origin 6 0.014 90 <0.001 83 <0.001

Location

� origin

<1 0.307 3 0.001 4 0.013

origin also rendered clear and significant differences

(Table 6). The mean body length was significantly higher

for specimens reared on E-algae (7.8 � 0.4 mm for E-Ulva

gammarids and 5.0 � 0.4 mm for E-Fucus gammarids, n = 3

aquaria) compared to those individuals reared on A-algae

(6.1 � 0.2 mm on A-Ulva and 3.8 � 0.1 on A-Fucus).

Gammarids fed U. lactuca also grew significantly bigger

than those offered F. serratus (Table 6). The nutrient analyses

(means for September 1999) showed that there were higher N

and P contents in E-Ulva (N = 4.11%, P = 0.21% of DW)

compared to A-Ulva (N = 3.41%, P = 0.11%) as well as in E-

Fucus (N = 3.46%, P = 0.63%) compared to A-Fucus

(N = 2.30%, P = 0.12%). Interestingly, the N content was as

high in E-Fucus as in A-Ulva (and the P content almost six

Table 4

Two-way factorial ANOVA on differences in macroalgal nutrient content in the m

Source Ulva Fucus

%TSS p %TSS

(a) N content

Mesocosm 49 0.006 2

Origin 22 0.035 92

Mesocosm � origin <1 0.627 <1

(b) P content

Mesocosm 60 <0.001 3

Origin 13 0.032 86

Mesocosm � origin 13 0.031 1

times higher), but even so, the gammarids were on average

1 mm longer in the aquaria with A-Ulva.

3.5. Aquarium experiment—gammaridean growth on

periphyton

The growth of G. locusta also seemed to be stimulated in

aquaria receiving E-periphyton compared to aquaria receiving

esocosm cage experiment, d.f. always = 1,8

Ceram. Rhodo.

p %TSS p %TSS p

0.097 5 0.414 1 0.499

<0.001 32 0.068 80 <0.001

0.905 5 0.410 <1 0.950

0.137 13 0.222 <1 0.550

<0.001 23 0.135 86 <0.001

0.305 <1 0.779 <1 0.654

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209206

Table 6

Results from a two-way factorial ANOVA on differences in growth rate of

juvenile G. locusta (mean body length) fed with U. lactuca and F. serratus of

different nutrient origin for 4 weeks, d.f. always = 1,8

Source %TSS p

Algal species 70 <0.001

Origin 21 0.002

Algae � origin <1 0.384

A-periphyton, although not significantly so (one-way ANOVA;

d.f. = 1,8; %TSS = 36; p = 0.066). The growth of G. locusta

over 4 weeks was much faster when based on periphyton (9.4–

10.3 mm) than when based on U. lactuca (6.1–7.7 mm) and F.

serratus (3.8–5.0 mm). The number of egg carrying females

was stimulated significantly (one-way ANOVA; d.f. = 1,8;

%TSS = 50; p = 0.022) in aquaria where animals were reared

on E-periphyton.

4. Discussion

All the mesocosm, field and aquarium experiments support a

conclusion of responses inG. locusta to elevated nutrient content

of algae and shifts to increased green algal cover with increasing

nutrient enrichment. The abundance of Gammarus spp. also

correlated positively to nutrient addition level, although not very

strongly. In contrast, there was a very strong positive correlation

betweenGammarus density and the amount of green algae. Path

analysis showed that the indirect effect of nutrient enrichment

(through green algae) on Gammarus density probably was more

important than the direct effect (by the nutrients themselves). It

was further seen that considerably moreG. locusta colonised red

algae than green and brown in the mesocosms and that

significantly greater numbers of G. locusta chose nutrient

enrichedE-algae than ambientA-algae in bothmesocosms and in

the field. Finally, a significantly faster growth rate for juvenileG.

locusta reared onE-algae (green and brown algae)was registered

compared to specimens reared on A-algae, with a significant

stimulation of the number of egg-carrying females when the

gammarids were reared on E-periphyton.

The increase in green algal cover along the nutrient gradient

was not unexpected as most conceptual marine eutrophication

models predict that ephemeral and epiphytic macroalgae will

become substantially more abundant with increasing nutrient

levels (Sand-Jensen and Borum, 1991; Duarte, 1995; Schramm,

1999). The stimulation of ephemeral algae is linked to

ecophysiological traits (growth rate, nutrient requirements

and uptake rates), which favour thin algae above thick algae at

higher nutrient levels (Pedersen and Borum, 1996, 1997). The

low cover of green algae observed in mesocosms in 1997 may

partly be explained by the weather conditions. The summer of

1997 was warm and dry in the Oslofjord area and the

background nutrient levels in the fjord were exceptionally low.

Severe nutrient limitation may therefore have limited the

amount of green algae in the mesocosms. The summers of

1998–1999 were, in turn, very rainy, resulting in high

background nutrient levels in the fjord (Kraufvelin et al., 2002).

The mesocosm gammarids responded both to the nutrient

enrichment and to the increased amounts of green algae.

Enhanced nutrient levels may be expected to improve food

quality for gammarids in several ways: increased N and P

content of algae, stimulation of preferred periphyton and green

algae, increased amounts of detritus and other food items, etc.

(Mattson, 1980). According to the path analysis, the effects of

nutrients on Gammarus were mostly indirect (through effects

on green algae): indirect effects contributing to 59% and direct

effects to 41%, respectively, of the total effect (0.76).

The gammaridean habitat preference (red > green >brown > periphyton) was almost the reverse of the growth rate

(periphyton > green > brown).Note that thegrowthon redalgae

was not examined here, since these red algae have hardly been

eaten at all in other experiments (Norderhaug, 2004; Pedersen

et al., unpubl.). This implies that red macroalgae may be more

important as a habitat than as a food source for these animals

(Dayton, 1971; Norderhaug et al., 2002; Christie andKraufvelin,

2004), but present study also suggests that the animals actively

search for food items of preferred nutritive quality. This mobility

is most pronounced at night for the majority of these grazing

macroinvertebrates, whereas the animals hide among red algae

andbeneath stones duringdaytimehours (JørgensenandChristie,

2003; Christie and Kraufvelin, 2004).

Algal habitat complexity often determines the associated

animal communities (Fretter andManley, 1977; Holmlund et al.,

1990; Gee andWarwick, 1994). The finer structure ofCeramium

and Rhodomela compared to F. serratus andU. lactuca seems to

imply that these red algal species are able to host 10–15 times

more gammarids, but also other taxa, per g DW (Fig. 3). Higher

animal abundance and diversity in a more complex habitat may

be a result of increased living space (Morse et al., 1985),

increased variety of food organisms (Fretter and Manley, 1977),

more suitable feeding surfaces (Gibbons, 1988a), and increased

protection from predation (Coull and Wells, 1983; Gibbons,

1988b; Pavia et al., 1999). Many studies of the factors affecting

host plant specialisation by herbivores have highlighted thevalue

of the plant both as habitat and as food, but have often failed to

distinguish between the relative importance of the two factors

(Sotka et al., 1999). This study suffers somewhat from a similar

problem, although results from parallel grazing experiments and

examination of algae from the Solbergstrand mesocosms (the

EULIT-project) support a conclusion that the red algae merely

served as a hiding place than as a food source for the majority of

the animals present. Among the dominant mesograzers, both G.

locusta and Idotea granulosa ate U. lactuca, F. serratus and F.

vesiculosus in feeding experiments, with a clear preference for

the former, whereas only very small amounts of the red algae

Ceramium were eaten (Pedersen et al., unpubl.). Nevertheless,

the red algae were by far the most popular habitat. Norderhaug

(2004) found that while amphipods grew well on the red alga

Palmaria palmata (Linn) Kuntze, which is a smooth and sheet-

like species similar to U. lactuca, at the same time were much

more abundant on other species of more bushy red algae, which

they in turn did not feed on.

The gammarids also clearly preferred caged algae to

attached ambient algae (consistently across the algal groups,

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209 207

Fig. 3). It is possible that caged algae placed out on the

mesocosm floors served as more efficient food traps or more

secure hiding places than the constantly moving/sweeping

attached algae on the mesocosm steps. The role of the cage

structures themselves in attracting animals can probably be

ruled out, however, due to the big differences in the number of

colonising gammarids between caged red algae and caged

green and brown algae.

It should be beneficial for animals to choose food items

higher in nitrogen content to save energy on foraging and

optimise fitness (Cruz-Rivera and Hay, 2000a,b). The

preference of E-algae above A-algae shown in this study

may depend on an ability to recognise high quality food (higher

N and P levels, more easily digestible, etc.), at ambient nutrient

concentrations in the field, in control mesocosms as well as in

aquaria, by using some kind of cues (Friedman and Strickler,

1975; Valiela and Rietsma, 1984; Valiela, 1995; Jackson and

Kiørboe, 2004). Previously, a preference among gammarids for

highly nutritive decomposing algal water above fresh algal

water has been found when the animals were exposed to both

types of water at ambient nutrient concentrations in laboratory

aquaria (Salovius and Kraufvelin, 2004). In nutrient enriched

mesocosms, however, an ability to detect high nutrient levels

may be of limited importance, since all surrounding algal

individuals probably contain sufficient amounts of nutrients.

This may explain the lack of preference for E-algae in the

nutrient enriched mesocosm.

In the mesocosm and in the field experiments, the

gammaridean preferences for and responses to certain algae

were linked positively to higher N and P contents. All nutrient

analyses showed higher N and P levels in algae from enriched

mesocosms, and only small changes took place in these levels

during the experimental period. An exception occurred for U.

lactuca in the 60 h cage colonisation experiment where nutrient

levels decreased in E-algae transplanted to the control

mesocosms. This may have interfered with the results

decreasing the amount of colonisers to U. lactuca. In the field

study, the higher colonisation rate by G. locusta to E-algae may

also have been underestimated due to the field sites being

different. The north site had higher background levels of N and

P (significant interaction effects in the two-way ANOVA on U.

lactuca nutrient content, Table 5), thus being more similar to a

nutrient enriched mesocosm, where no responses occurred,

while the southern site more resembled a control mesocosm.

With regard to gammaridean growth rate, the results were

very clear with the fastest growth rate on E-periphyton followed

by A-periphyton, E-Ulva, A-Ulva, E-Fucus, A-Fucus, and the

differences between the algal groups and between E- and A-

macroalgae being significant. The lack of a significant

difference in G. locusta growth between A- and E-periphyton

may be explained by the onset of reproduction. When female

gammarids mature and start to reproduce, their growth rate

decreases (Kolding and Fenchel, 1979).

What are the possible consequences of the documented

Gammarus spp. preferences and responses to E-algae? A

stimulated growth rate may lead to an earlier onset of

reproductive activity (shorter generation times) and in the end

more juveniles at higher nutrient levels (Hay, 1996; Cruz-

Rivera and Hay, 2000a, b), i.e. possibilities for a population

increase, although an accelerated life cycle also can mean a

reduced life span for individual gammarids (Neuparth et al.,

2002). Moreover, if E-algae attract more grazers and further

stimulate their growth, with a reproduction and population

increase, the suggested capabilities among mesograzers to

buffer moderate eutrophication effects (e.g. Geertz-Hansen

et al., 1993; Hillebrand et al., 2000; Worm et al., 2000) may

have been underestimated in the past. Such a buffering may

have occurred in the mesocosms/ecosystems studied within

EULITand may be one explanation to the few algal community

responses to the nutrient addition (Bokn et al., 2002, 2003).

Norderhaug (2003) found a higher degree of vulnerability to

grazing within kelp plants in spring when the C/N ratio was

lower. Increased gammaridean populations may then result in

overgrazing of the perennial algae (Kangas et al., 1982). On the

other hand, if the biomass of red ephemeral algae (or other

popular habitats for this animal group) decrease with increased

nutrient concentrations as indicated by Karez et al. (2004), the

G. locusta population may ultimately decline and its buffering

effects on the eutrophication process or detrimental effects on

the host perennial algae may decrease.

Eutrophication in rocky shore systems seems to provide

more and better suited plant food but less suitable habitats for

animals, where the latter may serve as a regulating mechanism

restricting grazers/overgrazing and even contribute to maintain

the ecological stability of the system (e.g. Thompson et al.,

2002; Bokn et al., 2003). Such mechanisms related to

ecosystem resistance and resilience are largely unknown, but

important future research fields regarding the functioning of

plant–animal assemblages. It would also be of interest to carry

out similar preference and growth experiments on Idotea spp.,

which may be both more stationary and more selective feeders

than G. locusta (Salemaa, 1987; Arrontes, 1990; Merilaita and

Jormalainen, 2000; Costa and Costa, 2000). In the mesocosm

cage colonisation experiments, Idotea spp. (i.e. I. granulosa

Rathke and Idotea pelagica Leach) showed an even stronger

preference for E-algae above A-algae than Gammarus spp. did

(Kraufvelin et al., unpubl. results). Finally, the large time lag

before the effects of eutrophication were reflected in these

artificial rocky shore communities needs to be mentioned. In

this paper, this was partly the case forGammarus density, but in

other parallel EULIT studies, this was even more evident for

canopy brown algae (Kraufvelin et al., 2002; Bokn et al., 2003).

Such time lags are sound pleas for more long-term experiments

(beyond the 3-year time frame of an EU project), especially

when dealing with larger-sized and longer-lived organisms,

subtle (non-toxic) community stressors, and creeping change.

Acknowledgements

The European Commission (through MAST III programme

MAS3-CT97-0153) financed most of this work. We also

received financial contribution from Abo Akademi University,

Carl Cedercreutz stipendiefond at Svenska Kulturfonden, Ella

och Georg Ehrnrooths Stiftelse, Svenska Litteratursallskapet i

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209208

Finland (Ingrid, Margit och Henrik Hoijers donationsfond II),

Oskar Oflunds Stiftelse, Societas pro Fauna et Flora Fennica,

Letterstedska Foreningen and the Academy of Finland (for the

first author), as well as from the Maj and Tor Nessling

Foundation project number 200006 (for the second author). We

thank the coordinator T.L. Bokn and the entire EULIT group for

a fruitful collaboration and S. Engelbert, F.L. Fotel, A. Hagg,

M. Olsen and B. Sinisalo for field assistance at Solbergstrand.

The editor J.E. Vermaat and two anonymous referees gave

valuable comments on the manuscript. K. O’Brien kindly

revised the language. The paper is contribution number 49 from

Marine Research Station Solbergstrand.

References

Arrontes, J., 1990. Diet, food preference, and digestive efficiency in intertidal

isopods inhabiting macroalgae. J. Exp. Mar. Biol. Ecol. 139, 231–249.

Balducci, C., Sfriso, A., Pavoni, B., 2001. Macrofauna impact onUlva rigidaC.

Ag. production and relationship with environmental variables in the lagoon

of Venice. Mar. Environ. Res. 52, 27–49.

Benedetti-Cecchi, L., Pannacciulli, F., Bulleri, F., Moschella, P.S., Airoldi, L.,

Relini, G., Cinelli, F., 2001. Predicting the consequences of anthropogenic

disturbance: large-scale loss of canopy algae on rocky shores. Mar. Ecol.

Prog. Ser. 214, 137–150.

Bokn, T.L., Hoell, E.E., Kersting, K., Moy, F.E., Sørensen, K., 2001. Methods

applied in the large littoral mesocosms study of nutrient enrichment in

rocky shore ecosystems—EULIT. Continental Shelf Res. 21, 1925–

1936.

Bokn, T.L., Moy, F.E., Christie, H., Karez, R., Kersting, K., Kraufvelin, P.,

Lindblad, C., Marba, N., Pedersen, M.F., Sørensen, K., 2002. Are rocky

shore ecosystems affected by nutrient enriched seawater? Some preliminary

results from a mesocosm experiment. Hydrobiologia 484, 167–175.

Bokn, T.L., Duarte, C.M., Pedersen, M.F., Marba, N., Moy, F.E., Barron, C.,

Bjerkeng, B., Borum, J., Christie, H., Engelbert, S., Fotel, F.L., Hoell, E.E.,

Karez, R., Kersting, K., Kraufvelin, P., Lindblad, C., Olsen, M., Sanderud,

K.A., Sommer, U., Sørensen, K., 2003. The response of experimental rocky

shore communities to nutrient additions. Ecosystems 6, 577–594.

Bryman, A., Cramer, D., 1990. Quantitative Data Analysis for Social Scientists.

Routledge, London.

Buschmann, A.H., 1990. Intertidal macroalgae as refuge and food for amphi-

poda in central Chile. Aquat. Bot. 36, 237–245.

Christie, H., Kraufvelin, P., 2004. Mechanisms regulating amphipod population

density within macroalgal communities with restricted predator impact. Sci.

Mar. 68 (Suppl. 1), 189–198.

Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutrophication

problem. Mar. Ecol. Prog. Ser. 210, 223–253.

Costa, F.O., Correia, A.D., Costa, M.H., 1996. Sensitivity of a marine amphipod

to non-contaminant variables and to copper in the sediments. Ecologie 27,

269–276.

Costa, F.O., Costa, M.H., 2000. Review of the ecology of Gammarus locusta

(L.) Pol. Arch. Hydrobiol. 47, 351–359.

Coull, B.C., Wells, J.B.J., 1983. Refuges from fish predation: experiments with

phytal meiofauna from the New Zealand rocky intertidal. Ecology 64,

1599–1609.

Cruz-Rivera, E., Hay, M.E., 2000a. Can quantity replace quality? Food choice,

compensatory feeding, and fitness of marine mesograzers. Ecology 81,

201–219.

Cruz-Rivera, E., Hay, M.E., 2000b. The effect of diet mixing on consumer

fitness: macroalgae, epiphytes, and animal matter as food for marine

amphipods. Oecologia 123, 252–264.

Cruz-Rivera, E., Hay, M.E., 2003. Prey nutritional quality interacts with

chemical defenses to affect consumer feeding and fitness. Ecol. Monogr.

73, 483–506.

Day, R.W., Quinn, G.P., 1989. Comparisons of treatments after an analysis of

variance in ecology. Ecol. Monogr. 59, 433–463.

Dayton, P.K., 1971. Competition, disturbance and community organization: the

provision and subsequent utilization of space in a rocky intertidal commu-

nity. Ecol. Monogr. 41, 351–389.

Duarte, C.M., 1995. Submerged aquatic vegetation in relation to different

nutrient regimes. Ophelia 41, 87–112.

Duffy, J.E., 1990. Amphipods on seaweeds: partners or pests? Oecologia 83,

267–276.

Duffy, J.E., Hay,M.E., 1991. Food and shelter as determinants of food choice by

an herbivorous marine amphipod. Ecology 72, 1286–1298.

Duggins, D., Eckman, J.E., Siddon, C.E., Klinger, T., 2001. Interactive roles of

mesograzers and current flow in survival of kelps. Mar. Ecol. Prog. Ser. 223,

143–155.

Dye, A.H., 1998. Dynamics of rocky intertidal communities: analyses of long

time series from South African shores. Estuar. Coast. Shelf Sci. 46, 287–

305.

Edgar, G.J., 1983. The ecology of south-east Tasmanian phytal animal com-

munities. IV. Factors affecting the distribution of amphitoid amphipods

among algae. J. Exp. Mar. Biol. Ecol. 70, 205–225.

Edgar, G.J., 1990. The influence of plant structure on the species richness,

biomass and secondary production of macroalgal assemblages associated

with Western Australian seagrass beds. J. Exp. Mar. Ecol. Biol. 137, 215–

240.

Edgar, G.J., Barrett, N.S., Morton, A.J., Samson, C.R., 2004. Effects of algal

canopy clearance on plant, fish and macroinvertebrate communities on

eastern Tasmanian reefs. J. Exp. Mar. Biol. Ecol. 312, 67–87.

Eggleston, D.B., Elis, W.E., Etherington, L.L., Dahlgren, C.P., Posey, M.H.,

1999. Organism responses to habitat fragmentation and diversity: habitat

colonization by estuarine macrofauna. J. Exp. Mar. Biol. Ecol. 236, 107–

132.

Everitt, B.S., Dunn, G., 1991. Applied Multivariate Data Analysis. Edward

Arnold, London.

Friedman, M.M., Strickler, J.R., 1975. Chemoreceptors and feeding in calanoid

copepods (Arthropoda: Crustacea). Proc. Natl Acad. Sci. 72, 4185–4188.

Fretter, V., Manley, R., 1977. Algal associations of Tricolia pullus, Lacuna

vincta and Cerithiopsis tubercularis (Gastropoda) with special reference to

the settlement of their larvae. J. Mar. Biol. Assoc. UK 57, 999–1017.

Gee, J.M., Warwick, R.M., 1994. Metazoan community structure in relation to

the fractal dimensions of marine macroalgae. Mar. Ecol. Prog. Ser. 103,

141–150.

Geertz-Hansen, O., Sand-Jensen, K., Hanen, D.F., Christiansen, A., 1993.

Growth and grazing control of abundance of the marine macroalgae Ulva

lactuca L in a eutrophic Danish estuary. Aquat. Bot. 46, 101–109.

Gibbons, M.J., 1988a. The impact of sediment accumulation, relative habitat

complexity and elevation on rocky shore meiofauna. J. Exp.Mar. Biol. Ecol.

122, 225–241.

Gibbons, M.J., 1988b. Impact of predation by juvenile Clinus superciliosus on

phytal meiofauna: are fish important as predators? Mar. Ecol. Prog. Ser. 45,

13–22.

Granado, I., Caballero, P., 2001. Feeding rates of Littorina striata and Osilinus

atratus in relation to nutritional quality and chemical defenses of seaweeds.

Mar. Biol. 138, 1213–1224.

Gray, J.S., Wu, R.S., Or, Y.Y., 2002. Effects of hypoxia and organic enrichment

on the coastal marine environment. Mar. Ecol. Prog. Ser. 238, 249–279.

Hay, M.E., 1996. Marine chemical ecology: what’s known and what’s next? J.

Exp. Mar. Biol. Ecol. 200, 103–134.

Hemmi, A., Jormalainen, V., 2002. Nutrient enhancement increases perfor-

mance of a marine herbivore via quality of its food algae. Ecology 83, 1052–

1064.

Hillebrand, H., Worm, B., Lotze, H., 2000. Marine microbenthic community

structure regulated by nitrogen loading and grazing pressure. Mar. Ecol.

Prog. Ser. 204, 27–38.

Hochberg, Y., 1988. A sharper Bonferroni procedure for multiple tests of

significance. Biometrika 75, 800–802.

Holmlund, M.B., Peterson, C.H., Hay, M.E., 1990. Does algal morphology

affect amphipod susceptibility to fish predation? J. Exp. Mar. Biol. Ecol.

139, 65–83.

Jackson, G.A., Kiørboe, T., 2004. Zooplankton use of chemodetection to find

and eat particles. Mar. Ecol. Prog. Ser. 269, 153–162.

P. Kraufvelin et al. / Aquatic Botany 84 (2006) 199–209 209

Jones, J.I., Young, J.O., Haynes, G.M., Moss, B., Eaton, J.W., Hardwick, K.J.,

1999. Do submerged aquatic plants influence their periphyton to enhance

the growth and reproduction of invertebrate mutualists? Oecologia 120,

463–474.

Jørgensen, N.M., Christie, H., 2003. Diurnal, horizontal and vertical dispersal of

kelp-associated fauna. Hydrobiologia 503, 69–76.

Kangas, P., Autio, H., Hallfors, G., Luther, J., Niemi, A., Salemaa, H., 1982. A

general model of the decline of Fucus vesiculosus at Tvarminne, south coast

of Finland in 1977–1981. Acta Bot. Fennica 118, 1–27.

Karez, R., Engelbert, S., Sommer, U., 2000. Co-consumption and protective

coating: two new proposed effects of epiphytes on their macroalgal hosts

in mesograzer–epiphyte–host interactions. Mar. Ecol. Prog. Ser. 205, 85–

93.

Karez, R., Engelbert, S., Kraufvelin, P., Pedersen, M.F., Sommer, U., 2004.

Biomass response and changes in composition of ephemeral macroalgal

assemblages along an experimental gradient of nutrient enrichment. Aquat.

Bot. 78, 103–117.

Kolding, S., Fenchel, T.M., 1979. Coexistence and life cycle characteristics of

five species of the amphipod genus Gammarus. Oikos 33, 323–327.

Kraufvelin, P., Salovius, S., 2004. Animal diversity in Baltic rocky shore

macroalgae: can Cladophora glomerata compensate for lost Fucus

vesiculosus? Estuar. Coast. Shelf Sci. 61, 369–378.

Kraufvelin, P., Christie, H., Olsen, M., 2002. Littoral macrofauna (secondary)

responses to experimental nutrient addition to rocky shore mesocosms and a

coastal lagoon. Hydrobiologia 484, 149–166.

Lopez-Rodriguez, M.C., Barbara, I., Perez-Cirera, J.L., 1999. Effects of

pollution on Fucus vesiculosus communities on the northwest Iberian

Atlantic coast. Ophelia 51, 129–141.

Mattson Jr., W.J., 1980. Herbivory in relation to plant nitrogen content. Ann.

Rev. Ecol. Syst. 12, 405–437.

Merilaita, S., Jormalainen, V., 2000. Different roles of feeding and protection in

diel microhabitat choice of sexes in Idotea baltica. Oecologia 122, 445–451.

Morse, D.R., Lawton, J.H., Dodson, M.M., Williamson, M.H., 1985. Fractal

dimension of vegetation and the distribution of arthropod body length.

Nature 314, 731–733.

Neuparth, T., Costa, F.O., Costa, M.H., 2002. Effects of temperature and salinity

on life history of the marine amphipod Gammarus locusta. Implications for

ecotoxicological testing. Ecotoxicology 11, 61–73.

Nicotri, M.E., 1980. Factors involved in herbivore food preference. J. Exp. Mar.

Biol. Ecol. 42, 13–26.

Norderhaug, K.M., 2003. Importance of macrofauna in transferring kelp forest

primary production to higher levels in the food web. PhD Thesis. University

of Oslo.

Norderhaug, K.M., 2004. Use of red algae as hosts by kelp-associated amphi-

pods. Mar. Biol. 144, 225–230.

Norderhaug, K.M., Christie, H., Rinde, E., 2002. Colonisation of kelp imitations

by epiphyte and holdfast fauna; a study of mobility patterns. Mar. Biol. 141,

965–973.

Pavia, H., Carr, H., Aberg, P., 1999. Habitat and feeding preferences of

crustacean mesoherbivores inhabiting the brown seaweed Ascophyllum

nodosum (L.) Le Jol and its epiphytic macroalgae. J. Exp. Mar. Biol. Ecol.

236, 15–32.

Pedersen, M.F., Borum, J., 1996. Nutrient control of algal growth in estuarine

waters. Nutrient limitation and the importance of nitrogen requirements and

nitrogen storage among phytoplankton and species of macroalgae. Mar.

Ecol. Prog. Ser. 142, 261–272.

Pedersen, M.F., Borum, J., 1997. Nutrient control of estuarine macroalgae:

growth strategy and the balance between nitrogen requirements and uptake.

Mar. Ecol. Prog. Ser. 161, 155–163.

Persson, A., Hansson, L.A., Bronmark, C., Lundberg, P., Petterson, L.B.,

Greenberg, L., Nilsson, P.A., Nystrom, P., Romare, P., Tranvik, L., 2001.

Effects of enrichment on simple aquatic food webs. Am. Nat. 157, 654–669.

Russell, B.D., Connell, S.D., 2005. A novel interaction between nutrients and

grazers alters relative dominance of marine habitats. Mar. Ecol. Prog. Ser.

289, 5–11.

Salemaa, H., 1987. Herbivory and microhabitat preferences of Idotea spp.

(Isopoda) in the Northern Baltic Sea. Ophelia 27, 1–15.

Salovius, S., Kraufvelin, P., 2004. The filamentous green alga Cladophora

glomerata as a habitat for littoral macrofauna in the northern Baltic Sea.

Ophelia 58, 65–78.

Sand-Jensen, K., Borum, J., 1991. Interactions among phytoplankton, periph-

yton and macrophytes in temperate freshwaters and estuaries. Aquat. Bot.

41, 137–175.

Schramm, W., 1999. Factors influencing seaweed responses to eutrophication:

some results from EU-project EUMAC. J. Appl. Phycol. 11, 69–78.

Shacklock, P.F., Doyle, R.W., 1983. Control of epiphytes in seaweed cultures

using grazers. Aquaculture 31, 141–151.

Smith, V.H., 2003. Eutrophication of freshwater and coastal marine ecosys-

tems—a gobal problem. Environ. Sci. Pollut. Res. 10, 126–139.

Sotka, E.E., Hay, M.E., Thomas, J.D., 1999. Host–plant specialisation by a non-

herbivorous amphipod: advantages for the amphipod and costs for the

seaweed. Oecologia 118, 471–482.

Strickland, J.D.H., Parsons, T.R., 1968. A practical handbook of seawater

analysis. Bull. Fish. Res. Bd. Can. 167, 1–311.

Thompson, R.C., Crowe, T.P., Hawkins, S.J., 2002. Rocky intertidal commu-

nities, past environmental changes, present status and predictions for the

next 25 years. Environ. Conserv. 29, 168–191.

Valiela, I., 1995. Marine Ecological Processes, 2nd ed. Springer-Verlag, New

York.

Valiela, I., Rietsma, C.S., 1984. Nitrogen, phenolic acids, and other feeding cues

for salt marsh detrivores. Oecologica 63, 350–356.

Worm, B., Sommer, U., 2000. Rapid direct and indirect effects of a single

nutrient pulse in a seaweed–epiphyte–grazer system. Mar. Ecol. Prog. Ser.

202, 282–288.

Worm, B., Lotze, H., Sommer, U., 2000. Coastal food web structure, carbon

storage, and nitrogen retention regulated by consumer pressure and nutrient

loading. Limnol. Oceanogr. 45, 339–349.

Worm, B., Lotze, H.K., Hillebrand, H., Sommer, U., 2002. Consumer versus

resource control of species diversity and ecosystem functioning. Nature 417,

848–851.