Moore et al 07
MARINE ECOLOGY PROGRESS SERIES
Vol. 334: 11–19, 2007 Published March 26
Mar Ecol Prog Ser
Role of biological habitat amelioration in altering
the relative responses of congeneric species to
climate change
P. Moore1, 2, 4,*, S. J. Hawkins1, 2, 3, R. C. Thompson2
1
Marine Biological Association of the UK, The Laboratory, Citadel Hill, Plymouth PL1 2PB, UK
2
Marine Biology and Ecology Research Centre, School of Biological Sciences, University of Plymouth,
Plymouth PL4 8AA, UK
3
School of Biological Sciences, University of Southampton, Southampton SO16 7PX, UK
4
Present address: Centre of Excellence for Coral Reef Studies, Centre for Marine Studies, University Queensland,
Brisbane, 4072 Queensland, Australia
ABSTRACT: The distribution of most species is expected to alter in response to climate change. Pre-
dictions for the extent of these range shifts are frequently based on ‘climate envelope’ approaches,
which often oversimplify species responses because many do not consider interactions between
physical and biological factors. The local persistence of some species, however, is likely to be strongly
modulated by microhabitat-forming organisms. Using congeneric patellid gastropods with northern/
boreal and southern/lusitanian distributions, we have demonstrated how the loss of habitat-forming
macroalgal species could modify species responses to climate change. The northern limpet Patella
vulgata preferentially aggregates beneath Fucus spp. When Fucus vesiculosus was experimentally
removed, to simulate a decline in macroalgal abundance in response to climatic warming, P. vulgata
suffered increased mortality or relocated home scars, often to nearby Fucus spp. patches. In contrast,
the southern limpet P. depressa did not aggregate beneath Fucus spp. and showed no response in
terms of movement or mortality to the loss of F. vesiculosus. Based on these results, we predict that
the loss of Fucus spp. will influence the relative abundance of these 2 limpet species, particularly at
the distributional limit of Fucus spp. In addition, differences in the aggregative behaviour of these
limpet species will result in changes in the spatial distribution of grazing in the intertidal, with likely
consequences for community dynamics. These outcomes could not be anticipated from predictions
based on direct responses to temperature alone, highlighting the need for biotic and abiotic factors to
be incorporated into predictions of species responses to climate change.
KEY WORDS: Biological interactions · Biologically generated habitat · Climate change · Climate
envelope · Limpets · Macroalgae
Resale or republication not permitted without written consent of the publisher
INTRODUCTION temperature, is mapped in ‘climate space’ (the area in
which a species can survive because the environment
Global air temperatures have risen by 0.6 ± 0.2°C in is suitable). If the climatic space alters, then it is pre-
the last 100 yr and further increases of between 2 and dicted that the geographical distribution of species
3°C are predicted by 2100 (Hulme et al. 2002), prompt- found within that climatic space will alter accordingly
ing the need to understand how species and assem- (Pearson & Dawson 2003). This approach can be useful
blages respond to climate change. To date, forecasts of (Hodkinson 1999, Pearson & Dawson 2003), but it has
species-level responses have used a ‘climate envelope’ also been criticised because it tends to predict a spe-
approach, whereby a single climatic variable, usually cies’ fundamental (or potential) niche and does not
*Email: pippa.moore@uq.edu.au © Inter-Research 2007 · www.int-res.com
12 Mar Ecol Prog Ser 334: 11–19, 2007
take into account biological interactions, dispersal northward range retraction has already started, with
capability, habitat quality, together with habitat avail- some cold-temperate canopy-forming species, such as
ability and connectivity, all of which influence a spe- Laminaria spp. (Breeman 1990, A. E. F. Murias Dos
cies’ realised niche (Davis et al. 1998, Pearson & Daw- Santos unpubl. data), Himanthalia elongata and Pel-
son 2003). Here, we use the rocky intertidal as a vetia canaliculata (A. E. F. Murias Dos Santos unpubl.
convenient model system to investigate the role of data), already becoming scarcer at their southern
behaviourally mediated direct and indirect biological range limits.
interactions in modifying species’ responses to climate Canopy-forming algae influence community struc-
change and the likely consequence for community ture by modifying the environment (Menge 1978).
structure and dynamics. Higher concentrations of microalgal food are found
The British Isles straddle 2 major marine biogeo- beneath algal canopies (Thompson et al. 2004), where
graphic zones, with both warm lusitanian and cool environmental conditions are ameliorated and habitats
boreal biota (Lewis 1964). As a consequence, many of are generated for other species (Leonard 2000), often
the species present are at either the northern or south- resulting in increased diversity (e.g. Fucus spp. at mid-
ern edge of their biogeographic ranges (Southward et shore heights in Britain, Thompson et al. 1996; and
al. 1995), resulting in sets of ecologically comparable Cystoseira spp. at low-shore levels in the Mediter-
species, with northern and southern centres of distrib- ranean, Benedetti-Cecchi et al. 2001). Hence, it is
ution co-existing and with abundances which have likely that the presence/absence of canopy-forming
been shown to fluctuate with climatic changes (South- algae will have a disproportionate effect on the distrib-
ward 1967, Southward et al. 1995). Therefore, south- ution and diversity of species in the intertidal, with
west Britain provides an ideal region in which to implications for assemblage dynamics.
investigate how species with northern and southern Limpets are key grazers on rocky shores in the NE
biogeographic distributions respond to climate change Atlantic, and strongly influence community structure
and, in particular, the way in which biotic interactions and dynamics by controlling macroalgal abundance
may modify these responses. (Southward & Southward 1978, Hawkins et al. 1992).
A general poleward movement of species’ ranges in Two species of limpet co-exist at midshore levels on
response to global warming has been predicted moderately exposed shores of south-west Britain:
(Parmesan 1996), although see Helmuth et al. (2006) Patella vulgata, a cold-temperate/boreal limpet, dis-
for discussion of local factors, such as time of low tributed from northern Norway to southern Portugal,
water, influencing species’ range shifts. Evidence of and P. depressa, a southern/lusitanian limpet, distrib-
changes in the relative abundance of species with uted from Senegal in West Africa to north Wales, UK
northern and southern biogeographic distributions, as (Southward et al. 1995). It is predicted that P. depressa
well as poleward and upward latitudinal shifts in spe- will become the dominant limpet on the shores of
cies’ ranges, have been observed in a wide range of south-west Britain, at the expense of P. vulgata, in
taxonomic groups during the twentieth century response to increased climatic warming (Southward et
(Walther et al. 2002). These shifts are evident in both al. 1995), and there is evidence to suggest this is
aquatic and terrestrial habitats, including coastal already occurring on some shores (S. J. Hawkins
waters where plankton (Beaugrand & Ibanez 2004), unpubl. data).
intertidal organisms (Southward et al. 1995) and fish Both species of limpet exhibit ‘homing’ behaviour,
(Genner et al. 2004, Perry et al. 2005) have all shown returning to the same home scar between foraging ex-
responses. cursions (Hawkins et al. 1992), and their spatial distrib-
Canopy-forming fucoid algae are conspicuous mem- ution influences the probability of Fucus spp. (here after
bers of the community of all but the most exposed Fucus) becoming established via escapes from grazing
rocky shores of the NE Atlantic (Lewis 1964). Midshore (Hartnoll & Hawkins 1985). Conversely, Fucus
fucoids are a cold temperate/boreal group of species influences the distribution of many intertidal species in-
that becomes less abundant with decreasing latitude cluding limpets (Menge 1978, Hawkins et al. 1992,
(Ballantine 1961). Thus, it is anticipated that fucoids Thompson et al. 1996, Leonard 2000, Jenkins et al. 2005).
will become restricted to more northerly latitudes or Thus, whilst Patella spp. control the initial development
buffered environments if temperatures rise (Franklin & of fucoid stands, once established, fucoids provide
Forster 1997). For example, ‘climate envelope’ models ‘nursery’ grounds for juvenile P. vulgata and a microhab-
predict that, with an increase in temperature of 1 to itat for adult P. vulgata that aggregate under patches of
2°C, Fucus vesiculosis, the dominant midshore algal Fucus. In contrast, P. depressa does not appear to aggre-
species on moderately exposed shores, would be lost gate under Fucus patches (S. J. Hawkins unpubl. data).
from most of south-west Britain (M. T. Burrows unpubl. Here, we explore the direct and indirect effects of
data). There is some evidence to suggest that this Fucus canopy loss on these 2 closely related limpet
Moore et al.: Biotic interactions and climate change 13
species to demonstrate how responses to climate in both 2002 and 2003. At midshore level, 20 areas with
change can be modulated by biological interactions patches of Fucus of approximately 0.09 m2 in size were
mediated by intrinsic differences in behaviour. Such randomly allocated to 2 treatments: Fucus present
interactions are likely to alter predictions for species’ (unmanipulated control) or Fucus removal. In the
range shifts based on the direct effects of climate Fucus removal treatments, Fucus canopy (including
change on setting distributional limits. The relation- holdfast) was removed from the substrate with a
ships between Fucus patches and both Patella vulgata scalpel. Ten areas of open rock were also selected to
and P. depressa were examined to determine the act as controls. Responses of P. vulgata and P. depressa
extent to which each species aggregates beneath were monitored independently of each other in each of
Fucus. Manipulative field experiments were then used 5 replicate plots for each of the 3 treatments. Five P.
to investigate the responses of both limpet species to vulgata or P. depressa were randomly selected within
the loss of Fucus. Responses were quantified in terms each experimental plot; these individuals were mea-
of changes in behaviour — in this case, both the dis- sured and double tagged with micro-numbers glued to
tance moved to locate a new home scar and mortality. the shell with cyanoacrylate.
Specifically, we examined the hypothesis that the loss The response variables measured were distance
of Fucus would have a much greater effect on the moved to a new home scar and mortality. To quantify
behaviour and mortality of P. vulgata than on that of relocation to new home scars the position of each
P. depressa, by effectively reducing P. vulgatas’ pre- limpet was recorded using coordinates from a 1 m2 grid
ferred habitat. separated into 0.03 m2 grid squares. Limpet positions
prior to treatment manipulations were taken to be the
limpets’ initial home scars, as limpets are generally
MATERIALS AND METHODS inactive and on their homes scars at midshore levels
while the tide is out. Analyses were carried out on the
Spatial distribution of northern and southern distance limpets moved from their initial home scar at
limpet species in relation to Fucus patches. The spa- the completion of the experiment 4 mo later. The posi-
tial distribution of the limpets Patella depressa and P. tion of each individual was then determined every 14 d
vulgata in relation to Fucus were compared indepen- during low water when the limpets were not active (i.e.
dently of each other at each of 2 locations: Trevone on their home scars) to ensure that micro-numbers
(50° 55’ N, 4° 98’ W) and Crackington Haven (50° 74’ N, were still attached to limpet shells and to monitor mor-
4° 64’ W), on the north coast of Cornwall, UK. The tality. In cases in which limpets had relocated from
abundance of P. vulgata and P. depressa was recorded their original home scar, the habitat to which they had
in Fucus patches and on areas of open rock using 10 moved was noted. Where limpets were missing from
randomly placed 0.5 × 0.5 m quadrats for each plots, a 15 min search of the surrounding area (approx-
species–habitat combination. The effect of macroalgal imately 9 m2) was made, and, if the limpets were not
canopy on the relative abundance of P. vulgata and found, they were assumed to have died. Total mortality
P. depressa was then compared using a 3-factor across the 4 mo of the experiment was used for
ANOVA, with the factor location (2 levels) considered analysis.
random and the factors habitat (2 levels: with or with- In 2002, a large number of limpets from all treat-
out macroalgal canopy) and species (2 levels: P. vul- ments at Trevone died, probably as a consequence of
gata or P. depressa) considered fixed. natural sand scour in some of the experimental areas;
To provide an indication of the differences in tem- therefore, only the Crackington Haven experiment
perature limpets may experience beneath Fucus was analysed for this period. In 2003, comparisons
patches and on open rock, 3 temperature data loggers were possible between Trevone and Crackington
(Thermochron® ibutton DS1921G) were allocated to Haven. Two-factor ANOVA was used to compare dif-
each of 3 areas of open rock and beneath Fucus ferences in response variables (distance moved from
patches at Crackington Haven during typical summer original home scar at the completion of the experi-
low-tide weather conditions in 2004. ment and mortality over the course of the experiment)
Limpet mortality and behaviour following loss of between the 3 treatments at Crackington Haven in
Fucus. The effect of the loss of Fucus vesiculosis (here- 2002. The factors treatment (3 levels) and species (2
after Fucus) on the mortality and behaviour of Patella levels) were considered fixed. In 2003, a third factor,
vulgata and P. depressa was investigated experimen- location (2 levels and random), was also examined to
tally at Trevone and Crackington Haven. Both shores establish spatial consistency. The number of individ-
are moderately exposed, with areas of open rock, bar- ual limpets available for analysis of distance moved to
nacle-covered rock and a mosaic of Fucus patches. a new home scar was unequal, due to differential
Experiments were run between June and September mortality between treatments, so data were randomly
14 Mar Ecol Prog Ser 334: 11–19, 2007
removed from treatments to create a balanced design. Table 1. ANOVA for the abundance of Patella vulgata and
For all analyses, heterogeneity of variance was exam- P. depressa beneath Fucus patches and on emergent rock at
2 locations (Crackington Haven and Trevone) in south-west
ined using Cochran’s test, and, where appropriate,
Britain [ln (x + 1) transformation; Cochran’s test: C = 0.2008,
data were log(x + 1) transformed. In order to increase non-significant]. Location × habitat × species was non-
the power when comparing factors of interest, post- significant (p > 0.25) and was thus pooled with the residual to
hoc pooling was utilised to remove non-significant increase the power of the test for the treatment × species
terms (p > 0.25). Student-Newman-Keuls (SNK) post- interaction
hoc tests were carried out on significant results (p <
0.05). Source SS df MS F p F vs.
Location (Lo) 3.84 1 3.84 15.54 < 0.01 1-Pooled
Habitat (Ha) 9.70 1 9.70 99.57 0.06 Lo × Ha
RESULTS Species (Sp) 30.46 1 30.46 504.05 < 0.05 Lo × Sp
Lo × Ha 0.10 1 0.10 0.39 0.53 1-Pooled
Spatial distribution of northern and southern limpet Lo × Sp 0.06 1 0.06 0.24 0.62 1-Pooled
Ha × Sp 33.07 1 33.07 133.87 < 0.01 1-Pooled
species in relation to Fucus patches
Lo × Ha × Sp 0.04 1 0.04 0.15 0.70 1-Pooled
Residual 17.99 72 0.25
There were significant differences in the spatial dis- Total 95.25 79
tribution of Patella vulgata and P. depressa in relation 1-Pooled 73 0.25
to Fucus patches at both Trevone and Crackington SNK tests
Haven. The abundance of P. vulgata was significantly P. vulgata, Fucus patch > P. vulgata, open rock
P. depressa, open rock > P. depressa, Fucus patch
higher beneath Fucus patches than on open rock
(Table 1). In contrast, the abundance of P. depressa
was significantly higher on open rock than beneath
Fucus patches (F1,73 = 133.87; p < 0.01; Fig. 1). 0.05; Table 3, Fig. 2d). The maximum distance an
Temperature data loggers confirmed the potential individual P. vulgata moved to relocate their home
for Fucus patches to ameliorate conditions during scar following Fucus removal was approximately
low tide, with temperatures beneath Fucus patches 2 m. Of the P. vulgata that survived following Fucus
(mean ± SE: 20.59 ± 0.11°C) on average 5°C cooler removal, 36% from Crackington Haven in 2002 and
than temperatures experienced on open rock (25.68 ± 25% from Trevone and Crackington Haven in 2003
0.22°C). relocated home scars to beneath another Fucus
patch.
In contrast, Patella depressa at both locations in 2003
Limpet mortality and behaviour following loss stayed loyal to their home scars in all treatments, irre-
of Fucus spective of the presence or absence of Fucus. At
Crackington Haven in 2002, P. depressa moved a
Patella vulgata experienced significantly higher small, but significant, distance from their home scars in
mortality in Fucus removal treatments (approximately unmanipulated Fucus patch treatments (mean ± SE:
40%) than in unmanipulated Fucus patches (approxi- 12.2 ± 2.9 cm) compared to in Fucus removal (3.3 ±
mately 20%) and on open rock (approximately 24%).
There was no difference in the levels of mortality for P.
Limpet abundance (0.25 m-2)
30 ** Fucus patch
depressa amongst the 3 treatments. These patterns
Open rock
were evident at Crackington Haven in 2002 (F2,17 = 25
4.66, p < 0.05; Table 2, Fig. 2a) and were consistent at **
20
both Trevone and Crackington Haven in 2003 (F2,50 =
3.75, p < 0.05; Table 2, Fig. 2b). At Crackington Haven 15
in 2002, P. depressa suffered significantly higher levels
of mortality compared to P. vulgata in the open rock 10
treatments.
5
The distance moved to a new home scar by Patella
vulgata following experimental removal of Fucus was 0
greater than that of P. depressa in all treatments and P. vulgata P. depressa
than that of P. vulgata under Fucus patches and on
Fig. 1. Patella vulgata and P. depressa. Effect of macroalgal
open rock. This behaviour was consistent at Crack- cover (Fucus vesiculosus) on the relative abundance of
ington Haven in 2002 (F2,66 = 8.97, p < 0.01; Table 3, limpets at Crackington Haven and Trevone (means ± 1 SE)
Fig. 2c) and across locations in 2003 (F2,194 = 3.4, p < (**p < 0.01)
Moore et al.: Biotic interactions and climate change 15
1.2 cm) and open rock treatments (2 ± 0.7 cm; Fig. 2a). shore if the relative abundance of these 2 species alters
However, all individuals remained beneath the Fucus as is expected with increased climatic warming.
patch in which they were initially tagged. Experimental removal of Fucus led to two-thirds of
Patella vulgata individuals at Trevone and Cracking-
ton Haven either relocating their home scars to
DISCUSSION beneath other Fucus patches or dying. This may have
been caused by the greater thermal and desiccation
Responses of northern and southern species of stress experienced on open rock compared to beneath
limpet to fucoid microhabitats Fucus patches. Similar subtle changes in temperature
have been shown to influence the survival of other
We have shown that Patella vulgata aggregates intertidal species such as barnacles (Bertness et al.
beneath Fucus patches. In contrast P. depressa appears 1999, Leonard 2000).
to prefer open rock habitats. Our findings support pre- In contrast, the southern/lusitanian congeneric
vious work for P. vulgata (Hartnoll & Hawkins 1985, limpet species Patella depressa was more abundant on
Hawkins et al. 1992), but data on the patterns of distri- open rock, and the removal of Fucus had little effect on
bution of P. depressa in relation to Fucus patches have its behaviour or mortality. P. depressa did, however,
not been previously reported. These differences in
behaviour in relation to Fucus patches may lead to
Table 3. ANOVA for distance limpets moved to a new home
changes in the spatial distribution of grazers on the
scar for 3 treatments: Fucus removal (FR), Fucus patch (FP)
and open rock (OR) for Crackington Haven, south-west
Britain in 2002 [ln(x + 1) transformation; Cochran’s test: C =
Table 2. ANOVA of limpet mortality in 3 treatments: Fucus 0.2154, non-significant] (Due to unequal levels of mortality n
removal (FR), Fucus patch (FP) and open rock (OR) for Crack- = 12 limpets treatment–1 were used to formally analyse the
ington Haven, south-west Britain, in 2002 (Cochran’s test: distance moved from a home scar.) and Trevone and Crack-
C = 0.4943, non-significant), and Trevone and Crackington ington Haven, south-west Britain, in 2003 [ln(x + 1) transfor-
Haven, south-west Britain, in 2003 (Cochran’s test: C = mation; Cochran’s test: C = 0.1496, non-significant). Location
0.1932, non-significant). Location × treatment × species was × habitat × species was non-significant (p > 0.25) and was thus
non-significant (p > 0.25) and was thus pooled with the resid- pooled with the residual to increase the power of the test for
ual to increase the power of the test for the treatment × the treatment × species interaction. (Due to unequal levels of
species interaction mortality n = 17 limpets treatment–1)
Source SS df MS F p F vs. Source SS df MS F p F vs.
Crackington Haven (2002) Crackington Haven (2002)
Treatment (Tr) 0.96 2 0.48 2.72 0.12 Residual Treatment (Tr) 35.07 2 17.53 4.28 0.02 Residual
Species (Sp) 0.02 1 0.02 0.09 0.77 Residual Species (Sp) 1.19 1 1.19 0.29 0.59 Residual
Tr × Sp 1.64 2 0.82 4.66 0.03 Residual Tr × Sp 73.51 2 36.76 8.97 < 0.01 Residual
Residual 2.12 12 0.18 Residual 270.57 66 4.10
Total 4.74 17 Total 380.34 71
SNK tests SNK tests
Tr × Sp interaction Sp × Tr interaction Tr × Sp interaction Sp × Tr interaction
P. vulgata: FR > FP = OR FR: P. vulgata = P. depressa P. vulgata: FR > FP = OR FR: P. vulgata > P. depressa
P. depressa: FR = FP = OR FP: P. vulgata = P. depressa P. depressa: OR < FP; FP: P. vulgata < P. depressa
OR: P. vulgata < P. depressa FR < FP; OR = FR OR: P. vulgata = P. depressa
Trevore and Crackington Haven (2003) Trevone and Crackington Haven (2003)
Location (Lo) 1.67 1 1.67 2.30 0.13 1-Pooled Location (Lo) 9.45 1 9.45 8.68 0.00 1-Pooled
Treatment (Tr) 3.23 2 1.62 1.59 0.39 Lo × Tr Treatment (Tr) 4.33 2 2.17 0.89 0.53 Lo × Tr
Species (Sp) 3.27 1 3.27 49.00 0.09 Lo × Sp Species (Sp) 7.39 1 7.39 35.55 0.11 Lo × Sp
Lo × Tr 2.03 2 1.02 1.40 0.26 1-Pooled Lo × Tr 4.86 2 2.43 2.23 0.11 1-Pooled
Lo × Sp 0.07 1 0.07 0.09 0.77 1-Pooled Lo × Sp 0.21 1 0.21 0.19 0.66 1-Pooled
Tr × Sp 5.43 2 2.72 3.75 0.03 1-Pooled Tr × Sp 7.39 2 3.7 3.4 0.04 1-Pooled
Lo × Tr × Sp 1.03 2 0.52 0.71 0.50 1-Pooled Lo × Tr × Sp 0.73 2 0.37 0.34 0.72 1-Pooled
Residual 35.20 48 0.73 Residual 210.50 192 1.1
Total 51.93 59 Total 244.86 203
1-Pooled 36.23 50 0.72 1-Pooled 211.23 194 1.09
SNK tests SNK tests
Tr × Sp interaction Sp × Tr interaction Tr × Sp interaction Sp × Tr interaction
P. vulgata: FR > FP = OR FR: P. vulgata = P. depressa P. vulgata: FR > FP = OR FR: P. vulgata > P. depressa
P. depressa: FR = FP = OR FP: P. vulgata = P. depressa P. depressa: FR = FP = OR FP: P. vulgata = P. depressa
OR: P. vulgata < P. depressa OR: P. vulgata = P. depressa
16 Mar Ecol Prog Ser 334: 11–19, 2007
a b
c d
Fig. 2. Patella vulgata and P. depressa. Response to the experimental loss of Fucus vesiculosus for: treatment × species interaction
for mortality at (a) Crackington Haven in 2002 and (b) Trevone and Crackington Haven in 2003; treatment × species interaction
for distance moved to a new home scar at (c) Crackington Haven in 2002 and (d) Trevone and Crackington Haven in 2003
(means ± 1 SE) (*p < 0.05; **p < 0.01, ns: non-significant)
suffer increased levels of mortality on open rock com- depressa. The amelioration of temperature extremes
pared to P. vulgata, but this perhaps reflects greater beneath canopy algae may explain the preferential
intrinsic rates of mortality of P. depressa compared to P. aggregation of P. vulgata beneath Fucus patches and
vulgata (Boaventura et al. 2002). The reasons for the the relocation of home scars beneath new Fucus
difference in behaviour between P. vulgata and P. patches as well as the increased mortality when
depressa in relation to Fucus are not clear. Fucus is canopy algae is experimentally removed. Hence, the
absent over much of the biogeographic range of P. local persistence of P. vulgata may be partly reliant on
depressa, so the lack of a response by P. depressa to the presence of its preferred habitat, Fucus patches.
the loss of Fucus patches may be a direct result of It is unlikely that selective predation resulted in the
the limited overlap in their biogeographic ranges. increased mortality of Patella vulgata in Fucus removal
Alternatively, there is some evidence to suggest that treatments. The main predators of limpets are birds,
P. depressa does less well beneath Fucus patches crabs, whelks and fish (Thompson et al. 2000). There is
compared to on open rock (Southward & Southward no evidence of any selection for P. vulgata or P.
1978, Moore et al. in press), and may actively avoid depressa by these predators (e.g. Coleman et al. 1999
creating home scars beneath Fucus patches. for birds and Thompson et al. 2000 for crabs), or, in the
In laboratory studies both species of limpet have case of oystercatchers, a preference for foraging on
been shown to withstand and recover from exposure open or Fucus-covered rock (Coleman et al. 1999).
to temperatures close to 43°C (lethal temperatures:
Patella vulgata, 42.8°C and P. depressa, 43.3°C; Evans
1948); however, no work has been done on the temper- Biotic interactions modify species responses to
ature tolerances of these 2 species in the field. Recent climate change
work has shown that congeneric species with cold/
boreal biogeographic distributions were more temper- Amelioration of environmental conditions by organ-
ature sensitive than those with warmer/lusitanian dis- isms such as canopy-forming algae has been well
tributions (Stillman 2003). In addition species may be documented. However, the role of these habitat-
able to initially survive exposure to extreme tempera- generating species will become more important as
tures, but thermal damage experienced as a result of warming increases, since limpets, littorinids, whelks,
the exposure may prove fatal in the future (Stenseng et anemones and numerous other species will increas-
al. 2005). Therefore, although both species have been ingly become restricted to habitats where climatic ex-
shown to survive similar temperature extremes, P. vul- tremes are buffered (Leonard 2000, Jenkins et al. 2005).
gata may still be more temperature sensitive than P. For example, in the USA, habitat amelioration has been
Moore et al.: Biotic interactions and climate change 17
shown to increase the survival and persistence of the algal patches common on many moderately exposed
northern barnacle Semibalanus balanoides, whose shores of the NE Atlantic (Lewis 1964, Hartnoll &
survivorship was increased by shading by macroalgae Hawkins 1985, Hawkins et al. 1992), which, in turn,
at hotter southern sites, but not at cooler northern sites influence the community structure of rocky shores,
(Bertness et al. 1999, Leonard 2000). Furthermore, a because a higher diversity of organisms is found
study of high-level salt marsh plants in New England beneath Fucus patches compared to in areas of adja-
found that the plants experienced facilitative effects of cent emergent rock (Thompson et al. 1996). P. vulgata
neighbouring plants at hotter southern sites, but not at plays a key role in structuring rocky shore communi-
cooler northern sites (Bertness & Ewanchuk 2002). ties because of its aggregative behaviour, which can
Therefore, if warming continues as predicted, the local lead to a patchy distribution of grazing intensity that
persistence of many species may become more enables Fucus escapes to occur (Hartnoll & Hawkins
dependent on the presence of others, and shifts in 1985). Our data show that P. depressa does not prefer-
species’ distributions may become magnified where entially aggregate beneath Fucus patches, and this is
biologically generated habitats are lost or reduced. likely to result in a more even distribution of grazing
If climatic warming increases, the abundance of activity. Initial studies also indicate that P. depressa
Patella depressa at sites in Britain is predicted to in- may be less effective at controlling macroalgal abun-
crease, while the abundances of P. vulgata and Fucus dance than P. vulgata on shores in south-west Britain
are predicted to decrease. However, when the interac- (Moore et al. in press). Hence, changes in the spatial
tion of these 2 limpet species with Fucus is incorpo- distribution of grazers and the level of grazing inten-
rated, it is apparent that a stepped response in the rela- sity could have broad-scale implications for macroalgal
tive abundance of the 2 limpet species is likely to occur, abundance. Manipulation of canopyforming algae has
particularly at the edge of the distributional boundary been shown to result in changes in the diversity and
of Fucus. At this boundary the abundance of P. vulgata abundance of other algal species (Dayton 1975,
may decrease sharply as a consequence of increased Benedetti-Cecchi et al. 2001) and results in a reduction
environmental stress resulting from the loss of its pre- in the diversity and abundance of invertebrates
ferred micro-habitat (Fucus canopy). In contrast, the (Benedetti-Cecchi et al. 2001). Therefore, changes in
abundance of P. depressa may rapidly increase at the macroalgal cover are likely to have broad-scale impli-
boundary as more of its preferred habitat (open rock) cations for rocky shore community dynamics.
becomes available. This switch in the relative abun-
dance of P. vulgata and P. depressa is apparent when
the proportion of the 2 limpet species is compared be- Predicting responses to climate change
tween shores in south-west Britain (mean proportion of
P. depressa, 0.30; Jenkins et al. 2001), where Fucus can Our findings are supported by work carried out in
be abundant, and shores 7° (approximately 800 km) mesocosms (Davis et al. 1998) and by experimental
further south in northern Spain (mean proportion of P. field studies (Bertness et al. 1999, Leonard 2000),
depressa, 0.81; Jenkins et al. 2001), where Fucus be- which have highlighted the need to incorporate biotic
comes rare on open coasts (Ballantine 1961). If, as pre- interactions into predictions of species’ future biogeo-
dicted, canopy algae, such as Fucus vesiculosis, re- graphic distributions. The present study has shown, in
spond to increased warming faster than P. vulgata, the case of Patella vulgata and P. depressa, that the
without incorporation of the modifying (‘engineering’) presence or absence of Fucus canopy may alter the
effect of Fucus on local microhabitats, predictions of fu- speed at which these 2 species respond to increased
ture range shifts are likely to be inaccurate. For exam- climatic warming. Therefore, if future predictions on
ple, ‘climate envelope’ models based solely on changes the effects of climate change do not include effects on
in sea-surface temperature and wave action forecast lit- the survival of Fucus, then it is unlikely that realistic
tle change in the distribution of P. vulgata in the British predictions of range shifts by these limpet species can
Isles (M. T. Burrows unpubl. data). This is unlikely to be made. Recent work has also shown that variability
be the case if canopy-forming algae disappear, as in the physical environment is often greater within
expected, on shores in south-west Britain. sites than that experienced over larger geographic
areas (Benedetti-Cecchi et al. 2000); therefore, future
predictions of species’ range shifts need to be made at
Community-level responses to local changes in appropriate spatial (and temporal) scales. ‘Climate
grazer distribution envelope’ models may be able to provide a first
approximation of species’ responses to climate change;
The grazing behaviour of Patella vulgata contributes however, these models have the potential to be highly
to the conspicuous mosaic of variously aged macro- inaccurate if they do not take into account biotic
18 Mar Ecol Prog Ser 334: 11–19, 2007
interactions and are undertaken at an inappropriate range in response to global warming. Nature 391:783–786
scale (Hellmuth et al. 2006). ‘Climate envelope’ models Dayton PK (1975) Experimental evaluation of ecological dom-
inance in a rocky intertidal algal community. Ecol Monogr
incorporating multiple species, to account for biotic
45:137–159
interactions, and measurements of the physical envi- Evans RG (1948) The lethal temperatures of some common
ronment made at the appropriate scales are therefore British littoral molluscs. J Anim Ecol 17:165–173
fundamental in providing more biologically realistic Franklin LA, Forster RM (1997) The changing irradiance envi-
predictions of species’ range and abundance shifts in ronment: consequences for marine macrophyte physiol-
ogy, productivity and ecology. Eur J Phycol 32:207–232
response to climatic change. Genner MJ, Sims DW, Wearmouth VJ, Southall EJ, South-
ward AJ, Henderson PA, Hawkins SJ (2004) Regional
climatic warming drives long-term community changes of
Acknowledgements. P.M. was funded by a joint studentship British marine fish. Proc R Soc Lond B 271:655–661
from the University of Plymouth and the Marine Biological Hartnoll RG, Hawkins SJ (1985) Patchiness and fluctuations
Association of the UK. S.J.H. was supported by NERC via a on moderately exposed rocky shores. Ophelia 24:53–63
NERC grant-in-aid funded fellowship. S.J.H., R.C.T. and P.M. Hawkins SJ, Hartnoll RG, Kain JM, Norton TA (1992)
were also funded by the MarClim consortium (Countryside Plant –animal interactions on hard substrata in the north-
Council for Wales; The Crown Estates; Department for Envi- east Atlantic. In: John DM, Hawkins SJ, Price JH (eds)
ronment, Food and Rural Affairs; English Nature; Environ- Plant animal interactions in the marine benthos. Systemat-
ment Agency; Joint Nature Conservation Committee; Scottish ics Association Special, Vol 46, Clarendon Press, Oxford,
Executive; Scottish Natural Heritage; States of Jersey and p 1–32
Worldwide Fund for Nature). We appreciate the comments of Helmuth B, Mieszkowska N, Moore P, Hawkins SJ (2006)
4 reviewers, who greatly improved the manuscript. Living on the edge of two changing worlds: forecasting
the responses of rocky intertidal ecosystems to climate
change. Annu Rev Ecol Evol Syst 37:373–404
LITERATURE CITED Hodkinson ID (1999) Species response to global environmen-
tal change or why ecophysiological models are important:
Ballantine WJ (1961) A biologically-defined exposure scale a reply to Davis et al. J Anim Ecol 68:1259–1262
for the comparative description of rocky shores. Field Stud Hulme M, Jenkins GJ, Lu X, Turnpenny JR and 8 others
1:73–84 (2002) Climate change scenarios for the United Kingdom:
Beaugrand G, Ibanez F (2004) Monitoring marine plankton the UKCIP02 scientific report. Tyndall Centre for Climate
ecosystems. II. Long-term changes in North Sea calanoid Change Research, School of Environmental Sciences,
copepods in relation to hydro-climatic variability. Mar University of East Anglia, Norwich
Ecol Prog Ser 284:35–47 Jenkins SR, Coleman RA, Della Santina P, Hawkins SJ, Bur-
Benedetti-Cecchi L, Bulleri F, Cinelli F (2000) The interplay of rows MT, Hartnoll RG (2005) Regional scale differences in
physical and biological factors in maintaining mid-shore the determinism of grazing effects in the rocky intertidal.
and low-shore assemblages on rocky coasts in the north- Mar Ecol Prog Ser 287:77–86
west Mediterranean. Oecologia 123:406–417 Jenkins SR, Arenas F, Arrontes J, Bussell J and 10 others
Benedetti-Cecchi L, Pannacciulli FG, Bulleri F, Moschella PS, (2001) European-scale analysis of seasonal variability in
Airoldi L, Relini G, Cinelli F (2001) Predicting the conse- limpet grazing activity and microalgal abundance. Mar
quences of anthropogenic disturbance: large-scale effects Ecol Prog Ser 211:193–203
of loss of canopy algae on rocky shores. Mar Ecol Prog Ser Leonard GH (2000) Latitudinal variation in species inter-
214:137–150 actions: a test in the New England rocky intertidal zone.
Bertness MD, Leonard GH, Levine JM, Bruno JF (1999) Cli- Ecology 81:1015–1030
mate-driven interactions among rocky intertidal organ- Lewis JR (1964) The ecology of rocky shores. English Univer-
isms caught between a rock and a hot place. Oecologia sities Press, London
120:446–450 Menge BA (1978) Predation intensity in a rocky intertidal
Bertness MD, Ewanchuk PJ (2002) Latitudinal and climate- community: effect of an algal canopy, wave action and
driven variation in the strength and nature of biological desiccation on predator feeding rates. Oecologia 34:17–35
interactions in New England salt marshes. Oecologia 132: Moore P, Thompson RC, Hawkins SJ (in press) Effects of
392–401 grazer identity on the probability of a canopy-forming
Boaventura D, Fonseca LCD, Hawkins SJ (2002) Analysis of macroalga. J Exp Mar Biol Ecol
competitive interactions between the limpets Patella Parmesan C (1996) Climate and species range. Nature 382:
depressa Pennant and Patella vulgata L. on the northern 765–766
coast of Portugal. J Exp Mar Biol Ecol 271:171–188 Pearson RG, Dawson TP (2003) Predicting the impacts of cli-
Breeman AM (1990) Expected effects of changing sea-water mate change on the distribution of species: Are bioclimate
temperatures on the geographic distribution of seaweed envelope models useful? Global Ecol Biogeogr 12:361–371
species. In: Beukema JJ, Wolff WJ, Brouns JJWM (eds) Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate
Expected effects of climate change on marine coastal change and distribution shifts in marine fishes. Science
ecosystems. Kluwer Academic Publishers, Dordrecht, 308:1912–1915
p 69–76 Southward AJ (1967) Recent changes in abundance of inter-
Coleman RA, Goss-Custard JD, Durell SEAV, Hawkins SJ tidal barnacles in south-west England: a possible effect of
(1999) Limpet Patella spp. consumption by oystercatchers climatic deterioration. J Mar Biol Assoc UK 47:81–95
Haematopus ostralegus: a preference for solitary prey Southward AJ, Southward EC (1978) Recolonization of rocky
items. Mar Ecol Prog Ser 193:253–261 shores in Cornwall after use of toxic dispersants to clean
Davis AJ, Jenkinson LS, Lawton JH, Shorrocks B, Wood S up the Torrey Canyon spill. J Fish Res Board Can 35:
(1998) Making mistakes when predicting shifts in species 682–706
Moore et al.: Biotic interactions and climate change 19
Southward AJ, Hawkins SJ, Burrows MT (1995) Seventy (1996) Biologically generated habitat provision and diver-
years’ observations of changes in distribution and abun- sity of rocky shore organisms at a hierarchy of spatial
dance of zooplankton and intertidal organisms in the scales. J Exp Mar Biol Ecol 202:73–84
western English Channel in relation to rising sea temper- Thompson RC, Jenkins SR, Bussell JA (2000) A method for
ature. J Therm Biol 20:127–155 recording predator–prey encounters between crabs and
Stenseng E, Braby CE, Somero GN (2005) Evolutionary and limpets using wax replicas. J Mar Biol Assoc UK 80:633–638
acclimation-induced variation in the thermal limits of Thompson RC, Norton TA, Hawkins SJ (2004) Physical stress
heart function in congeneric marine snails (genus Tegula): and biological control regulate the balance between
implications for vertical zonation. Biol Bull (Woods Hole) producers and consumers in marine intertidal biofilms.
208:138–144 Ecology 85:1372–1382
Stillman JH (2003) Acclimation capacity underlies suscepti- Walther GR, Post E, Convey P, Menzel A and 5 others (2002)
bility to climate change. Science 301:65 Ecological responses to recent climate change. Nature
Thompson RC, Wilson BJ, Tobin ML, Hill AS, Hawkins SJ 416:389–395
Editorial responsibility: Howard Browman (Associate Editor- Submitted: March 17, 2006; Accepted: September 1, 2006
in-Chief), Storebø, Norway Proofs received from author(s): March 18, 2007
Vol. 334: 11–19, 2007 Published March 26
Mar Ecol Prog Ser
Role of biological habitat amelioration in altering
the relative responses of congeneric species to
climate change
P. Moore1, 2, 4,*, S. J. Hawkins1, 2, 3, R. C. Thompson2
1
Marine Biological Association of the UK, The Laboratory, Citadel Hill, Plymouth PL1 2PB, UK
2
Marine Biology and Ecology Research Centre, School of Biological Sciences, University of Plymouth,
Plymouth PL4 8AA, UK
3
School of Biological Sciences, University of Southampton, Southampton SO16 7PX, UK
4
Present address: Centre of Excellence for Coral Reef Studies, Centre for Marine Studies, University Queensland,
Brisbane, 4072 Queensland, Australia
ABSTRACT: The distribution of most species is expected to alter in response to climate change. Pre-
dictions for the extent of these range shifts are frequently based on ‘climate envelope’ approaches,
which often oversimplify species responses because many do not consider interactions between
physical and biological factors. The local persistence of some species, however, is likely to be strongly
modulated by microhabitat-forming organisms. Using congeneric patellid gastropods with northern/
boreal and southern/lusitanian distributions, we have demonstrated how the loss of habitat-forming
macroalgal species could modify species responses to climate change. The northern limpet Patella
vulgata preferentially aggregates beneath Fucus spp. When Fucus vesiculosus was experimentally
removed, to simulate a decline in macroalgal abundance in response to climatic warming, P. vulgata
suffered increased mortality or relocated home scars, often to nearby Fucus spp. patches. In contrast,
the southern limpet P. depressa did not aggregate beneath Fucus spp. and showed no response in
terms of movement or mortality to the loss of F. vesiculosus. Based on these results, we predict that
the loss of Fucus spp. will influence the relative abundance of these 2 limpet species, particularly at
the distributional limit of Fucus spp. In addition, differences in the aggregative behaviour of these
limpet species will result in changes in the spatial distribution of grazing in the intertidal, with likely
consequences for community dynamics. These outcomes could not be anticipated from predictions
based on direct responses to temperature alone, highlighting the need for biotic and abiotic factors to
be incorporated into predictions of species responses to climate change.
KEY WORDS: Biological interactions · Biologically generated habitat · Climate change · Climate
envelope · Limpets · Macroalgae
Resale or republication not permitted without written consent of the publisher
INTRODUCTION temperature, is mapped in ‘climate space’ (the area in
which a species can survive because the environment
Global air temperatures have risen by 0.6 ± 0.2°C in is suitable). If the climatic space alters, then it is pre-
the last 100 yr and further increases of between 2 and dicted that the geographical distribution of species
3°C are predicted by 2100 (Hulme et al. 2002), prompt- found within that climatic space will alter accordingly
ing the need to understand how species and assem- (Pearson & Dawson 2003). This approach can be useful
blages respond to climate change. To date, forecasts of (Hodkinson 1999, Pearson & Dawson 2003), but it has
species-level responses have used a ‘climate envelope’ also been criticised because it tends to predict a spe-
approach, whereby a single climatic variable, usually cies’ fundamental (or potential) niche and does not
*Email: pippa.moore@uq.edu.au © Inter-Research 2007 · www.int-res.com
12 Mar Ecol Prog Ser 334: 11–19, 2007
take into account biological interactions, dispersal northward range retraction has already started, with
capability, habitat quality, together with habitat avail- some cold-temperate canopy-forming species, such as
ability and connectivity, all of which influence a spe- Laminaria spp. (Breeman 1990, A. E. F. Murias Dos
cies’ realised niche (Davis et al. 1998, Pearson & Daw- Santos unpubl. data), Himanthalia elongata and Pel-
son 2003). Here, we use the rocky intertidal as a vetia canaliculata (A. E. F. Murias Dos Santos unpubl.
convenient model system to investigate the role of data), already becoming scarcer at their southern
behaviourally mediated direct and indirect biological range limits.
interactions in modifying species’ responses to climate Canopy-forming algae influence community struc-
change and the likely consequence for community ture by modifying the environment (Menge 1978).
structure and dynamics. Higher concentrations of microalgal food are found
The British Isles straddle 2 major marine biogeo- beneath algal canopies (Thompson et al. 2004), where
graphic zones, with both warm lusitanian and cool environmental conditions are ameliorated and habitats
boreal biota (Lewis 1964). As a consequence, many of are generated for other species (Leonard 2000), often
the species present are at either the northern or south- resulting in increased diversity (e.g. Fucus spp. at mid-
ern edge of their biogeographic ranges (Southward et shore heights in Britain, Thompson et al. 1996; and
al. 1995), resulting in sets of ecologically comparable Cystoseira spp. at low-shore levels in the Mediter-
species, with northern and southern centres of distrib- ranean, Benedetti-Cecchi et al. 2001). Hence, it is
ution co-existing and with abundances which have likely that the presence/absence of canopy-forming
been shown to fluctuate with climatic changes (South- algae will have a disproportionate effect on the distrib-
ward 1967, Southward et al. 1995). Therefore, south- ution and diversity of species in the intertidal, with
west Britain provides an ideal region in which to implications for assemblage dynamics.
investigate how species with northern and southern Limpets are key grazers on rocky shores in the NE
biogeographic distributions respond to climate change Atlantic, and strongly influence community structure
and, in particular, the way in which biotic interactions and dynamics by controlling macroalgal abundance
may modify these responses. (Southward & Southward 1978, Hawkins et al. 1992).
A general poleward movement of species’ ranges in Two species of limpet co-exist at midshore levels on
response to global warming has been predicted moderately exposed shores of south-west Britain:
(Parmesan 1996), although see Helmuth et al. (2006) Patella vulgata, a cold-temperate/boreal limpet, dis-
for discussion of local factors, such as time of low tributed from northern Norway to southern Portugal,
water, influencing species’ range shifts. Evidence of and P. depressa, a southern/lusitanian limpet, distrib-
changes in the relative abundance of species with uted from Senegal in West Africa to north Wales, UK
northern and southern biogeographic distributions, as (Southward et al. 1995). It is predicted that P. depressa
well as poleward and upward latitudinal shifts in spe- will become the dominant limpet on the shores of
cies’ ranges, have been observed in a wide range of south-west Britain, at the expense of P. vulgata, in
taxonomic groups during the twentieth century response to increased climatic warming (Southward et
(Walther et al. 2002). These shifts are evident in both al. 1995), and there is evidence to suggest this is
aquatic and terrestrial habitats, including coastal already occurring on some shores (S. J. Hawkins
waters where plankton (Beaugrand & Ibanez 2004), unpubl. data).
intertidal organisms (Southward et al. 1995) and fish Both species of limpet exhibit ‘homing’ behaviour,
(Genner et al. 2004, Perry et al. 2005) have all shown returning to the same home scar between foraging ex-
responses. cursions (Hawkins et al. 1992), and their spatial distrib-
Canopy-forming fucoid algae are conspicuous mem- ution influences the probability of Fucus spp. (here after
bers of the community of all but the most exposed Fucus) becoming established via escapes from grazing
rocky shores of the NE Atlantic (Lewis 1964). Midshore (Hartnoll & Hawkins 1985). Conversely, Fucus
fucoids are a cold temperate/boreal group of species influences the distribution of many intertidal species in-
that becomes less abundant with decreasing latitude cluding limpets (Menge 1978, Hawkins et al. 1992,
(Ballantine 1961). Thus, it is anticipated that fucoids Thompson et al. 1996, Leonard 2000, Jenkins et al. 2005).
will become restricted to more northerly latitudes or Thus, whilst Patella spp. control the initial development
buffered environments if temperatures rise (Franklin & of fucoid stands, once established, fucoids provide
Forster 1997). For example, ‘climate envelope’ models ‘nursery’ grounds for juvenile P. vulgata and a microhab-
predict that, with an increase in temperature of 1 to itat for adult P. vulgata that aggregate under patches of
2°C, Fucus vesiculosis, the dominant midshore algal Fucus. In contrast, P. depressa does not appear to aggre-
species on moderately exposed shores, would be lost gate under Fucus patches (S. J. Hawkins unpubl. data).
from most of south-west Britain (M. T. Burrows unpubl. Here, we explore the direct and indirect effects of
data). There is some evidence to suggest that this Fucus canopy loss on these 2 closely related limpet
Moore et al.: Biotic interactions and climate change 13
species to demonstrate how responses to climate in both 2002 and 2003. At midshore level, 20 areas with
change can be modulated by biological interactions patches of Fucus of approximately 0.09 m2 in size were
mediated by intrinsic differences in behaviour. Such randomly allocated to 2 treatments: Fucus present
interactions are likely to alter predictions for species’ (unmanipulated control) or Fucus removal. In the
range shifts based on the direct effects of climate Fucus removal treatments, Fucus canopy (including
change on setting distributional limits. The relation- holdfast) was removed from the substrate with a
ships between Fucus patches and both Patella vulgata scalpel. Ten areas of open rock were also selected to
and P. depressa were examined to determine the act as controls. Responses of P. vulgata and P. depressa
extent to which each species aggregates beneath were monitored independently of each other in each of
Fucus. Manipulative field experiments were then used 5 replicate plots for each of the 3 treatments. Five P.
to investigate the responses of both limpet species to vulgata or P. depressa were randomly selected within
the loss of Fucus. Responses were quantified in terms each experimental plot; these individuals were mea-
of changes in behaviour — in this case, both the dis- sured and double tagged with micro-numbers glued to
tance moved to locate a new home scar and mortality. the shell with cyanoacrylate.
Specifically, we examined the hypothesis that the loss The response variables measured were distance
of Fucus would have a much greater effect on the moved to a new home scar and mortality. To quantify
behaviour and mortality of P. vulgata than on that of relocation to new home scars the position of each
P. depressa, by effectively reducing P. vulgatas’ pre- limpet was recorded using coordinates from a 1 m2 grid
ferred habitat. separated into 0.03 m2 grid squares. Limpet positions
prior to treatment manipulations were taken to be the
limpets’ initial home scars, as limpets are generally
MATERIALS AND METHODS inactive and on their homes scars at midshore levels
while the tide is out. Analyses were carried out on the
Spatial distribution of northern and southern distance limpets moved from their initial home scar at
limpet species in relation to Fucus patches. The spa- the completion of the experiment 4 mo later. The posi-
tial distribution of the limpets Patella depressa and P. tion of each individual was then determined every 14 d
vulgata in relation to Fucus were compared indepen- during low water when the limpets were not active (i.e.
dently of each other at each of 2 locations: Trevone on their home scars) to ensure that micro-numbers
(50° 55’ N, 4° 98’ W) and Crackington Haven (50° 74’ N, were still attached to limpet shells and to monitor mor-
4° 64’ W), on the north coast of Cornwall, UK. The tality. In cases in which limpets had relocated from
abundance of P. vulgata and P. depressa was recorded their original home scar, the habitat to which they had
in Fucus patches and on areas of open rock using 10 moved was noted. Where limpets were missing from
randomly placed 0.5 × 0.5 m quadrats for each plots, a 15 min search of the surrounding area (approx-
species–habitat combination. The effect of macroalgal imately 9 m2) was made, and, if the limpets were not
canopy on the relative abundance of P. vulgata and found, they were assumed to have died. Total mortality
P. depressa was then compared using a 3-factor across the 4 mo of the experiment was used for
ANOVA, with the factor location (2 levels) considered analysis.
random and the factors habitat (2 levels: with or with- In 2002, a large number of limpets from all treat-
out macroalgal canopy) and species (2 levels: P. vul- ments at Trevone died, probably as a consequence of
gata or P. depressa) considered fixed. natural sand scour in some of the experimental areas;
To provide an indication of the differences in tem- therefore, only the Crackington Haven experiment
perature limpets may experience beneath Fucus was analysed for this period. In 2003, comparisons
patches and on open rock, 3 temperature data loggers were possible between Trevone and Crackington
(Thermochron® ibutton DS1921G) were allocated to Haven. Two-factor ANOVA was used to compare dif-
each of 3 areas of open rock and beneath Fucus ferences in response variables (distance moved from
patches at Crackington Haven during typical summer original home scar at the completion of the experi-
low-tide weather conditions in 2004. ment and mortality over the course of the experiment)
Limpet mortality and behaviour following loss of between the 3 treatments at Crackington Haven in
Fucus. The effect of the loss of Fucus vesiculosis (here- 2002. The factors treatment (3 levels) and species (2
after Fucus) on the mortality and behaviour of Patella levels) were considered fixed. In 2003, a third factor,
vulgata and P. depressa was investigated experimen- location (2 levels and random), was also examined to
tally at Trevone and Crackington Haven. Both shores establish spatial consistency. The number of individ-
are moderately exposed, with areas of open rock, bar- ual limpets available for analysis of distance moved to
nacle-covered rock and a mosaic of Fucus patches. a new home scar was unequal, due to differential
Experiments were run between June and September mortality between treatments, so data were randomly
14 Mar Ecol Prog Ser 334: 11–19, 2007
removed from treatments to create a balanced design. Table 1. ANOVA for the abundance of Patella vulgata and
For all analyses, heterogeneity of variance was exam- P. depressa beneath Fucus patches and on emergent rock at
2 locations (Crackington Haven and Trevone) in south-west
ined using Cochran’s test, and, where appropriate,
Britain [ln (x + 1) transformation; Cochran’s test: C = 0.2008,
data were log(x + 1) transformed. In order to increase non-significant]. Location × habitat × species was non-
the power when comparing factors of interest, post- significant (p > 0.25) and was thus pooled with the residual to
hoc pooling was utilised to remove non-significant increase the power of the test for the treatment × species
terms (p > 0.25). Student-Newman-Keuls (SNK) post- interaction
hoc tests were carried out on significant results (p <
0.05). Source SS df MS F p F vs.
Location (Lo) 3.84 1 3.84 15.54 < 0.01 1-Pooled
Habitat (Ha) 9.70 1 9.70 99.57 0.06 Lo × Ha
RESULTS Species (Sp) 30.46 1 30.46 504.05 < 0.05 Lo × Sp
Lo × Ha 0.10 1 0.10 0.39 0.53 1-Pooled
Spatial distribution of northern and southern limpet Lo × Sp 0.06 1 0.06 0.24 0.62 1-Pooled
Ha × Sp 33.07 1 33.07 133.87 < 0.01 1-Pooled
species in relation to Fucus patches
Lo × Ha × Sp 0.04 1 0.04 0.15 0.70 1-Pooled
Residual 17.99 72 0.25
There were significant differences in the spatial dis- Total 95.25 79
tribution of Patella vulgata and P. depressa in relation 1-Pooled 73 0.25
to Fucus patches at both Trevone and Crackington SNK tests
Haven. The abundance of P. vulgata was significantly P. vulgata, Fucus patch > P. vulgata, open rock
P. depressa, open rock > P. depressa, Fucus patch
higher beneath Fucus patches than on open rock
(Table 1). In contrast, the abundance of P. depressa
was significantly higher on open rock than beneath
Fucus patches (F1,73 = 133.87; p < 0.01; Fig. 1). 0.05; Table 3, Fig. 2d). The maximum distance an
Temperature data loggers confirmed the potential individual P. vulgata moved to relocate their home
for Fucus patches to ameliorate conditions during scar following Fucus removal was approximately
low tide, with temperatures beneath Fucus patches 2 m. Of the P. vulgata that survived following Fucus
(mean ± SE: 20.59 ± 0.11°C) on average 5°C cooler removal, 36% from Crackington Haven in 2002 and
than temperatures experienced on open rock (25.68 ± 25% from Trevone and Crackington Haven in 2003
0.22°C). relocated home scars to beneath another Fucus
patch.
In contrast, Patella depressa at both locations in 2003
Limpet mortality and behaviour following loss stayed loyal to their home scars in all treatments, irre-
of Fucus spective of the presence or absence of Fucus. At
Crackington Haven in 2002, P. depressa moved a
Patella vulgata experienced significantly higher small, but significant, distance from their home scars in
mortality in Fucus removal treatments (approximately unmanipulated Fucus patch treatments (mean ± SE:
40%) than in unmanipulated Fucus patches (approxi- 12.2 ± 2.9 cm) compared to in Fucus removal (3.3 ±
mately 20%) and on open rock (approximately 24%).
There was no difference in the levels of mortality for P.
Limpet abundance (0.25 m-2)
30 ** Fucus patch
depressa amongst the 3 treatments. These patterns
Open rock
were evident at Crackington Haven in 2002 (F2,17 = 25
4.66, p < 0.05; Table 2, Fig. 2a) and were consistent at **
20
both Trevone and Crackington Haven in 2003 (F2,50 =
3.75, p < 0.05; Table 2, Fig. 2b). At Crackington Haven 15
in 2002, P. depressa suffered significantly higher levels
of mortality compared to P. vulgata in the open rock 10
treatments.
5
The distance moved to a new home scar by Patella
vulgata following experimental removal of Fucus was 0
greater than that of P. depressa in all treatments and P. vulgata P. depressa
than that of P. vulgata under Fucus patches and on
Fig. 1. Patella vulgata and P. depressa. Effect of macroalgal
open rock. This behaviour was consistent at Crack- cover (Fucus vesiculosus) on the relative abundance of
ington Haven in 2002 (F2,66 = 8.97, p < 0.01; Table 3, limpets at Crackington Haven and Trevone (means ± 1 SE)
Fig. 2c) and across locations in 2003 (F2,194 = 3.4, p < (**p < 0.01)
Moore et al.: Biotic interactions and climate change 15
1.2 cm) and open rock treatments (2 ± 0.7 cm; Fig. 2a). shore if the relative abundance of these 2 species alters
However, all individuals remained beneath the Fucus as is expected with increased climatic warming.
patch in which they were initially tagged. Experimental removal of Fucus led to two-thirds of
Patella vulgata individuals at Trevone and Cracking-
ton Haven either relocating their home scars to
DISCUSSION beneath other Fucus patches or dying. This may have
been caused by the greater thermal and desiccation
Responses of northern and southern species of stress experienced on open rock compared to beneath
limpet to fucoid microhabitats Fucus patches. Similar subtle changes in temperature
have been shown to influence the survival of other
We have shown that Patella vulgata aggregates intertidal species such as barnacles (Bertness et al.
beneath Fucus patches. In contrast P. depressa appears 1999, Leonard 2000).
to prefer open rock habitats. Our findings support pre- In contrast, the southern/lusitanian congeneric
vious work for P. vulgata (Hartnoll & Hawkins 1985, limpet species Patella depressa was more abundant on
Hawkins et al. 1992), but data on the patterns of distri- open rock, and the removal of Fucus had little effect on
bution of P. depressa in relation to Fucus patches have its behaviour or mortality. P. depressa did, however,
not been previously reported. These differences in
behaviour in relation to Fucus patches may lead to
Table 3. ANOVA for distance limpets moved to a new home
changes in the spatial distribution of grazers on the
scar for 3 treatments: Fucus removal (FR), Fucus patch (FP)
and open rock (OR) for Crackington Haven, south-west
Britain in 2002 [ln(x + 1) transformation; Cochran’s test: C =
Table 2. ANOVA of limpet mortality in 3 treatments: Fucus 0.2154, non-significant] (Due to unequal levels of mortality n
removal (FR), Fucus patch (FP) and open rock (OR) for Crack- = 12 limpets treatment–1 were used to formally analyse the
ington Haven, south-west Britain, in 2002 (Cochran’s test: distance moved from a home scar.) and Trevone and Crack-
C = 0.4943, non-significant), and Trevone and Crackington ington Haven, south-west Britain, in 2003 [ln(x + 1) transfor-
Haven, south-west Britain, in 2003 (Cochran’s test: C = mation; Cochran’s test: C = 0.1496, non-significant). Location
0.1932, non-significant). Location × treatment × species was × habitat × species was non-significant (p > 0.25) and was thus
non-significant (p > 0.25) and was thus pooled with the resid- pooled with the residual to increase the power of the test for
ual to increase the power of the test for the treatment × the treatment × species interaction. (Due to unequal levels of
species interaction mortality n = 17 limpets treatment–1)
Source SS df MS F p F vs. Source SS df MS F p F vs.
Crackington Haven (2002) Crackington Haven (2002)
Treatment (Tr) 0.96 2 0.48 2.72 0.12 Residual Treatment (Tr) 35.07 2 17.53 4.28 0.02 Residual
Species (Sp) 0.02 1 0.02 0.09 0.77 Residual Species (Sp) 1.19 1 1.19 0.29 0.59 Residual
Tr × Sp 1.64 2 0.82 4.66 0.03 Residual Tr × Sp 73.51 2 36.76 8.97 < 0.01 Residual
Residual 2.12 12 0.18 Residual 270.57 66 4.10
Total 4.74 17 Total 380.34 71
SNK tests SNK tests
Tr × Sp interaction Sp × Tr interaction Tr × Sp interaction Sp × Tr interaction
P. vulgata: FR > FP = OR FR: P. vulgata = P. depressa P. vulgata: FR > FP = OR FR: P. vulgata > P. depressa
P. depressa: FR = FP = OR FP: P. vulgata = P. depressa P. depressa: OR < FP; FP: P. vulgata < P. depressa
OR: P. vulgata < P. depressa FR < FP; OR = FR OR: P. vulgata = P. depressa
Trevore and Crackington Haven (2003) Trevone and Crackington Haven (2003)
Location (Lo) 1.67 1 1.67 2.30 0.13 1-Pooled Location (Lo) 9.45 1 9.45 8.68 0.00 1-Pooled
Treatment (Tr) 3.23 2 1.62 1.59 0.39 Lo × Tr Treatment (Tr) 4.33 2 2.17 0.89 0.53 Lo × Tr
Species (Sp) 3.27 1 3.27 49.00 0.09 Lo × Sp Species (Sp) 7.39 1 7.39 35.55 0.11 Lo × Sp
Lo × Tr 2.03 2 1.02 1.40 0.26 1-Pooled Lo × Tr 4.86 2 2.43 2.23 0.11 1-Pooled
Lo × Sp 0.07 1 0.07 0.09 0.77 1-Pooled Lo × Sp 0.21 1 0.21 0.19 0.66 1-Pooled
Tr × Sp 5.43 2 2.72 3.75 0.03 1-Pooled Tr × Sp 7.39 2 3.7 3.4 0.04 1-Pooled
Lo × Tr × Sp 1.03 2 0.52 0.71 0.50 1-Pooled Lo × Tr × Sp 0.73 2 0.37 0.34 0.72 1-Pooled
Residual 35.20 48 0.73 Residual 210.50 192 1.1
Total 51.93 59 Total 244.86 203
1-Pooled 36.23 50 0.72 1-Pooled 211.23 194 1.09
SNK tests SNK tests
Tr × Sp interaction Sp × Tr interaction Tr × Sp interaction Sp × Tr interaction
P. vulgata: FR > FP = OR FR: P. vulgata = P. depressa P. vulgata: FR > FP = OR FR: P. vulgata > P. depressa
P. depressa: FR = FP = OR FP: P. vulgata = P. depressa P. depressa: FR = FP = OR FP: P. vulgata = P. depressa
OR: P. vulgata < P. depressa OR: P. vulgata = P. depressa
16 Mar Ecol Prog Ser 334: 11–19, 2007
a b
c d
Fig. 2. Patella vulgata and P. depressa. Response to the experimental loss of Fucus vesiculosus for: treatment × species interaction
for mortality at (a) Crackington Haven in 2002 and (b) Trevone and Crackington Haven in 2003; treatment × species interaction
for distance moved to a new home scar at (c) Crackington Haven in 2002 and (d) Trevone and Crackington Haven in 2003
(means ± 1 SE) (*p < 0.05; **p < 0.01, ns: non-significant)
suffer increased levels of mortality on open rock com- depressa. The amelioration of temperature extremes
pared to P. vulgata, but this perhaps reflects greater beneath canopy algae may explain the preferential
intrinsic rates of mortality of P. depressa compared to P. aggregation of P. vulgata beneath Fucus patches and
vulgata (Boaventura et al. 2002). The reasons for the the relocation of home scars beneath new Fucus
difference in behaviour between P. vulgata and P. patches as well as the increased mortality when
depressa in relation to Fucus are not clear. Fucus is canopy algae is experimentally removed. Hence, the
absent over much of the biogeographic range of P. local persistence of P. vulgata may be partly reliant on
depressa, so the lack of a response by P. depressa to the presence of its preferred habitat, Fucus patches.
the loss of Fucus patches may be a direct result of It is unlikely that selective predation resulted in the
the limited overlap in their biogeographic ranges. increased mortality of Patella vulgata in Fucus removal
Alternatively, there is some evidence to suggest that treatments. The main predators of limpets are birds,
P. depressa does less well beneath Fucus patches crabs, whelks and fish (Thompson et al. 2000). There is
compared to on open rock (Southward & Southward no evidence of any selection for P. vulgata or P.
1978, Moore et al. in press), and may actively avoid depressa by these predators (e.g. Coleman et al. 1999
creating home scars beneath Fucus patches. for birds and Thompson et al. 2000 for crabs), or, in the
In laboratory studies both species of limpet have case of oystercatchers, a preference for foraging on
been shown to withstand and recover from exposure open or Fucus-covered rock (Coleman et al. 1999).
to temperatures close to 43°C (lethal temperatures:
Patella vulgata, 42.8°C and P. depressa, 43.3°C; Evans
1948); however, no work has been done on the temper- Biotic interactions modify species responses to
ature tolerances of these 2 species in the field. Recent climate change
work has shown that congeneric species with cold/
boreal biogeographic distributions were more temper- Amelioration of environmental conditions by organ-
ature sensitive than those with warmer/lusitanian dis- isms such as canopy-forming algae has been well
tributions (Stillman 2003). In addition species may be documented. However, the role of these habitat-
able to initially survive exposure to extreme tempera- generating species will become more important as
tures, but thermal damage experienced as a result of warming increases, since limpets, littorinids, whelks,
the exposure may prove fatal in the future (Stenseng et anemones and numerous other species will increas-
al. 2005). Therefore, although both species have been ingly become restricted to habitats where climatic ex-
shown to survive similar temperature extremes, P. vul- tremes are buffered (Leonard 2000, Jenkins et al. 2005).
gata may still be more temperature sensitive than P. For example, in the USA, habitat amelioration has been
Moore et al.: Biotic interactions and climate change 17
shown to increase the survival and persistence of the algal patches common on many moderately exposed
northern barnacle Semibalanus balanoides, whose shores of the NE Atlantic (Lewis 1964, Hartnoll &
survivorship was increased by shading by macroalgae Hawkins 1985, Hawkins et al. 1992), which, in turn,
at hotter southern sites, but not at cooler northern sites influence the community structure of rocky shores,
(Bertness et al. 1999, Leonard 2000). Furthermore, a because a higher diversity of organisms is found
study of high-level salt marsh plants in New England beneath Fucus patches compared to in areas of adja-
found that the plants experienced facilitative effects of cent emergent rock (Thompson et al. 1996). P. vulgata
neighbouring plants at hotter southern sites, but not at plays a key role in structuring rocky shore communi-
cooler northern sites (Bertness & Ewanchuk 2002). ties because of its aggregative behaviour, which can
Therefore, if warming continues as predicted, the local lead to a patchy distribution of grazing intensity that
persistence of many species may become more enables Fucus escapes to occur (Hartnoll & Hawkins
dependent on the presence of others, and shifts in 1985). Our data show that P. depressa does not prefer-
species’ distributions may become magnified where entially aggregate beneath Fucus patches, and this is
biologically generated habitats are lost or reduced. likely to result in a more even distribution of grazing
If climatic warming increases, the abundance of activity. Initial studies also indicate that P. depressa
Patella depressa at sites in Britain is predicted to in- may be less effective at controlling macroalgal abun-
crease, while the abundances of P. vulgata and Fucus dance than P. vulgata on shores in south-west Britain
are predicted to decrease. However, when the interac- (Moore et al. in press). Hence, changes in the spatial
tion of these 2 limpet species with Fucus is incorpo- distribution of grazers and the level of grazing inten-
rated, it is apparent that a stepped response in the rela- sity could have broad-scale implications for macroalgal
tive abundance of the 2 limpet species is likely to occur, abundance. Manipulation of canopyforming algae has
particularly at the edge of the distributional boundary been shown to result in changes in the diversity and
of Fucus. At this boundary the abundance of P. vulgata abundance of other algal species (Dayton 1975,
may decrease sharply as a consequence of increased Benedetti-Cecchi et al. 2001) and results in a reduction
environmental stress resulting from the loss of its pre- in the diversity and abundance of invertebrates
ferred micro-habitat (Fucus canopy). In contrast, the (Benedetti-Cecchi et al. 2001). Therefore, changes in
abundance of P. depressa may rapidly increase at the macroalgal cover are likely to have broad-scale impli-
boundary as more of its preferred habitat (open rock) cations for rocky shore community dynamics.
becomes available. This switch in the relative abun-
dance of P. vulgata and P. depressa is apparent when
the proportion of the 2 limpet species is compared be- Predicting responses to climate change
tween shores in south-west Britain (mean proportion of
P. depressa, 0.30; Jenkins et al. 2001), where Fucus can Our findings are supported by work carried out in
be abundant, and shores 7° (approximately 800 km) mesocosms (Davis et al. 1998) and by experimental
further south in northern Spain (mean proportion of P. field studies (Bertness et al. 1999, Leonard 2000),
depressa, 0.81; Jenkins et al. 2001), where Fucus be- which have highlighted the need to incorporate biotic
comes rare on open coasts (Ballantine 1961). If, as pre- interactions into predictions of species’ future biogeo-
dicted, canopy algae, such as Fucus vesiculosis, re- graphic distributions. The present study has shown, in
spond to increased warming faster than P. vulgata, the case of Patella vulgata and P. depressa, that the
without incorporation of the modifying (‘engineering’) presence or absence of Fucus canopy may alter the
effect of Fucus on local microhabitats, predictions of fu- speed at which these 2 species respond to increased
ture range shifts are likely to be inaccurate. For exam- climatic warming. Therefore, if future predictions on
ple, ‘climate envelope’ models based solely on changes the effects of climate change do not include effects on
in sea-surface temperature and wave action forecast lit- the survival of Fucus, then it is unlikely that realistic
tle change in the distribution of P. vulgata in the British predictions of range shifts by these limpet species can
Isles (M. T. Burrows unpubl. data). This is unlikely to be made. Recent work has also shown that variability
be the case if canopy-forming algae disappear, as in the physical environment is often greater within
expected, on shores in south-west Britain. sites than that experienced over larger geographic
areas (Benedetti-Cecchi et al. 2000); therefore, future
predictions of species’ range shifts need to be made at
Community-level responses to local changes in appropriate spatial (and temporal) scales. ‘Climate
grazer distribution envelope’ models may be able to provide a first
approximation of species’ responses to climate change;
The grazing behaviour of Patella vulgata contributes however, these models have the potential to be highly
to the conspicuous mosaic of variously aged macro- inaccurate if they do not take into account biotic
18 Mar Ecol Prog Ser 334: 11–19, 2007
interactions and are undertaken at an inappropriate range in response to global warming. Nature 391:783–786
scale (Hellmuth et al. 2006). ‘Climate envelope’ models Dayton PK (1975) Experimental evaluation of ecological dom-
inance in a rocky intertidal algal community. Ecol Monogr
incorporating multiple species, to account for biotic
45:137–159
interactions, and measurements of the physical envi- Evans RG (1948) The lethal temperatures of some common
ronment made at the appropriate scales are therefore British littoral molluscs. J Anim Ecol 17:165–173
fundamental in providing more biologically realistic Franklin LA, Forster RM (1997) The changing irradiance envi-
predictions of species’ range and abundance shifts in ronment: consequences for marine macrophyte physiol-
ogy, productivity and ecology. Eur J Phycol 32:207–232
response to climatic change. Genner MJ, Sims DW, Wearmouth VJ, Southall EJ, South-
ward AJ, Henderson PA, Hawkins SJ (2004) Regional
climatic warming drives long-term community changes of
Acknowledgements. P.M. was funded by a joint studentship British marine fish. Proc R Soc Lond B 271:655–661
from the University of Plymouth and the Marine Biological Hartnoll RG, Hawkins SJ (1985) Patchiness and fluctuations
Association of the UK. S.J.H. was supported by NERC via a on moderately exposed rocky shores. Ophelia 24:53–63
NERC grant-in-aid funded fellowship. S.J.H., R.C.T. and P.M. Hawkins SJ, Hartnoll RG, Kain JM, Norton TA (1992)
were also funded by the MarClim consortium (Countryside Plant –animal interactions on hard substrata in the north-
Council for Wales; The Crown Estates; Department for Envi- east Atlantic. In: John DM, Hawkins SJ, Price JH (eds)
ronment, Food and Rural Affairs; English Nature; Environ- Plant animal interactions in the marine benthos. Systemat-
ment Agency; Joint Nature Conservation Committee; Scottish ics Association Special, Vol 46, Clarendon Press, Oxford,
Executive; Scottish Natural Heritage; States of Jersey and p 1–32
Worldwide Fund for Nature). We appreciate the comments of Helmuth B, Mieszkowska N, Moore P, Hawkins SJ (2006)
4 reviewers, who greatly improved the manuscript. Living on the edge of two changing worlds: forecasting
the responses of rocky intertidal ecosystems to climate
change. Annu Rev Ecol Evol Syst 37:373–404
LITERATURE CITED Hodkinson ID (1999) Species response to global environmen-
tal change or why ecophysiological models are important:
Ballantine WJ (1961) A biologically-defined exposure scale a reply to Davis et al. J Anim Ecol 68:1259–1262
for the comparative description of rocky shores. Field Stud Hulme M, Jenkins GJ, Lu X, Turnpenny JR and 8 others
1:73–84 (2002) Climate change scenarios for the United Kingdom:
Beaugrand G, Ibanez F (2004) Monitoring marine plankton the UKCIP02 scientific report. Tyndall Centre for Climate
ecosystems. II. Long-term changes in North Sea calanoid Change Research, School of Environmental Sciences,
copepods in relation to hydro-climatic variability. Mar University of East Anglia, Norwich
Ecol Prog Ser 284:35–47 Jenkins SR, Coleman RA, Della Santina P, Hawkins SJ, Bur-
Benedetti-Cecchi L, Bulleri F, Cinelli F (2000) The interplay of rows MT, Hartnoll RG (2005) Regional scale differences in
physical and biological factors in maintaining mid-shore the determinism of grazing effects in the rocky intertidal.
and low-shore assemblages on rocky coasts in the north- Mar Ecol Prog Ser 287:77–86
west Mediterranean. Oecologia 123:406–417 Jenkins SR, Arenas F, Arrontes J, Bussell J and 10 others
Benedetti-Cecchi L, Pannacciulli FG, Bulleri F, Moschella PS, (2001) European-scale analysis of seasonal variability in
Airoldi L, Relini G, Cinelli F (2001) Predicting the conse- limpet grazing activity and microalgal abundance. Mar
quences of anthropogenic disturbance: large-scale effects Ecol Prog Ser 211:193–203
of loss of canopy algae on rocky shores. Mar Ecol Prog Ser Leonard GH (2000) Latitudinal variation in species inter-
214:137–150 actions: a test in the New England rocky intertidal zone.
Bertness MD, Leonard GH, Levine JM, Bruno JF (1999) Cli- Ecology 81:1015–1030
mate-driven interactions among rocky intertidal organ- Lewis JR (1964) The ecology of rocky shores. English Univer-
isms caught between a rock and a hot place. Oecologia sities Press, London
120:446–450 Menge BA (1978) Predation intensity in a rocky intertidal
Bertness MD, Ewanchuk PJ (2002) Latitudinal and climate- community: effect of an algal canopy, wave action and
driven variation in the strength and nature of biological desiccation on predator feeding rates. Oecologia 34:17–35
interactions in New England salt marshes. Oecologia 132: Moore P, Thompson RC, Hawkins SJ (in press) Effects of
392–401 grazer identity on the probability of a canopy-forming
Boaventura D, Fonseca LCD, Hawkins SJ (2002) Analysis of macroalga. J Exp Mar Biol Ecol
competitive interactions between the limpets Patella Parmesan C (1996) Climate and species range. Nature 382:
depressa Pennant and Patella vulgata L. on the northern 765–766
coast of Portugal. J Exp Mar Biol Ecol 271:171–188 Pearson RG, Dawson TP (2003) Predicting the impacts of cli-
Breeman AM (1990) Expected effects of changing sea-water mate change on the distribution of species: Are bioclimate
temperatures on the geographic distribution of seaweed envelope models useful? Global Ecol Biogeogr 12:361–371
species. In: Beukema JJ, Wolff WJ, Brouns JJWM (eds) Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate
Expected effects of climate change on marine coastal change and distribution shifts in marine fishes. Science
ecosystems. Kluwer Academic Publishers, Dordrecht, 308:1912–1915
p 69–76 Southward AJ (1967) Recent changes in abundance of inter-
Coleman RA, Goss-Custard JD, Durell SEAV, Hawkins SJ tidal barnacles in south-west England: a possible effect of
(1999) Limpet Patella spp. consumption by oystercatchers climatic deterioration. J Mar Biol Assoc UK 47:81–95
Haematopus ostralegus: a preference for solitary prey Southward AJ, Southward EC (1978) Recolonization of rocky
items. Mar Ecol Prog Ser 193:253–261 shores in Cornwall after use of toxic dispersants to clean
Davis AJ, Jenkinson LS, Lawton JH, Shorrocks B, Wood S up the Torrey Canyon spill. J Fish Res Board Can 35:
(1998) Making mistakes when predicting shifts in species 682–706
Moore et al.: Biotic interactions and climate change 19
Southward AJ, Hawkins SJ, Burrows MT (1995) Seventy (1996) Biologically generated habitat provision and diver-
years’ observations of changes in distribution and abun- sity of rocky shore organisms at a hierarchy of spatial
dance of zooplankton and intertidal organisms in the scales. J Exp Mar Biol Ecol 202:73–84
western English Channel in relation to rising sea temper- Thompson RC, Jenkins SR, Bussell JA (2000) A method for
ature. J Therm Biol 20:127–155 recording predator–prey encounters between crabs and
Stenseng E, Braby CE, Somero GN (2005) Evolutionary and limpets using wax replicas. J Mar Biol Assoc UK 80:633–638
acclimation-induced variation in the thermal limits of Thompson RC, Norton TA, Hawkins SJ (2004) Physical stress
heart function in congeneric marine snails (genus Tegula): and biological control regulate the balance between
implications for vertical zonation. Biol Bull (Woods Hole) producers and consumers in marine intertidal biofilms.
208:138–144 Ecology 85:1372–1382
Stillman JH (2003) Acclimation capacity underlies suscepti- Walther GR, Post E, Convey P, Menzel A and 5 others (2002)
bility to climate change. Science 301:65 Ecological responses to recent climate change. Nature
Thompson RC, Wilson BJ, Tobin ML, Hill AS, Hawkins SJ 416:389–395
Editorial responsibility: Howard Browman (Associate Editor- Submitted: March 17, 2006; Accepted: September 1, 2006
in-Chief), Storebø, Norway Proofs received from author(s): March 18, 2007