Forgot your password?
Document Actions
Cervin 04
Journal of Experimental Marine Biology and Ecology
302 (2004) 35 – 49
www.elsevier.com/locate/jembe
Effects of small-scale disturbances of canopy and
grazing on intertidal assemblages on the
Swedish west coast
˚
Gunnar Cervin a,*, Mats Lindegarth b, Rosa M. Viejo a,1, Per Aberg a
a
Department of Marine Ecology, Marine Botany, Goteborg University, Box 461, Goteborg SE-405 30, Sweden
¨ ¨
b
Department of Marine Ecology, Tjarno Marine Biological Laboratory, Goteborg University,
¨ ¨ ¨
Stromstad SE-452 96, Sweden
¨
Received 13 March 2003; received in revised form 23 July 2003; accepted 4 September 2003
Abstract
The effects of small-scale disturbances (80 Â 30-cm plots) of canopy and grazers on intertidal
assemblages were investigated in this 4-year experiment on sheltered rocky shores on the Swedish
west coast. Canopy disturbances due to ice scouring were mimicked by removal of adult plants of the
seaweed Ascophyllum nodosum (L.) Le Joli. Density of the main epilithic grazing gastropods,
Littorina spp., was lowered by exclosure and handpicking. Based on earlier experiments in other
areas, the general hypothesis was that canopy removal and grazer exclosure, alone or in combination,
should increase the recruitment of A. nodosum or other fucoid juveniles, and change the structure of
the understorey assemblage.
There was an effect of canopy removal on the development of this assemblage, lasting for more
than 31 months. Both increased and decreased abundances of species were found as short-term
effects, but there was also a longer-term effect with increased abundance. Grazer exclosure was only
effective in combination with canopy removal, causing a short-term increase in ephemeral green
algae. Short-term effects of canopy removal were also the increase in recruitment of Semibalanus
balanoides (Linnaeus) and the decrease of the red alga Hildenbrandia rubra (Sommerfelt)
Meneghini. Fast recruitment and growth of fucoid species (Fucus serratus L. and F. vesiculosus L.)
restored the canopy and conditions of the understorey within 18 months. Thus, the canopy removal
changed the physical conditions for the understorey, making it possible for other species to coexist in
this community. Surprisingly, no effect of canopy removal or grazer exclusion was found on the
recruitment of juvenile A. nodosum, neither by canopy removal nor grazer exclosure. The lack of
* Corresponding author. Tel.: +46-31-7732709; fax: +46-31-7732727.
E-mail address: gunnar.Cervin@marbot.gu.se (G. Cervin).
1 ´ ´
Current address: Area de Biodiversidad y Conservacion, Escuela Superior de Ciencias, Experimentales y
´ ´ ´
Tecnologıa, Universidad Rey Juan Carlos, Tulipan s/n, Mostoles, Madrid E-28933, Spain.
0022-0981/$ - see front matter D 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jembe.2003.09.022
36 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
such effects might be due to the early mortality caused by other grazers (small, mobile crustaceans),
or to the low density of periwinkles on these shores. However, despite the patchy and generally low
recruitment of A. nodosum juveniles, observations suggested that the cover of A. nodosum in
manipulated patches would return to initial levels, either by recruitment or regrowth of small
holdfasts and from growth of edge plants.
D 2003 Elsevier B.V. All rights reserved.
Keywords: Adult canopy; Ascophyllum nodosum; Grazing; Fucus; Recruitment; Littorina
1. Introduction
Rocky shores as well as most natural communities are characterized by spatial
heterogeneity and temporal dynamics (Sousa, 1984). This may be a response to
physical disturbances and biological interactions, which commonly cause communities
to undergo unpredictable changes in species abundances, or a predictable temporal
sequence of species replacements. These changes often involve the appearance or
dominance of plant species with progressively greater maximum size, age, and shade
tolerance, and progressively lower maximum growth rates and dispersal abilities
(McCook, 1994). Disturbances, such as formation of gaps in spatial cover or increased
levels of resources, facilitate the establishment of early colonizers, even when they are
eventually excluded from the patch by competitively dominant species. The presence
of dominant, structure-forming species, on the other hand, may facilitate the growth
and/or survival of associated or ‘‘accompanying’’ species in natural assemblages, acting
as ‘‘physical engineers’’ (sensu Jones et al., 1994), ameliorating the environment. For
example, red algal species that are otherwise found sublittoraly can prosper under
dense fucoid canopies in the intertidal.
In sheltered rocky shores of the temperate region of the north Atlantic, the seaweed
Ascophyllum nodosum (L.) Le Joli is one of the major canopy-forming species in the mid-
intertidal zone. Turf forming and particularly encrusting species dominate the understorey
in shores on the Swedish west coast (G.C., personal observation). Other canopy-forming
species above the A. nodosum zone are Fucus spiralis L., and below there are F. serratus
L. and F. vesiculosus L. (Lewis, 1964). These are, however, considered competitively
inferior to A. nodosum in the mid-intertidal (Lewis, 1964). Nevertheless, F. serratus and F.
vesiculosus are sometimes also found as smaller or larger patches within the A. nodosum
canopy (Jenkins et al., 1999b). An important agent of disturbance of these canopies in the
NW Atlantic and Swedish west coast is ice scouring, creating gaps of different sizes
˚
(Mathieson et al., 1982; Aberg, 1992; Dudgeon and Petraitis, 2001), with small-scale
˚
losses typically occurring on sheltered shores of the Swedish west coast (P.A., personal
observation). The losses of A. nodosum due to ice scouring may be more than 50% of the
˚
biomass during years of extreme ice cover (Mathieson et al., 1982; Aberg, 1992). Previous
studies have indicated that the effect of gaps on the recruitment of A. nodosum and the
development of these communities are size-dependent (Dudgeon and Petraitis, 2001). The
presence of grazers in combination with disturbances may affect the sequence and rate of
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 37
colonization by ephemerals and recruitment of fucoids. Studies on the Isle of Man
suggested interactions among canopy, red algal turf, and limpets (Jenkins et al., 1999b).
For example, in areas with dense cover of A. nodosum over rocks kept bare by grazing
limpets, Jenkins et al. (1999b) found that combined removal of both canopy and grazers
caused a massive growth of Fucus spp. Previous studies on the Swedish west coast, where
limpets are lacking, have demonstrated that high densities of L. littorea (Linnaeus)
˚
increase the mortality of A. nodosum juveniles (Cervin and Aberg, 1997). Complex
interactions of canopy disturbance, juvenile densities, and Littorina spp. have also been
demonstrated for the survival of postsettlement stages of A. nodosum, and it was
hypothesized that grazing also may affect the survival of older juvenile stages (Viejo et
al., 1999). However, grazing by the dominant littorinid [Littorina saxatilis (Olivi)] on
semiexposed shores on the Swedish west coast does not appear to significantly affect algal
assemblages (Lindegarth et al., 2001).
This paper presents the results of a manipulative experiment on sheltered shores of
the Swedish west coast. The experiment was designed to test the effect of small-scale
disturbances to the canopy species A. nodosum typical for this area (canopy removal)
and grazing by periwinkles (grazing exclosure) on the recruitment of fucoids,
particularly A. nodosum, and the development of associated assemblages. For A.
nodosum and Fucus spp., the specific hypothesis is that we expect higher recruitment
in canopy removal and grazer exclosure in combination. This is the first time that such
an experiment has been done on sheltered shores on the Swedish west coast, with
almost atidal conditions and irregular heavy ice winters, which make it different from
similar experiments done in other parts of the world.
2. Materials and methods
2.1. Study area
The study started in April 1997 and was conducted in the northernmost part of the
Swedish west coast, close to the Tjarno Marine Biological Laboratory (TMBL) (N58j52.6V,
¨ ¨
E11j9.0V). This area consists of numerous granite islands and islets. The surface water is
influenced by fully saline North Sea water, brackish water from the Baltic Sea, and
freshwater from the river Glomma in southeastern Norway, giving a surface salinity from
15 to 30 psu in the study area. The air temperature frequently drops below 0 jC, and ice
winters are common. The regular tidal range is less than 0.3 m, but the water level can vary
up to 2 m, depending on atmospheric pressure and winds (Johannesson, 1989).
2.2. Experimental procedures
Effects of the presence of macroalgal canopy and grazing periwinkles on the
recruitment of canopy-forming fucoids and the development of the understorey were
tested in a two-factor experiment. The experimental factors were Canopy {Intact vs.
Removed (CI vs. CR)} and Grazer {Fences, Procedural Fence Control, and Unmanip-
ulated Control (FE, PC, and CO)}. Three replicates were used for each treatment. The
38 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
experimental plots were randomly allocated to areas with dense populations of A.
nodosum on different islands within 2 km of TMBL. Due to the narrow intertidal zone
that restricts the vertical distribution of A. nodosum, the plots were 80 cm wide by 30
cm high.
In mid April 1997, before A. nodosum release their gametes, adult plants including
their holdfasts were removed in nine plots (CR). Nine other plots were marked but
canopies were left intact (CI). Adult plants around each cleared plot were trimmed to
prevent large plants from reaching into the plots. To prevent periwinkles from entering the
plots, fences (FE) of stainless wire mesh with a mesh size of 0.5 cm were used
(80 Â 30 Â 5 cm). A procedural fence control (PC) consisting of partial barriers that
protected the corners of plots was also included in the experiment. These were included to
test for experimental artefacts, such as changes in flow conditions and shading, that might
be caused by fences, without preventing access for grazing snails. Thus, if there were
significant artefacts due to fences, controls without any fences would differ from both
fence treatments (CO p PC = FE). If, on the other hand, there were no experimental
artefacts but effects of removing the grazers, plots with fences would differ from the plots
where snails had access (CO = PC p FE). Fences were attached to the rock using stainless
screws and plastic expanders, and two screws marked the control plots without fences
(CO).
The number of Littorina spp. on those shores where the experiment was performed
was very variable, with a mean of 17 L. littorea mÀ 2 and 17 L. mariae Sacchi and
Rastelli/L. obtusata (Linnaeus) mÀ 2. During the experiment, snails were observed
inside and climbing over experimental fences. Therefore, plots with fences were cleared
of gastropods every 2 days during the first 4 weeks, and weekly for another 2 months.
The periwinkle grazers were not completely excluded from these areas, but there was a
clear reduction in densities during the first 3 months after gamete release. On the
Swedish west coast, A. nodosum reproduce during a short period of 1 –2 weeks, but
depending on water temperature and other physical factors such as periods of low
water in combination with warm air temperature, the onset of the short reproductive
period can vary among years from mid April to late in May. The other canopy-forming
species, F. serratus and F. vesiculosus, reproduce in winter to early spring and April to
June, respectively (Rueness, 1977). Since the treatments were not maintained over the
whole experimental time, experimental manipulation can be seen as a pulse-type
perturbation.
The percentage cover of different species was measured by the intersection of 30
random points in each plot in April, July, and October 1997; November 1998; and
November 1999. Algae were identified to species or genus, and animals were usually
identified to species. Five randomly placed subplots (5 Â 5 cm) were used to estimate the
density of juvenile A. nodosum in July and October 1997. In November 1998 and 1999,
and September 2001, the whole area of the plots was searched for juvenile A. nodosum.
2.3. Statistical analyses
Patterns of whole assemblages were explored at individual times (t = 0, 3, 6, 19, and 31
months) using metric multidimensional scaling (MDS) (using routines in Anderson,
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 39
Fig. 1. MDS of assemblages at different times of sampling. Empty and filled symbols are plots with and without
canopy, respectively. Diamonds, squares, and triangles are plots with fences, control fences, and plots without
fences, respectively. Numbers indicate cumulative percent variability explained by the two first principal
components.
40
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
Table 1
NPMANOVA on benthic assemblages at different times of sampling
Source df Start 3 months 6 months 19 months 31 months 53 months
(April 1997) (July 1997) (October 1997) (November 1998) (November 1999) (September 2001)
MS F p MS F p MS F p MS F p MS F p MS F p
Canopy 1 628 0.80 >0.5 5474 5.92 < 0.01 10,832 10.04 < 0.01 4305 3.14 < 0.02 4641 5.99 < 0.01 Not analysed
Fence 2 442 0.57 >0.5 510 0.55 >0.5 1401 1.30 >0.2 742 0.54 >0.5 1583 2.04 >0.09
Canopy  Fence 2 1666 2.13 < 0.05 1221 1.32 >0.2 1203 1.11 >0.2 1781 1.30 >0.2 1162 1.50 >0.2
Residual 12 782 924 1079 1373 775
See text for further details about analytical procedures.
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 41
2000a). Explicit tests of hypotheses about effects of experimental treatments on whole
assemblages were tested using nonparametric multivariate analysis of variance (NPMA-
NOVA) and a posteriori pairwise t tests (Anderson, 2000b). Tests for effects of treatments
were done using random permutation of data among treatments (Anderson, 2001). All
multivariate analyses were done using Bray –Curtis distances on fourth-root transformed
data (Clarke and Warwick, 1994). Strength of treatment effects was quantified using
average Bray – Curtis dissimilarities among and within experimental treatments at each
time of sampling.
Univariate analyses were done using two-way ANOVA. Heterogeneity of variances was
tested for using Cochran’s test. If necessary, data were transformed to remove heteroge-
neity. In some instances, transformation of data was not successful in reducing the
heterogeneity of variances. In these cases, we did the ANOVA on untransformed data,
but these tests should be treated with some caution since heterogeneous variances can
increase the risk of Type I error.
The dominant components in these assemblages and those included in multivariate
analyses were: Ahnfeltia plicata (Hudson) Fries, Calothrix spp., Ceramium spp.,
Chondrus crispus Stackhouse, Cladophora rupestris (L.) Kutzing, Dynamena pumila
¨
(Linnaeus), Enteromorpha spp., Electra pilosa (Linnaeus), Fucus spp., Halichondria
panicea (Pallas), Hildenbrandia rubra (Sommerfelt) Meneghini, Hydroidea spp.,
Mytilus edulis Linnaeus, Phycodrys rubens (Linnaeus) Batters, Phymatolithon spp.,
Polysiphonia spp., Ralfsia verrucosa (Areschoug) Areschoug, Sagartiidae, Semibanla-
nus balanoides (Linnaeus), Sphacelaria spp., Spirorbis spirorbis (Linnaeus), Ulothrix
spp., Urospora spp., and Ulva spp. The species that were manipulated (A. nodosum
and Littorina spp.) were not included in the multivariate analyses. Univariate analyses
were done on juvenile A. nodosum and on commonly found species in the plots.
Fig. 2. Dissimilarity between (filled circles) and within (empty circles) canopy treatments at different monitoring
dates.
42
Table 2
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
ANOVAs of specific parameters at different monitoring dates
Source df Start 3 months 6 months 19 months 31 months 53 months
(April 1997) (July 1997) (October 1997) (November 1998) (November 1999) (September 2001)
MS F p MS F p MS F p MS F p MS F p MS F p
(a) The number of A. nodosum juveniles
Canopy 1 Not measured 0.06 4.17 >0.05 0.009 0.50 >0.4 56.9 1.48 >0.2 14.2 1.14 >0.3 220 2.35 >0.1
Fence 2 0.002 0.17 >0.8 0.03 1.50 >0.2 92.2 2.41 >0.1 0.39 0.031 >0.9 26.0 0.28 >0.7
Canopy  Fence 2 0.002 0.17 >0.8 0.06 3.50 >0.06 74.1 1.94 >0.1 5.06 0.40 >0.6 24.0 0.26 >0.7
Residual 12 0.01 0.02 38.2 12.5 93.7
Cochran’s C test ( p) < 0.05 ns < 0.05 ns < 0.05
(b) The cover of Fucus spp.
Canopy 1 22.22 0.391 >0.5 672.2 2.637 >0.1 70.44 19.15 < 0.001 7160 21.83 < 0.001 13,889 15.16 < 0.005 1422 2.18 >0.1
Fence 2 56.17 0.989 >0.4 68.52 0.269 >0.7 2.386 0.6489 >0.5 152.0 0.4635 >0.6 605.6 0.661 >0.5 1184 1.82 >0.2
Canopy  Fence 2 124.1 2.185 >0.1 12.96 0.0508 >0.9 2.918 0.7935 >0.4 77.56 0.2365 >0.7 205.6 0.224 >0.8 310 0.47 >0.6
Residual 12 56.8 255 3.68 328 917 653
Cochran’s C test ( p) < 0.05 < 0.05 ns after ns ns ns
sqrt(x + 1)
transformation
(c) The cover of ephemeral green algae
Canopy 1 0.62 0.03 >0.8 1867 8.72 < 0.05 5.56 1.80 >0.2 0.89 1.80 >0.2 Not measured, Not measured
as the percentage
cover was less
than one
Fence 2 7.41 0.34 >0.7 1613 7.53 < 0.01 1.85 0.60 >0.5 0.30 0.60 >0.5
Canopy  Fence 2 32.1 1.49 >0.2 1671 7.80 < 0.01 1.85 0.60 >0.5 0.30 0.60 >0.5
Residual 12 21.6 214 3.09 0.49
Cochran’s C test ( p) < 0.05 < 0.05 < 0.05 < 0.05
(d) The cover of H. rubra
Canopy 1 158 0.41 >0.5 3024 8.93 < 0.05 6296 8.38 < 0.05 601 0.58 >0.4 987 1.60 >0.2 Not measured
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
Fence 2 1067 2.76 >0.1 169 0.50 >0.6 121 0.16 >0.8 384 0.37 >0.7 61.7 0.10 >0.9
Canopy  Fence 2 549 1.42 >0.2 230 0.68 >0.5 128 0.17 >0.8 281 0.27 >0.7 151 0.24 >0.7
Residual 12 387 339 751 1035 617
Cochran’s C test ns ns ns ns ns
(e) The cover of S. balanoides
Canopy 1 1543 8.87 < 0.05 7.70 12.6 < 0.005 183 1.01 >0.3 0.00 0.00 1 0.00 0.00 1 Not measured
Fence 2 85.8 0.49 >0.6 1.78 2.91 >0.09 178 0.98 >0.4 0.61 0.50 >0.6 0.62 0.50 >0.6
Canopy  Fence 2 260 1.49 >0.2 1.78 2.91 >0.09 30.3 0.17 >0.8 1.85 1.50 >0.2 1.85 1.50 >0.2
Residual 12 174 0.61 181 1.23 1.23
Cochran’s C test ns ns after ns ns
sqrt(x + 1)
transformation
43
44 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
3. Results
3.1. Multivariate analyses
Two-dimensional ordinations of assemblages using metric MDS explained between
63% and 75% of the total variability among plots at individual times. The ordinations
indicated that there were differences between experimental treatments, but these changed
among times (Fig. 1). Different assemblages developed in plots with and without a
canopy of A. nodosum. This effect was most evident after 6 months of the start of the
experiment (i.e., October 1997). Actually, multivariate analyses of effects at individual
times showed that there were significant effects of canopy removal at all times of
sampling, except at the start of the experiment (Table 1). Thus, it can be concluded that
the abundance and composition of these assemblages were affected by the removal of the
canopy already after 3 months and that the effect persisted for at least 31 months.
Quantitative estimates of average dissimilarities indicate that those within experimental
treatments changed little, but dissimilarities between treatments peaked after 6 months
(Fig. 2). No effects of grazing treatment were detected (Table 1). This is consistent with
visual impressions represented by the MDS (Fig. 1). One surprising result was that there
was a significant interaction effect at the start of the experiment (Table 1). No significant
differences were, however, detected using a posteriori tests. Nevertheless, the largest
difference at the start of the experiment was observed in plots without fences, where there
was an average dissimilarity of between 48% and 34% within canopy treatments. This
difference between canopy treatments was clearly smaller than what was observed at
subsequent times of sampling (e.g., more than 60% difference between canopy and no
canopy for all fencing treatments in October 1997; Fig. 2).
3.2. Univariate analyses
No effect of canopy removal or grazer exclusion was found on the recruitment of
juvenile A. nodosum after 32 months, nor after an additional time of sampling after 53
months (Table 2a, Fig. 3A). Plants surrounding plots that had been removed of canopy
were trimmed at the start so that they would not reach into the plots. These trimmed
plants had grown in size so that they, in November 1999, had begun to cover the plots.
At the start of the experiment and after 3 months, there were no effects of any of the
experimental treatments on the cover of Fucus spp. In October 1997, however, a large
number of juvenile Fucus spp. had recruited into plots where the canopy was removed,
especially in the plots where L. littorea was also excluded (Fig. 3B). The percent cover
of Fucus spp. was significantly higher in the plots without a canopy of A. nodosum than
in those with an intact canopy, but no significant effects of grazer exclosure were
detected (Table 2b, Fig. 3B). There was also a significant effect of canopy removal on
Fig. 3. (A – E) The number of A. nodosum juveniles < 20 mm per 0.8 Â 0.3 m, and the percentage cover of Fucus
spp., ephemeral green algae (mainly Enteromorpha spp. and Ulva spp.), H. rubra, and S. balanoides at the
monitoring dates. Less than 1% cover is not shown (symbol meanings: CI = canopy intact, CR = canopy removal,
FE = fence, PC = procedural fence control, CO = control).
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 45
46 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
the cover of Fucus spp. in November 1998 and 1999, and in plots with canopy removed,
the percent cover of Fucus spp. had increased at these dates. In September 2001,
however, the cover of Fucus spp. had declined in these plots and no significant effect of
canopy removal was found (Table 2b, Fig. 3B).
In July 1997, there was a significant interactive effect of canopy removal and grazer
exclosure on the cover of the ephemeral green algae (mainly Enteromorpha spp. and
Ulva spp.). In plots where grazers were excluded and canopy was removed, the mean
cover of ephemeral green algae was 58.9 F 20.6% (mean F S.E.), while in the proce-
dural controls and controls, there was very little growth of ephemeral green algae (Table
2c, Fig. 3D). Thus, at the only time, and for the only taxon that showed an effect of
grazing treatment, there was no indication of procedural artefacts (i.e., FE > CO = PC for
CR; FE = CO = PC for CI). The effect was not found at later dates.
There was also a significant effect of canopy removal on the red algal crust H. rubra
(Table 2d, Fig. 3D) and the barnacle Semibalanus balanoides (Table 2e, Fig. 3E) in July
and October 1997. The percent cover of H. rubra was lower in plots with no canopy, while
the percent cover of barnacles was higher in these plots. After October 1997, no significant
effect of canopy removal was detected for either the cover of H. rubra or barnacles (Table
2d and e, Fig. 3D and E). It is notable that this coincided with the development of a Fucus
spp. canopy in the plots where the canopy of A. nodosum had been removed (Fig. 3B).
4. Discussion
The results of the study showed that small-scale disturbances in the canopy of A.
nodosum significantly affected the development of intertidal assemblages. Some of these
effects persisted for more than 2 years, while others were more short-lived.
Short-term effects of small-scale disturbances were both positive and negative. The
abundance of the barnacle S. balanoides increased as a result of canopy disturbance,
recruiting in greater numbers in the absence of the canopy. It is possible that the canopy
prevented recruitment by acting as a physical barrier, or by sweeping away settling larvae.
The increase in cover of S. balanoides was short-lived, as the later appearance of a Fucus
spp. canopy reduced subsequent recruitment and increased the mortality of barnacles. The
three fucoid species A. nodosum, F. serratus, and F. spiralis have presently been shown to
have a negative effect on settlement and early recruitment of S. balanoides (Jenkins et al.,
1999a). Not only sweeping of algal fronds are important, but also chemical or physical
cues from canopy algae may discourage the larvae (Jenkins et al., 1999a).
The presence of adult plants of A. nodosum for understorey species is similar to other
dominant species with high biomass and persistent structures, such as trees or corals,
which can act as ‘‘physical engineers’’ (sensu Jones et al., 1994), ameliorating the
environment. As a consequence, species like H. rubra are affected negatively by removal
of the canopy. Colonization of Fucus spp. will restore the canopy and ameliorate the
environment for these understorey species. Another short-term effect of canopy removal,
in combination with grazer exclusion, was the development of a cover of ephemeral green
algae. This growth of ephemeral green algae also showed that the grazer exclosure was
effective during the first 3 months of the experiment. However, the removal of grazers was
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 47
only intense during the first months of the experiment and strong effects of grazers may
not be so likely at later times.
A longer-term effect of the canopy removal of A. nodosum was the development of a
canopy of Fucus spp., which, as juveniles, are fast-growing species compared to A.
nodosum. The high number of recruiting juvenile Fucus spp. and their fast growth are the
reasons why we measured them as percentage cover. F. serratus and F. vesiculosus are
both competitively inferior canopy species to A. nodosum (Jenkins et al., 1999c), and, as
such, do prosper in the absence of an A. nodosum canopy. In intertidal rocky shores,
ephemeral species such as Ulva spp. and Enteromorpha spp. are commonly early
colonizers after disturbances, and a general pattern is that they inhibit or at least retard
the appearance of canopy-forming species, such a Fucus spp. (e.g., Lubchenco, 1983). The
mechanism of inhibition in succession (envisioned by Connell and Slatyer, 1977) did not
seem to work on intertidal sheltered shores on the Swedish west coast. The presence of
ephemeral green algae did not prevent the development of fucoid species. If anything, the
effect was to increase the rate of development of a canopy of Fucus spp. The first summer
was hot with extreme low water, which caused the plants to be emerged for several days
(authors, personal observation), and thus the ephemeral green algae might have protected
recruiting fucoids from desiccation. The effect of ephemeral green algae may vary from
year to year, depending on temperature and atmospheric pressure. The indirect effect of
grazers would then retard the succession by eliminating the protective cover of ephemeral
green algae, in concordance with the model of Farrell (1991). Previous results have also
suggested a positive effect of ephemeral green algae for the survival of the early stages of
A. nodosum germlings (Viejo et al., 1999).
The experiment showed that small-scale disturbance of the A. nodosum canopy created
the opportunity for colonization of other species, such as Fucus spp. and, in interaction
with grazer exclusion, colonisation of ephemeral green algae. Ice scouring may thus open
‘‘windows’’ for coexistence of species, even if A. nodosum is a superior competitor on
sheltered shores. In this assemblage, as in many others, the natural disturbances facilitate
the coexistence of species. Previous studies have also suggested that if the size of the
disturbance is large enough, the removal of adult A. nodosum may initiate the establish-
ment of alternative, successional endpoints, such as assemblages dominated by barnacles
or mussels (Petraitis and Latham, 1999). This would probably not occur on the Swedish
west coast, due to the highly variable exposures at a smaller scale on those shores, always
leaving some remnants of the original population within dispersal distance.
On these sheltered rocky shores on the Swedish west coast where A. nodosum is
dominant, small-scale disturbances of Littorina spp. did not increase the recruitment of
juvenile A. nodosum. This is different from previous and similar small-scale studies
made on the Isle of Man, where removal of the main grazer Patella vulgata caused much
higher numbers of juvenile A. nodosum to recruit within 2 years, and the first year also
in interaction with canopy removal (Cervin et al., in preparation). These areas have the
˚
same naturally patchy distribution of juvenile A. nodosum (Aberg and Pavia, 1997), and
we believe that similar plant – animal interactions would work in these areas. The access
of A. nodosum zygotes should be massive, as the egg rain is estimated to be 2.5 Â 109
˚
eggs mÀ 2 (Aberg and Pavia, 1997) and as the fertilisation probably is high since other
studies of fucoids have shown high fertilisation success (>90%) (e.g., Brawley, 1992;
48 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
˜
Serrao et al., 1996, 1999). Furthermore, previous studies in the NW Atlantic have shown
that dispersal of A. nodosum zygotes is very variable, but that more than 50% of the
zygotes disperse more than 6 m (Dudgeon et al., 2001), with an estimated maximum of
30 m (Chapman, 1995). Moreover, previous studies have also indicated that sexual
recruitment may be affected by the absence of a canopy of adults and by the size of
these clearings (e.g., Dudgeon et al., 2001; Dudgeon and Petraitis, 2001). Viejo et al.
(1999) have shown that canopy removal of the same size as in this experiment does not
alter the postsettlement survival of A. nodosum juveniles. Neither is there a higher
˚
presettlement mortality in this size of canopy removal, according to earlier studies (P.A.,
unpublished data). However, despite zygote availability, canopy removal, and successful
exclusion of L. littorea during the first 3 months after spawning of A. nodosum gametes,
no effect could be seen at the monitoring dates. Why were no effects of canopy removal
or grazer exclusion, alone or in combination, detected here? Although this was a long-
time experiment, the temporal scale might be too short for such a long-lived plant as A.
nodosum. The differences could be at the level of settlement patterns, where mortality
occurs before the zygotes have managed to establish, due to grazing by fast-moving
crustaceans not included in the treatments. The differences could also be that the effect
of L. littorea is significant only at densities higher than the 17 mÀ 2 found on these
specific shores, or if they can be excluded for a longer time than the 3 months in this
study, as the juveniles of A. nodosum have slow growth and therefore are exposed to the
action of grazers for several years. Qualitative observations on shores in Roscoff
(Brittany, France) and on the Isle of Man show that, in some cases, large numbers of
˚
A. nodosum juveniles are found in patches of Fucus spp. juveniles (P.A., personal
observation). Despite the dense growth of Fucus spp., there was no such effect in this
study. However, the results in September 2001 with decreasing cover of Fucus spp. due
to growth of surrounding adult A. nodosum plants, regrowth of small holdfast, and
patchily distributed juvenile A. nodosum found in the plots despite the presence of
Littorina spp. suggested that the small patches would return to their initial stage.
Acknowledgements
We are grateful to the staff at Tjarno Maine Biological Laboratory for their help,
¨ ¨
hospitality, and patience. We are also grateful to Go
¨ran Nylund for help with monitoring
in the field. This study was funded by a studentship to G.C. from the Faculty of Science,
Goteborg University, as part of a PhD thesis in Marine Botany. Additional funding
¨
for long-term experimental work was provided by the EU-Mast III project Eurorock
MAS3-CT95-0012.[AU]
References
˚
Aberg, P., 1992. A demographic study of two populations of the seaweed Ascophyllum nodosum. Ecology 73,
1473 – 1487.
˚
Aberg, P., Pavia, H., 1997. Temporal and multiple scale spatial variation in juvenile and adult abundance of the
brown alga Ascophyllum nodosum. Mar. Ecol. Prog. Ser. 158, 111 – 119.
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 49
Anderson, M.J., 2000a. PCOORD: A FORTRAN Computer Program for Principal Coordinate Analysis. Depart-
ment of Statistics, University of Auckland.
Anderson, M.J., 2000b. NPMANOVA: A FORTRAN Computer Program for Non-Parametric Multivariate
Analysis of Variance (for Any Two-Factor ANOVA Design) Using Permutation Tests. Department of
Statistics, University of Auckland.
Anderson, M.J., 2001. Permutation tests for univariate or multivariate analysis of variance and regression. Can. J.
Fish. Aquat. Sci. 58, 623 – 639.
Brawley, S.H., 1992. Fertilization in natural populations of the dioecious brown alga Fucus ceranoides and the
importance of the polyspermy block. Mar. Biol. 113, 145 – 157.
˚
Cervin, G., Aberg, P., 1997. Do littorinids affect the survival of Ascophyllum nodosum germlings? J. Exp. Mar.
Biol. Ecol. 218, 35 – 47.
Chapman, A.R.O., 1995. Functional ecology of fucoid algae: twenty-three years of progress. Phycologia 34, 1 – 32.
Clarke, K.R., Warwick, R.M., 1994. Change in Marine Communities: An Approach to Statistical Analysis and
Interpretation Plymouth Marine Laboratory, Natural Environmental Research Council, UK. 144 pp.
Connell, J.H., Slatyer, R.O., 1977. Mechanisms of succession in natural communities and their role in community
stability and organization. Am. Nat. 111, 1119 – 1144.
Dudgeon, S., Petraitis, P.S., 2001. Scale-dependent recruitment and divergence of intertidal communities. Ecol-
ogy 82, 991 – 1006.
Dudgeon, S., Kubler, J.E., Wright, W.A., Vadas, R.L., Petraitis, P.S., 2001. Natural variability in zygote dispersal
of Ascophyllum nodosum at small spatial scales. Funct. Ecol. 15, 595 – 604.
Farrell, T.M., 1991. Models and mechanisms of succession: an example from a rocky intertidal community. Ecol.
Monogr. 61, 95 – 113.
Jenkins, S.R., Norton, T.A., Hawkins, S.J., 1999a. Settlement and post-settlement interactions between Semi-
balanus balanoides (L.) (Crustacea: Cirripedia) and three species of fucoid canopy algae. J. Exp. Mar. Biol.
Ecol. 236, 49 – 67.
Jenkins, S.R., Hawkins, S.J., Norton, T.A., 1999b. Direct and indirect effects of a macroalgal canopy and limpet
grazing in structuring a sheltered inter-tidal community. Mar. Ecol. Prog. Ser. 188, 81 – 92.
Jenkins, S.R., Norton, T.A., Hawkins, S.J., 1999c. Interactions between canopy forming algae in the eulittoral
zone of sheltered rocky shores on the Isle of Man. J. Mar. Biol. Assoc. U.K. 79, 341 – 349.
Johannesson, K., 1989. The bare zone of Swedish rocky shores: why is it there? Oikos 54, 77 – 86.
Jones, C.G., Lawton, J.H., Shachack, M., 1994. Organisms as ecosystem engineers. Oikos 69, 373 – 386.
Lewis, J.R., 1964. The Ecology of Rocky Shores. The English Universities Press, London. 314 pp.
˚
Lindegarth, M., Aberg, P., Cervin, G., Nilsson, P., 2001. Effects of grazing on the structure of mid-shore,
intertidal assemblages on moderately exposed rocky shores of the Swedish west coast. Mar. Ecol. Prog.
Ser. 212, 29 – 38.
Lubchenco, J., 1983. Littorina and Fucus: effects of herbivores, substratum heterogeneity, and plant escapes
during succession. Ecology 64, 1116 – 1123.
Mathieson, A.C., Penniman, C.A., Busse, P.K., Tveter-Gallagher, E., 1982. Effects of ice on Ascophyllum
nodosum within the Great Bay estuary system of New Hampshire – Maine. J. Phycol. 18, 331 – 336.
McCook, L.J., 1994. Understanding ecological community succession. Vegetation 110, 115 – 147.
Petraitis, P.S., Latham, R.E., 1999. The importance of scale in testing the origins of alternative community states.
Ecology 80, 429 – 442.
Rueness, J., 1977. Norsk Algeflora. Scandinavian University Books, Oslo.
˜
Serrao, E.A., Pearson, G., Kautsky, L., Brawley, S.H., 1996. Successful external fertilization in turbulent environ-
ments. Proc. Natl. Acad. Sci. U. S. A. 93, 5286.
˜
Serrao, E.A., Brawley, S.H., Hedman, J., Kautsky, L., Samuelsson, G., 1999. Reproductive success of Fucus
vesiculosus (Phaeophyceae) in the Baltic Sea. J. Phycol. 35, 254 – 269.
Sousa, W.P., 1984. The role of disturbance in natural communities. Annu. Rev. Ecol. Syst. 15, 353 – 391.
˚
Viejo, R.M., Aberg, P., Cervin, G., Lindegarth, M., 1999. The interactive effects of adult canopy, germling
density and grazing on germling survival of the rockweed Ascophyllum nodosum. Mar. Ecol. Prog. Ser. 187,
113 – 120.
302 (2004) 35 – 49
www.elsevier.com/locate/jembe
Effects of small-scale disturbances of canopy and
grazing on intertidal assemblages on the
Swedish west coast
˚
Gunnar Cervin a,*, Mats Lindegarth b, Rosa M. Viejo a,1, Per Aberg a
a
Department of Marine Ecology, Marine Botany, Goteborg University, Box 461, Goteborg SE-405 30, Sweden
¨ ¨
b
Department of Marine Ecology, Tjarno Marine Biological Laboratory, Goteborg University,
¨ ¨ ¨
Stromstad SE-452 96, Sweden
¨
Received 13 March 2003; received in revised form 23 July 2003; accepted 4 September 2003
Abstract
The effects of small-scale disturbances (80 Â 30-cm plots) of canopy and grazers on intertidal
assemblages were investigated in this 4-year experiment on sheltered rocky shores on the Swedish
west coast. Canopy disturbances due to ice scouring were mimicked by removal of adult plants of the
seaweed Ascophyllum nodosum (L.) Le Joli. Density of the main epilithic grazing gastropods,
Littorina spp., was lowered by exclosure and handpicking. Based on earlier experiments in other
areas, the general hypothesis was that canopy removal and grazer exclosure, alone or in combination,
should increase the recruitment of A. nodosum or other fucoid juveniles, and change the structure of
the understorey assemblage.
There was an effect of canopy removal on the development of this assemblage, lasting for more
than 31 months. Both increased and decreased abundances of species were found as short-term
effects, but there was also a longer-term effect with increased abundance. Grazer exclosure was only
effective in combination with canopy removal, causing a short-term increase in ephemeral green
algae. Short-term effects of canopy removal were also the increase in recruitment of Semibalanus
balanoides (Linnaeus) and the decrease of the red alga Hildenbrandia rubra (Sommerfelt)
Meneghini. Fast recruitment and growth of fucoid species (Fucus serratus L. and F. vesiculosus L.)
restored the canopy and conditions of the understorey within 18 months. Thus, the canopy removal
changed the physical conditions for the understorey, making it possible for other species to coexist in
this community. Surprisingly, no effect of canopy removal or grazer exclusion was found on the
recruitment of juvenile A. nodosum, neither by canopy removal nor grazer exclosure. The lack of
* Corresponding author. Tel.: +46-31-7732709; fax: +46-31-7732727.
E-mail address: gunnar.Cervin@marbot.gu.se (G. Cervin).
1 ´ ´
Current address: Area de Biodiversidad y Conservacion, Escuela Superior de Ciencias, Experimentales y
´ ´ ´
Tecnologıa, Universidad Rey Juan Carlos, Tulipan s/n, Mostoles, Madrid E-28933, Spain.
0022-0981/$ - see front matter D 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jembe.2003.09.022
36 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
such effects might be due to the early mortality caused by other grazers (small, mobile crustaceans),
or to the low density of periwinkles on these shores. However, despite the patchy and generally low
recruitment of A. nodosum juveniles, observations suggested that the cover of A. nodosum in
manipulated patches would return to initial levels, either by recruitment or regrowth of small
holdfasts and from growth of edge plants.
D 2003 Elsevier B.V. All rights reserved.
Keywords: Adult canopy; Ascophyllum nodosum; Grazing; Fucus; Recruitment; Littorina
1. Introduction
Rocky shores as well as most natural communities are characterized by spatial
heterogeneity and temporal dynamics (Sousa, 1984). This may be a response to
physical disturbances and biological interactions, which commonly cause communities
to undergo unpredictable changes in species abundances, or a predictable temporal
sequence of species replacements. These changes often involve the appearance or
dominance of plant species with progressively greater maximum size, age, and shade
tolerance, and progressively lower maximum growth rates and dispersal abilities
(McCook, 1994). Disturbances, such as formation of gaps in spatial cover or increased
levels of resources, facilitate the establishment of early colonizers, even when they are
eventually excluded from the patch by competitively dominant species. The presence
of dominant, structure-forming species, on the other hand, may facilitate the growth
and/or survival of associated or ‘‘accompanying’’ species in natural assemblages, acting
as ‘‘physical engineers’’ (sensu Jones et al., 1994), ameliorating the environment. For
example, red algal species that are otherwise found sublittoraly can prosper under
dense fucoid canopies in the intertidal.
In sheltered rocky shores of the temperate region of the north Atlantic, the seaweed
Ascophyllum nodosum (L.) Le Joli is one of the major canopy-forming species in the mid-
intertidal zone. Turf forming and particularly encrusting species dominate the understorey
in shores on the Swedish west coast (G.C., personal observation). Other canopy-forming
species above the A. nodosum zone are Fucus spiralis L., and below there are F. serratus
L. and F. vesiculosus L. (Lewis, 1964). These are, however, considered competitively
inferior to A. nodosum in the mid-intertidal (Lewis, 1964). Nevertheless, F. serratus and F.
vesiculosus are sometimes also found as smaller or larger patches within the A. nodosum
canopy (Jenkins et al., 1999b). An important agent of disturbance of these canopies in the
NW Atlantic and Swedish west coast is ice scouring, creating gaps of different sizes
˚
(Mathieson et al., 1982; Aberg, 1992; Dudgeon and Petraitis, 2001), with small-scale
˚
losses typically occurring on sheltered shores of the Swedish west coast (P.A., personal
observation). The losses of A. nodosum due to ice scouring may be more than 50% of the
˚
biomass during years of extreme ice cover (Mathieson et al., 1982; Aberg, 1992). Previous
studies have indicated that the effect of gaps on the recruitment of A. nodosum and the
development of these communities are size-dependent (Dudgeon and Petraitis, 2001). The
presence of grazers in combination with disturbances may affect the sequence and rate of
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 37
colonization by ephemerals and recruitment of fucoids. Studies on the Isle of Man
suggested interactions among canopy, red algal turf, and limpets (Jenkins et al., 1999b).
For example, in areas with dense cover of A. nodosum over rocks kept bare by grazing
limpets, Jenkins et al. (1999b) found that combined removal of both canopy and grazers
caused a massive growth of Fucus spp. Previous studies on the Swedish west coast, where
limpets are lacking, have demonstrated that high densities of L. littorea (Linnaeus)
˚
increase the mortality of A. nodosum juveniles (Cervin and Aberg, 1997). Complex
interactions of canopy disturbance, juvenile densities, and Littorina spp. have also been
demonstrated for the survival of postsettlement stages of A. nodosum, and it was
hypothesized that grazing also may affect the survival of older juvenile stages (Viejo et
al., 1999). However, grazing by the dominant littorinid [Littorina saxatilis (Olivi)] on
semiexposed shores on the Swedish west coast does not appear to significantly affect algal
assemblages (Lindegarth et al., 2001).
This paper presents the results of a manipulative experiment on sheltered shores of
the Swedish west coast. The experiment was designed to test the effect of small-scale
disturbances to the canopy species A. nodosum typical for this area (canopy removal)
and grazing by periwinkles (grazing exclosure) on the recruitment of fucoids,
particularly A. nodosum, and the development of associated assemblages. For A.
nodosum and Fucus spp., the specific hypothesis is that we expect higher recruitment
in canopy removal and grazer exclosure in combination. This is the first time that such
an experiment has been done on sheltered shores on the Swedish west coast, with
almost atidal conditions and irregular heavy ice winters, which make it different from
similar experiments done in other parts of the world.
2. Materials and methods
2.1. Study area
The study started in April 1997 and was conducted in the northernmost part of the
Swedish west coast, close to the Tjarno Marine Biological Laboratory (TMBL) (N58j52.6V,
¨ ¨
E11j9.0V). This area consists of numerous granite islands and islets. The surface water is
influenced by fully saline North Sea water, brackish water from the Baltic Sea, and
freshwater from the river Glomma in southeastern Norway, giving a surface salinity from
15 to 30 psu in the study area. The air temperature frequently drops below 0 jC, and ice
winters are common. The regular tidal range is less than 0.3 m, but the water level can vary
up to 2 m, depending on atmospheric pressure and winds (Johannesson, 1989).
2.2. Experimental procedures
Effects of the presence of macroalgal canopy and grazing periwinkles on the
recruitment of canopy-forming fucoids and the development of the understorey were
tested in a two-factor experiment. The experimental factors were Canopy {Intact vs.
Removed (CI vs. CR)} and Grazer {Fences, Procedural Fence Control, and Unmanip-
ulated Control (FE, PC, and CO)}. Three replicates were used for each treatment. The
38 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
experimental plots were randomly allocated to areas with dense populations of A.
nodosum on different islands within 2 km of TMBL. Due to the narrow intertidal zone
that restricts the vertical distribution of A. nodosum, the plots were 80 cm wide by 30
cm high.
In mid April 1997, before A. nodosum release their gametes, adult plants including
their holdfasts were removed in nine plots (CR). Nine other plots were marked but
canopies were left intact (CI). Adult plants around each cleared plot were trimmed to
prevent large plants from reaching into the plots. To prevent periwinkles from entering the
plots, fences (FE) of stainless wire mesh with a mesh size of 0.5 cm were used
(80 Â 30 Â 5 cm). A procedural fence control (PC) consisting of partial barriers that
protected the corners of plots was also included in the experiment. These were included to
test for experimental artefacts, such as changes in flow conditions and shading, that might
be caused by fences, without preventing access for grazing snails. Thus, if there were
significant artefacts due to fences, controls without any fences would differ from both
fence treatments (CO p PC = FE). If, on the other hand, there were no experimental
artefacts but effects of removing the grazers, plots with fences would differ from the plots
where snails had access (CO = PC p FE). Fences were attached to the rock using stainless
screws and plastic expanders, and two screws marked the control plots without fences
(CO).
The number of Littorina spp. on those shores where the experiment was performed
was very variable, with a mean of 17 L. littorea mÀ 2 and 17 L. mariae Sacchi and
Rastelli/L. obtusata (Linnaeus) mÀ 2. During the experiment, snails were observed
inside and climbing over experimental fences. Therefore, plots with fences were cleared
of gastropods every 2 days during the first 4 weeks, and weekly for another 2 months.
The periwinkle grazers were not completely excluded from these areas, but there was a
clear reduction in densities during the first 3 months after gamete release. On the
Swedish west coast, A. nodosum reproduce during a short period of 1 –2 weeks, but
depending on water temperature and other physical factors such as periods of low
water in combination with warm air temperature, the onset of the short reproductive
period can vary among years from mid April to late in May. The other canopy-forming
species, F. serratus and F. vesiculosus, reproduce in winter to early spring and April to
June, respectively (Rueness, 1977). Since the treatments were not maintained over the
whole experimental time, experimental manipulation can be seen as a pulse-type
perturbation.
The percentage cover of different species was measured by the intersection of 30
random points in each plot in April, July, and October 1997; November 1998; and
November 1999. Algae were identified to species or genus, and animals were usually
identified to species. Five randomly placed subplots (5 Â 5 cm) were used to estimate the
density of juvenile A. nodosum in July and October 1997. In November 1998 and 1999,
and September 2001, the whole area of the plots was searched for juvenile A. nodosum.
2.3. Statistical analyses
Patterns of whole assemblages were explored at individual times (t = 0, 3, 6, 19, and 31
months) using metric multidimensional scaling (MDS) (using routines in Anderson,
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 39
Fig. 1. MDS of assemblages at different times of sampling. Empty and filled symbols are plots with and without
canopy, respectively. Diamonds, squares, and triangles are plots with fences, control fences, and plots without
fences, respectively. Numbers indicate cumulative percent variability explained by the two first principal
components.
40
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
Table 1
NPMANOVA on benthic assemblages at different times of sampling
Source df Start 3 months 6 months 19 months 31 months 53 months
(April 1997) (July 1997) (October 1997) (November 1998) (November 1999) (September 2001)
MS F p MS F p MS F p MS F p MS F p MS F p
Canopy 1 628 0.80 >0.5 5474 5.92 < 0.01 10,832 10.04 < 0.01 4305 3.14 < 0.02 4641 5.99 < 0.01 Not analysed
Fence 2 442 0.57 >0.5 510 0.55 >0.5 1401 1.30 >0.2 742 0.54 >0.5 1583 2.04 >0.09
Canopy  Fence 2 1666 2.13 < 0.05 1221 1.32 >0.2 1203 1.11 >0.2 1781 1.30 >0.2 1162 1.50 >0.2
Residual 12 782 924 1079 1373 775
See text for further details about analytical procedures.
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 41
2000a). Explicit tests of hypotheses about effects of experimental treatments on whole
assemblages were tested using nonparametric multivariate analysis of variance (NPMA-
NOVA) and a posteriori pairwise t tests (Anderson, 2000b). Tests for effects of treatments
were done using random permutation of data among treatments (Anderson, 2001). All
multivariate analyses were done using Bray –Curtis distances on fourth-root transformed
data (Clarke and Warwick, 1994). Strength of treatment effects was quantified using
average Bray – Curtis dissimilarities among and within experimental treatments at each
time of sampling.
Univariate analyses were done using two-way ANOVA. Heterogeneity of variances was
tested for using Cochran’s test. If necessary, data were transformed to remove heteroge-
neity. In some instances, transformation of data was not successful in reducing the
heterogeneity of variances. In these cases, we did the ANOVA on untransformed data,
but these tests should be treated with some caution since heterogeneous variances can
increase the risk of Type I error.
The dominant components in these assemblages and those included in multivariate
analyses were: Ahnfeltia plicata (Hudson) Fries, Calothrix spp., Ceramium spp.,
Chondrus crispus Stackhouse, Cladophora rupestris (L.) Kutzing, Dynamena pumila
¨
(Linnaeus), Enteromorpha spp., Electra pilosa (Linnaeus), Fucus spp., Halichondria
panicea (Pallas), Hildenbrandia rubra (Sommerfelt) Meneghini, Hydroidea spp.,
Mytilus edulis Linnaeus, Phycodrys rubens (Linnaeus) Batters, Phymatolithon spp.,
Polysiphonia spp., Ralfsia verrucosa (Areschoug) Areschoug, Sagartiidae, Semibanla-
nus balanoides (Linnaeus), Sphacelaria spp., Spirorbis spirorbis (Linnaeus), Ulothrix
spp., Urospora spp., and Ulva spp. The species that were manipulated (A. nodosum
and Littorina spp.) were not included in the multivariate analyses. Univariate analyses
were done on juvenile A. nodosum and on commonly found species in the plots.
Fig. 2. Dissimilarity between (filled circles) and within (empty circles) canopy treatments at different monitoring
dates.
42
Table 2
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
ANOVAs of specific parameters at different monitoring dates
Source df Start 3 months 6 months 19 months 31 months 53 months
(April 1997) (July 1997) (October 1997) (November 1998) (November 1999) (September 2001)
MS F p MS F p MS F p MS F p MS F p MS F p
(a) The number of A. nodosum juveniles
Canopy 1 Not measured 0.06 4.17 >0.05 0.009 0.50 >0.4 56.9 1.48 >0.2 14.2 1.14 >0.3 220 2.35 >0.1
Fence 2 0.002 0.17 >0.8 0.03 1.50 >0.2 92.2 2.41 >0.1 0.39 0.031 >0.9 26.0 0.28 >0.7
Canopy  Fence 2 0.002 0.17 >0.8 0.06 3.50 >0.06 74.1 1.94 >0.1 5.06 0.40 >0.6 24.0 0.26 >0.7
Residual 12 0.01 0.02 38.2 12.5 93.7
Cochran’s C test ( p) < 0.05 ns < 0.05 ns < 0.05
(b) The cover of Fucus spp.
Canopy 1 22.22 0.391 >0.5 672.2 2.637 >0.1 70.44 19.15 < 0.001 7160 21.83 < 0.001 13,889 15.16 < 0.005 1422 2.18 >0.1
Fence 2 56.17 0.989 >0.4 68.52 0.269 >0.7 2.386 0.6489 >0.5 152.0 0.4635 >0.6 605.6 0.661 >0.5 1184 1.82 >0.2
Canopy  Fence 2 124.1 2.185 >0.1 12.96 0.0508 >0.9 2.918 0.7935 >0.4 77.56 0.2365 >0.7 205.6 0.224 >0.8 310 0.47 >0.6
Residual 12 56.8 255 3.68 328 917 653
Cochran’s C test ( p) < 0.05 < 0.05 ns after ns ns ns
sqrt(x + 1)
transformation
(c) The cover of ephemeral green algae
Canopy 1 0.62 0.03 >0.8 1867 8.72 < 0.05 5.56 1.80 >0.2 0.89 1.80 >0.2 Not measured, Not measured
as the percentage
cover was less
than one
Fence 2 7.41 0.34 >0.7 1613 7.53 < 0.01 1.85 0.60 >0.5 0.30 0.60 >0.5
Canopy  Fence 2 32.1 1.49 >0.2 1671 7.80 < 0.01 1.85 0.60 >0.5 0.30 0.60 >0.5
Residual 12 21.6 214 3.09 0.49
Cochran’s C test ( p) < 0.05 < 0.05 < 0.05 < 0.05
(d) The cover of H. rubra
Canopy 1 158 0.41 >0.5 3024 8.93 < 0.05 6296 8.38 < 0.05 601 0.58 >0.4 987 1.60 >0.2 Not measured
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
Fence 2 1067 2.76 >0.1 169 0.50 >0.6 121 0.16 >0.8 384 0.37 >0.7 61.7 0.10 >0.9
Canopy  Fence 2 549 1.42 >0.2 230 0.68 >0.5 128 0.17 >0.8 281 0.27 >0.7 151 0.24 >0.7
Residual 12 387 339 751 1035 617
Cochran’s C test ns ns ns ns ns
(e) The cover of S. balanoides
Canopy 1 1543 8.87 < 0.05 7.70 12.6 < 0.005 183 1.01 >0.3 0.00 0.00 1 0.00 0.00 1 Not measured
Fence 2 85.8 0.49 >0.6 1.78 2.91 >0.09 178 0.98 >0.4 0.61 0.50 >0.6 0.62 0.50 >0.6
Canopy  Fence 2 260 1.49 >0.2 1.78 2.91 >0.09 30.3 0.17 >0.8 1.85 1.50 >0.2 1.85 1.50 >0.2
Residual 12 174 0.61 181 1.23 1.23
Cochran’s C test ns ns after ns ns
sqrt(x + 1)
transformation
43
44 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
3. Results
3.1. Multivariate analyses
Two-dimensional ordinations of assemblages using metric MDS explained between
63% and 75% of the total variability among plots at individual times. The ordinations
indicated that there were differences between experimental treatments, but these changed
among times (Fig. 1). Different assemblages developed in plots with and without a
canopy of A. nodosum. This effect was most evident after 6 months of the start of the
experiment (i.e., October 1997). Actually, multivariate analyses of effects at individual
times showed that there were significant effects of canopy removal at all times of
sampling, except at the start of the experiment (Table 1). Thus, it can be concluded that
the abundance and composition of these assemblages were affected by the removal of the
canopy already after 3 months and that the effect persisted for at least 31 months.
Quantitative estimates of average dissimilarities indicate that those within experimental
treatments changed little, but dissimilarities between treatments peaked after 6 months
(Fig. 2). No effects of grazing treatment were detected (Table 1). This is consistent with
visual impressions represented by the MDS (Fig. 1). One surprising result was that there
was a significant interaction effect at the start of the experiment (Table 1). No significant
differences were, however, detected using a posteriori tests. Nevertheless, the largest
difference at the start of the experiment was observed in plots without fences, where there
was an average dissimilarity of between 48% and 34% within canopy treatments. This
difference between canopy treatments was clearly smaller than what was observed at
subsequent times of sampling (e.g., more than 60% difference between canopy and no
canopy for all fencing treatments in October 1997; Fig. 2).
3.2. Univariate analyses
No effect of canopy removal or grazer exclusion was found on the recruitment of
juvenile A. nodosum after 32 months, nor after an additional time of sampling after 53
months (Table 2a, Fig. 3A). Plants surrounding plots that had been removed of canopy
were trimmed at the start so that they would not reach into the plots. These trimmed
plants had grown in size so that they, in November 1999, had begun to cover the plots.
At the start of the experiment and after 3 months, there were no effects of any of the
experimental treatments on the cover of Fucus spp. In October 1997, however, a large
number of juvenile Fucus spp. had recruited into plots where the canopy was removed,
especially in the plots where L. littorea was also excluded (Fig. 3B). The percent cover
of Fucus spp. was significantly higher in the plots without a canopy of A. nodosum than
in those with an intact canopy, but no significant effects of grazer exclosure were
detected (Table 2b, Fig. 3B). There was also a significant effect of canopy removal on
Fig. 3. (A – E) The number of A. nodosum juveniles < 20 mm per 0.8 Â 0.3 m, and the percentage cover of Fucus
spp., ephemeral green algae (mainly Enteromorpha spp. and Ulva spp.), H. rubra, and S. balanoides at the
monitoring dates. Less than 1% cover is not shown (symbol meanings: CI = canopy intact, CR = canopy removal,
FE = fence, PC = procedural fence control, CO = control).
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 45
46 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
the cover of Fucus spp. in November 1998 and 1999, and in plots with canopy removed,
the percent cover of Fucus spp. had increased at these dates. In September 2001,
however, the cover of Fucus spp. had declined in these plots and no significant effect of
canopy removal was found (Table 2b, Fig. 3B).
In July 1997, there was a significant interactive effect of canopy removal and grazer
exclosure on the cover of the ephemeral green algae (mainly Enteromorpha spp. and
Ulva spp.). In plots where grazers were excluded and canopy was removed, the mean
cover of ephemeral green algae was 58.9 F 20.6% (mean F S.E.), while in the proce-
dural controls and controls, there was very little growth of ephemeral green algae (Table
2c, Fig. 3D). Thus, at the only time, and for the only taxon that showed an effect of
grazing treatment, there was no indication of procedural artefacts (i.e., FE > CO = PC for
CR; FE = CO = PC for CI). The effect was not found at later dates.
There was also a significant effect of canopy removal on the red algal crust H. rubra
(Table 2d, Fig. 3D) and the barnacle Semibalanus balanoides (Table 2e, Fig. 3E) in July
and October 1997. The percent cover of H. rubra was lower in plots with no canopy, while
the percent cover of barnacles was higher in these plots. After October 1997, no significant
effect of canopy removal was detected for either the cover of H. rubra or barnacles (Table
2d and e, Fig. 3D and E). It is notable that this coincided with the development of a Fucus
spp. canopy in the plots where the canopy of A. nodosum had been removed (Fig. 3B).
4. Discussion
The results of the study showed that small-scale disturbances in the canopy of A.
nodosum significantly affected the development of intertidal assemblages. Some of these
effects persisted for more than 2 years, while others were more short-lived.
Short-term effects of small-scale disturbances were both positive and negative. The
abundance of the barnacle S. balanoides increased as a result of canopy disturbance,
recruiting in greater numbers in the absence of the canopy. It is possible that the canopy
prevented recruitment by acting as a physical barrier, or by sweeping away settling larvae.
The increase in cover of S. balanoides was short-lived, as the later appearance of a Fucus
spp. canopy reduced subsequent recruitment and increased the mortality of barnacles. The
three fucoid species A. nodosum, F. serratus, and F. spiralis have presently been shown to
have a negative effect on settlement and early recruitment of S. balanoides (Jenkins et al.,
1999a). Not only sweeping of algal fronds are important, but also chemical or physical
cues from canopy algae may discourage the larvae (Jenkins et al., 1999a).
The presence of adult plants of A. nodosum for understorey species is similar to other
dominant species with high biomass and persistent structures, such as trees or corals,
which can act as ‘‘physical engineers’’ (sensu Jones et al., 1994), ameliorating the
environment. As a consequence, species like H. rubra are affected negatively by removal
of the canopy. Colonization of Fucus spp. will restore the canopy and ameliorate the
environment for these understorey species. Another short-term effect of canopy removal,
in combination with grazer exclusion, was the development of a cover of ephemeral green
algae. This growth of ephemeral green algae also showed that the grazer exclosure was
effective during the first 3 months of the experiment. However, the removal of grazers was
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 47
only intense during the first months of the experiment and strong effects of grazers may
not be so likely at later times.
A longer-term effect of the canopy removal of A. nodosum was the development of a
canopy of Fucus spp., which, as juveniles, are fast-growing species compared to A.
nodosum. The high number of recruiting juvenile Fucus spp. and their fast growth are the
reasons why we measured them as percentage cover. F. serratus and F. vesiculosus are
both competitively inferior canopy species to A. nodosum (Jenkins et al., 1999c), and, as
such, do prosper in the absence of an A. nodosum canopy. In intertidal rocky shores,
ephemeral species such as Ulva spp. and Enteromorpha spp. are commonly early
colonizers after disturbances, and a general pattern is that they inhibit or at least retard
the appearance of canopy-forming species, such a Fucus spp. (e.g., Lubchenco, 1983). The
mechanism of inhibition in succession (envisioned by Connell and Slatyer, 1977) did not
seem to work on intertidal sheltered shores on the Swedish west coast. The presence of
ephemeral green algae did not prevent the development of fucoid species. If anything, the
effect was to increase the rate of development of a canopy of Fucus spp. The first summer
was hot with extreme low water, which caused the plants to be emerged for several days
(authors, personal observation), and thus the ephemeral green algae might have protected
recruiting fucoids from desiccation. The effect of ephemeral green algae may vary from
year to year, depending on temperature and atmospheric pressure. The indirect effect of
grazers would then retard the succession by eliminating the protective cover of ephemeral
green algae, in concordance with the model of Farrell (1991). Previous results have also
suggested a positive effect of ephemeral green algae for the survival of the early stages of
A. nodosum germlings (Viejo et al., 1999).
The experiment showed that small-scale disturbance of the A. nodosum canopy created
the opportunity for colonization of other species, such as Fucus spp. and, in interaction
with grazer exclusion, colonisation of ephemeral green algae. Ice scouring may thus open
‘‘windows’’ for coexistence of species, even if A. nodosum is a superior competitor on
sheltered shores. In this assemblage, as in many others, the natural disturbances facilitate
the coexistence of species. Previous studies have also suggested that if the size of the
disturbance is large enough, the removal of adult A. nodosum may initiate the establish-
ment of alternative, successional endpoints, such as assemblages dominated by barnacles
or mussels (Petraitis and Latham, 1999). This would probably not occur on the Swedish
west coast, due to the highly variable exposures at a smaller scale on those shores, always
leaving some remnants of the original population within dispersal distance.
On these sheltered rocky shores on the Swedish west coast where A. nodosum is
dominant, small-scale disturbances of Littorina spp. did not increase the recruitment of
juvenile A. nodosum. This is different from previous and similar small-scale studies
made on the Isle of Man, where removal of the main grazer Patella vulgata caused much
higher numbers of juvenile A. nodosum to recruit within 2 years, and the first year also
in interaction with canopy removal (Cervin et al., in preparation). These areas have the
˚
same naturally patchy distribution of juvenile A. nodosum (Aberg and Pavia, 1997), and
we believe that similar plant – animal interactions would work in these areas. The access
of A. nodosum zygotes should be massive, as the egg rain is estimated to be 2.5 Â 109
˚
eggs mÀ 2 (Aberg and Pavia, 1997) and as the fertilisation probably is high since other
studies of fucoids have shown high fertilisation success (>90%) (e.g., Brawley, 1992;
48 G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49
˜
Serrao et al., 1996, 1999). Furthermore, previous studies in the NW Atlantic have shown
that dispersal of A. nodosum zygotes is very variable, but that more than 50% of the
zygotes disperse more than 6 m (Dudgeon et al., 2001), with an estimated maximum of
30 m (Chapman, 1995). Moreover, previous studies have also indicated that sexual
recruitment may be affected by the absence of a canopy of adults and by the size of
these clearings (e.g., Dudgeon et al., 2001; Dudgeon and Petraitis, 2001). Viejo et al.
(1999) have shown that canopy removal of the same size as in this experiment does not
alter the postsettlement survival of A. nodosum juveniles. Neither is there a higher
˚
presettlement mortality in this size of canopy removal, according to earlier studies (P.A.,
unpublished data). However, despite zygote availability, canopy removal, and successful
exclusion of L. littorea during the first 3 months after spawning of A. nodosum gametes,
no effect could be seen at the monitoring dates. Why were no effects of canopy removal
or grazer exclusion, alone or in combination, detected here? Although this was a long-
time experiment, the temporal scale might be too short for such a long-lived plant as A.
nodosum. The differences could be at the level of settlement patterns, where mortality
occurs before the zygotes have managed to establish, due to grazing by fast-moving
crustaceans not included in the treatments. The differences could also be that the effect
of L. littorea is significant only at densities higher than the 17 mÀ 2 found on these
specific shores, or if they can be excluded for a longer time than the 3 months in this
study, as the juveniles of A. nodosum have slow growth and therefore are exposed to the
action of grazers for several years. Qualitative observations on shores in Roscoff
(Brittany, France) and on the Isle of Man show that, in some cases, large numbers of
˚
A. nodosum juveniles are found in patches of Fucus spp. juveniles (P.A., personal
observation). Despite the dense growth of Fucus spp., there was no such effect in this
study. However, the results in September 2001 with decreasing cover of Fucus spp. due
to growth of surrounding adult A. nodosum plants, regrowth of small holdfast, and
patchily distributed juvenile A. nodosum found in the plots despite the presence of
Littorina spp. suggested that the small patches would return to their initial stage.
Acknowledgements
We are grateful to the staff at Tjarno Maine Biological Laboratory for their help,
¨ ¨
hospitality, and patience. We are also grateful to Go
¨ran Nylund for help with monitoring
in the field. This study was funded by a studentship to G.C. from the Faculty of Science,
Goteborg University, as part of a PhD thesis in Marine Botany. Additional funding
¨
for long-term experimental work was provided by the EU-Mast III project Eurorock
MAS3-CT95-0012.[AU]
References
˚
Aberg, P., 1992. A demographic study of two populations of the seaweed Ascophyllum nodosum. Ecology 73,
1473 – 1487.
˚
Aberg, P., Pavia, H., 1997. Temporal and multiple scale spatial variation in juvenile and adult abundance of the
brown alga Ascophyllum nodosum. Mar. Ecol. Prog. Ser. 158, 111 – 119.
G. Cervin et al. / J. Exp. Mar. Biol. Ecol. 302 (2004) 35–49 49
Anderson, M.J., 2000a. PCOORD: A FORTRAN Computer Program for Principal Coordinate Analysis. Depart-
ment of Statistics, University of Auckland.
Anderson, M.J., 2000b. NPMANOVA: A FORTRAN Computer Program for Non-Parametric Multivariate
Analysis of Variance (for Any Two-Factor ANOVA Design) Using Permutation Tests. Department of
Statistics, University of Auckland.
Anderson, M.J., 2001. Permutation tests for univariate or multivariate analysis of variance and regression. Can. J.
Fish. Aquat. Sci. 58, 623 – 639.
Brawley, S.H., 1992. Fertilization in natural populations of the dioecious brown alga Fucus ceranoides and the
importance of the polyspermy block. Mar. Biol. 113, 145 – 157.
˚
Cervin, G., Aberg, P., 1997. Do littorinids affect the survival of Ascophyllum nodosum germlings? J. Exp. Mar.
Biol. Ecol. 218, 35 – 47.
Chapman, A.R.O., 1995. Functional ecology of fucoid algae: twenty-three years of progress. Phycologia 34, 1 – 32.
Clarke, K.R., Warwick, R.M., 1994. Change in Marine Communities: An Approach to Statistical Analysis and
Interpretation Plymouth Marine Laboratory, Natural Environmental Research Council, UK. 144 pp.
Connell, J.H., Slatyer, R.O., 1977. Mechanisms of succession in natural communities and their role in community
stability and organization. Am. Nat. 111, 1119 – 1144.
Dudgeon, S., Petraitis, P.S., 2001. Scale-dependent recruitment and divergence of intertidal communities. Ecol-
ogy 82, 991 – 1006.
Dudgeon, S., Kubler, J.E., Wright, W.A., Vadas, R.L., Petraitis, P.S., 2001. Natural variability in zygote dispersal
of Ascophyllum nodosum at small spatial scales. Funct. Ecol. 15, 595 – 604.
Farrell, T.M., 1991. Models and mechanisms of succession: an example from a rocky intertidal community. Ecol.
Monogr. 61, 95 – 113.
Jenkins, S.R., Norton, T.A., Hawkins, S.J., 1999a. Settlement and post-settlement interactions between Semi-
balanus balanoides (L.) (Crustacea: Cirripedia) and three species of fucoid canopy algae. J. Exp. Mar. Biol.
Ecol. 236, 49 – 67.
Jenkins, S.R., Hawkins, S.J., Norton, T.A., 1999b. Direct and indirect effects of a macroalgal canopy and limpet
grazing in structuring a sheltered inter-tidal community. Mar. Ecol. Prog. Ser. 188, 81 – 92.
Jenkins, S.R., Norton, T.A., Hawkins, S.J., 1999c. Interactions between canopy forming algae in the eulittoral
zone of sheltered rocky shores on the Isle of Man. J. Mar. Biol. Assoc. U.K. 79, 341 – 349.
Johannesson, K., 1989. The bare zone of Swedish rocky shores: why is it there? Oikos 54, 77 – 86.
Jones, C.G., Lawton, J.H., Shachack, M., 1994. Organisms as ecosystem engineers. Oikos 69, 373 – 386.
Lewis, J.R., 1964. The Ecology of Rocky Shores. The English Universities Press, London. 314 pp.
˚
Lindegarth, M., Aberg, P., Cervin, G., Nilsson, P., 2001. Effects of grazing on the structure of mid-shore,
intertidal assemblages on moderately exposed rocky shores of the Swedish west coast. Mar. Ecol. Prog.
Ser. 212, 29 – 38.
Lubchenco, J., 1983. Littorina and Fucus: effects of herbivores, substratum heterogeneity, and plant escapes
during succession. Ecology 64, 1116 – 1123.
Mathieson, A.C., Penniman, C.A., Busse, P.K., Tveter-Gallagher, E., 1982. Effects of ice on Ascophyllum
nodosum within the Great Bay estuary system of New Hampshire – Maine. J. Phycol. 18, 331 – 336.
McCook, L.J., 1994. Understanding ecological community succession. Vegetation 110, 115 – 147.
Petraitis, P.S., Latham, R.E., 1999. The importance of scale in testing the origins of alternative community states.
Ecology 80, 429 – 442.
Rueness, J., 1977. Norsk Algeflora. Scandinavian University Books, Oslo.
˜
Serrao, E.A., Pearson, G., Kautsky, L., Brawley, S.H., 1996. Successful external fertilization in turbulent environ-
ments. Proc. Natl. Acad. Sci. U. S. A. 93, 5286.
˜
Serrao, E.A., Brawley, S.H., Hedman, J., Kautsky, L., Samuelsson, G., 1999. Reproductive success of Fucus
vesiculosus (Phaeophyceae) in the Baltic Sea. J. Phycol. 35, 254 – 269.
Sousa, W.P., 1984. The role of disturbance in natural communities. Annu. Rev. Ecol. Syst. 15, 353 – 391.
˚
Viejo, R.M., Aberg, P., Cervin, G., Lindegarth, M., 1999. The interactive effects of adult canopy, germling
density and grazing on germling survival of the rockweed Ascophyllum nodosum. Mar. Ecol. Prog. Ser. 187,
113 – 120.