ii''-'GM` I'4 to Oll ii

. 011 MCA The role of biological interactions in modifying the effects of climate

change on intertidal assemblages

by

Philippa Jane Moore

A thesis submitted to the University of Plymouth in partial fulfilment for the degree of

Doctor of Philosophy

School of Biological Sciences Faculty of Science

In collaboration with the Marine Biological Association of the UK

July 2005 This copy of the thesis has been supplied on condition that anyone who consults it is understood to recognisethat its copyrightrests with its author and that no quotationfrom the thesis and no informationderived from it may be publishedwithout the author's prior consent. The remains of an experiment following flash floods at Crackington Haven on the 16th of August 2004, which resulted in cars and half the carpark ending up on the shore. An example of more extreme weather events with anthropogenic mediated climate change? Philippa Jane Moore

The role of biological interactions in modifying the effects of climate change on intertidal assemblages

Abstract

The geographic distribution of most species is expected to alter as a consequence of global climate change. Predictions for the extent of these range shifts are frequently based on anticipated changes in temperature using a 'climate envelope' approach, which oversimplifies predictions because it does not consider interactions with other physical and biological factors. The aim of this thesis was to investigate how biological interactions modulate species responses to climate change.

On many rocky shores in the NE Atlantic the interaction between limpets, barnacles and canopy forming macroalgae have an important role in structuring rocky shore communities. In particular, limpets control the abundance of macroalgae on the shore through their grazing activities. Through descriptive studies and manipulative experiments the behaviour of a northern/boreal species of limpet, Patella vulgata and a southern/lusitanian species of limpet, P. depressa were compared in relation to canopy forming algae (Fucus patches). As a result of differences in the spatial distribution, behaviour and grazing activity of these two species, if as predicted, there are changes in their relative abundance it is likely there will be implications for rocky shore community dynamics.

The second part of my thesis investigatedintra- and interspecificcompetition between two co- existing barnacle taxa with northern and southern centres of distribution. It is predicted that Increased warming will result in a reduction in the abundance of Semibalanus balanoides either as a direct result of increased temperatures or due to an increase in the number of poor spawning years. My results suggest that as a consequence of the gregarious nature of settling S. balanoides cyprids, recruitment success may be reduced irrespective of the numbers of cyprids in the plankton. This will result in more space becoming available for the competing and later settling Chthamalus spp, resulting in a change in barnacle population structure.

The likely impactsof populationchanges and species range shifts in responseto increased warming are discussed, with particular emphasis on how the interaction between limpets, barnacles and Fucus may alter. The implications of altered species interactions are then discussed in terms of the effects on community dynamics and ecosystem functioning. Finally the role of biotic interactions in modulating species responses to climate change are discussed with reference to the use of the 'climate envelope' approach in making predictions of species range shifts. List of Contents Abstract List of figures v List of tables vii Acknowledgements Ix Authors declaration x

Chapter One General Introduction

1.1 Introduction 1 1.2 Anthropogenic climate change 2 1.3 The biological consequences of climate change 3 1.4 Evidence of species responses to climatic fluctuations 5 1.5 Rationale of thesis 9

Chapter Two The association between two limpet species and patches of fucoid macroalgae

2.1 Introduction 12 2.2 Materials and methods 16 2.2.1 The distribution of Patella depressa in relation to Fucus patches 16 2.2.2 The spatial distribution and population structure of P. vulgata and 16 P. depressa in relation to Fucus patches 2.3 Results 18 2.3.1 The distribution of Patella depressa in relation to Fucus patches 18 2.3.2 The spatial distribution and population structure of P. vulgata and 19 P. depressa in relation to Fucus patches 2.4 Discussion 27

Chapter Three Behaviour of a northern and southern species of limpet in relation to Fucus patches

3.1 Introduction 30 3.2 Materials and methods 33 3.2.1 Behaviour of a northern and southern species of limpet in relation 33 to Fucus patches 3.2.2 Limpet behaviour following the experimental loss of Fucus 34 3.3 Results 36 3.3.1 Behaviour of a northern and southern species of limpet in relation 36 to Fucus patches 3.3.2 Limpet behaviour following loss of Fucus patches 38 3.4 Discussion 43 3.4.1 Behaviour of a northern and southern species of limpet in relation 43 to Fucus patches 3.4.2 Biotic interactions modify species responses to climate change 44 3.4.3 Community level responses to local changes in grazer distribution 45

Chapter Four The effects of temperature and gonad ripening on the grazing activity of P. vulgata and P. depressa

4.1 Introduction 48 4.2 Materials and methods 51 4.3 Results 54 4.3.1 Grazing activity 54 4.3.2 Fucus growth 56 4.3.3 Reproductive cycles 57 4.3.4 Limpet growth 58 iii 4.3.5 Temperature, Grazing and Reproduction 58 4.4 Discussion 66 4.4.1 Comparisons of the grazing activity of two limpet species with 66 different biogeographicranges 4.4.2 Phenology and climate change 67 4.4.3 Impacts of differences in limpet grazing activity for community 79 structure

Chapter Five The intra- and Interspecific Interactions during settlement and recruitment in a northern and southern species of barnacle

5.1 Introduction 72 5.1.1 The effect of adult abundance of the barnacle, Semibalanus 75 balanoides, on the recruitment success of its gregarious settling young 5.1.2 The effects of poor 0+ Semibalanus recruitment on the subsequent 76 recruitment success of Chthamalus spat 5.2 Materials and Methods 77 5.2.1 The effect of adult abundance of the barnacle, Semibalanus 77 balanoides, on the recruitment success of its gregarious settling young 5.2.2 The effects of poor 0+ Semibalanus recruitment on the subsequent 79 recruitment success of Chthamalusspat 5.3 Results 80 5.3.1 The effect of adult abundance of the barnacle, Semibalanus 80 balanoides, on the recruitmentsuccess of its gregarious settling young 5.3.2 The effects of poor 0+ Semibalanusrecruitment on the subsequent 84 recruitment success of Chthamalusspat 5.4 Discussion 86 5.4.1 Limitations of work 86 5.4.2 Effects of barnacle settlement factor on Semibalanusrecruitment 87 5.4.3 The effect of adult abundance of the barnacle, Semibalanus 87 balanoides, on the recruitment success of its gregarious settling young 5.4.4 The effects of poor 0+ Semibalanusrecruitment on the subsequent 89 recruitment success of Chthamalus spat 5.4.5 The consequences of climate change on the inter- and intraspecific 90 interactions of a northern and southern species of barnacle

Chapter Six General Discussion 6.1 Thesis overview and summary 92 6.2 Consequences of increased climatic warming on rocky shore 95 community structure and dynamics 6.3 Consequences of changes in species abundances and ranges 100 for rocky shore ecosystem functioning 6.4 Changes In the variance of climatic events Increases the 101 difficulty in predicting species responses to climate change 6.5 `Climate envelope' approach vs biotic Interactions 102

References 104 Appendix 1 Limpet abundance in treatment enclosures 125 Appendix 2 Settlement factor preparation 126

iv List of figures

Chapter Two

Figure 2.1 Location of study sites in south and southwest Britain. Figure 2.2 Proportion of P. depressa to P. vulgata found beneath Fucus patches and on open rock at five locations around the coast of south and southwest Britain. Error bars ±I SE. ** P<0.01. Figure 2.3 The effect of macroalgal cover on the relative abundance of P. vulgata and P. depressa. This effect was consistent at both Crackington Haven and Trevone therefore data from both locations have been combined. Error bars ±1 SE. ** P<0.01. Figure 2.4 The proportion of P. depressa to P. vulgata with increasing cover of Fucus at a) Newquay and b) Trevone. Figure 2.5 The distance limpets were found from their nearest conspecific neighbour at Trevone. Figure 2.6 Cumulative size frequencies of P. vulgata inside and outside Fucus patches at five locations around the south and southwest coast of Britain. Open rock - bold line, Fucus patches - dotted line. Figure 2.7 Cumulative size frequency distribution of P. depressa on open rock at five locations on the south and southwest coast of Britain. Chapter Three

Figure3.1 Locationof study sites on the north coast of Cornwall,UK. Figure 3.2 Mean number of limpets relocating their home scars to beneath uncolonised Fucus patches. As results were consistent at both Crackington Haven and Trevone the data from both locations have been combined. Error bars ±1 SE. " P<0.01. Figure 3.3 Mean distance moved from home scar for P. depressa and P. vulgata for three treatments: Fucus removal (FR), Fucus patch (FR) and Open rock (OR). Error bars ±1 SE. a) Means for Treatment x species interaction at Crackington Haven for 2002 and 2003 as pattern was consistent among years. b) Means for Treatment x species interaction at Trevone and Crackington Haven in 2003 as pattern was consistent among sites. ** P<0.01; ns=not significant. Figure 3.4 Limpet mortality in three treatments: Fucus removal (FR), Fucus patch (FP) and Open rock (OR). a) Means for Treatment x species interaction at Crackington Haven for 2002 and 2003 as pattern was consistent among years. b) Means for Treatment x species interaction at Trevone and Crackington Haven in 2003 as pattern was consistent among sites. ** P<0.01; ns = not significant. Figure 3.5 Changes in the abundance of a) two species of limpet (P. vulgata a northern/boreal species and P. depressa a southern/lusitanian species) and b) Fucus vesiculosus with increasing temperature c) the interactive effect of Fucus vesiculosus and temperature on the abundance of P. vu!gata and P. depressa.

Chapter Four

Figure4.1 Locationof study sites in southwestBritain. Figure 4.2 Mean percentage area of wax discs scraped for three treatments: P. vulgata only, P. depressa only and both species present for a) Kingsand and b) Andurn Point. Error bars ±1 SE. Arrow indicates the time when Fucus cover reached 40% in P. depressa only treatments at Kingsand. Figure 4.3 Mean percentage area of wax disc scraped in three treatments; P. vulgata only; P. depressa only and both species together for Andurn Point and Kingsand before Fucus cover reached 40% in P. depressa only enclosures at Kingsand and after Fucus cover reached 40% cover at Kingsand. Error bars ± 1SE. * P<0.05, ** P<0.01, ns = not significant. Figure 4.4 Fucus growth in three treatments: a) P. depressa only, b) P. vulgata only, c) both species together at Kingsand and Andurn Point. Pictures taken 15 months after initiation of experiment. Wax discs placed in pre-drilled holes to quantify limpet grazing activity are also visible in the treatment enclosures.

V Figure 4.5 Mean percentage cover of Fucus across the whole shore (Sh) and in three treatments: P. depressa only (Pd), P. vulgata only (Pv) and both species present (Bo) for Kingsand and Andurn Point. Error bars ± 1SE. ** P<0.01, ns = not significant. Figure4.6 The proportionof P. vulgataand P. depressain advancedgonad states (gonadstages 4 &5 after Orton et al. 1956)for a) Kingsandand b) Andurn Point. Figure 4.7 Ford-Walford plots of growth increment plotted against initial shell length for P. depressa and P. vulgata from three treatments: P. depressa only; P. vulgata only and both species together. P. vulgata from treatments with both species present (Both Pv); P. vulgata only treatments (Pv); P. depressa from treatments with both species present (Both Pd) and P. depressa only treatments (Pd). ANCOVA indicates no significant difference between the slopes of the regressions for each species- treatment plot (F 3,3430.78; P>0.05). Chapter Five

Figure 5.1 Location of barnacle study sites in southwest Britain. Figure 5.2 The abundance of Semibalanus to recruit to settlement plates (with barnacle settlement factor and without barnacle settlement factor) for three densities of adult con-specifics attached: high density (5cm 2), low density (1cm 2) and no adult conspecifcs for a) Trevone: results were consistent for treatments with barnacle settlement factor and those without, therefore data has been combined b) Looe: results were consistent for treatments with barnacle settlement factor and those without, therefore data has been combined c) Newquay. Error bars ± ISE. ** P<0.01, ns = not significant. Figure 5.3 Chthamalus spat recruitment to four treatments: bare rock, control, high density (5cm'Z) and low density (1cm 2) for a) Cellar Beach and Looe and b) CrackingtonHaven. Error bars ±1 SE. ** P<0.01, ns = not significant.

Chapter Six Figure 6.1 Interactions between a) P. vulgata and Fucus vesiculosus and b) P. depressa and Fucus vesiculosus. - negative interaction; + positive interaction. The width of arrows represents the strength of the interaction. Dotted lines are indirect interactions, while full lines are direct interactions between species. Figure6.2 Interactionsbetween Limpets, barnacles and Fucus on a) shores dominatedby Semibalanusand P. vulgataand b) shoreswith P. vulgataand P. depressaand dominatedby Chthamalus. - negativeinteraction; + positiveinteraction. The width of arrows representsthe strength of the interaction. Dotted lines are indirect interactions, while full lines are direct interactions between species.

Appendix One Figure 1 Limpet abundancein treatmentenclosures at Andurn Point and Kingsand. ** P<0.01

vi List of tables

Chapter two

Table 2.1 Analysis of variance of the proportion of P. depressa to P. vulgata found beneath Fucus patches and on open rock at five locations around the south and southwest coasts of Britain. Variances were homogenous following arcsine transformation (Cochran's test 0.36; ns). Table 2.2 Analysis of variance of the abundance of P. vulgata and P. depressa beneath Fucus and on open rock at Crackington Haven and Trevone. Variances were homogenous following Ln (X + 1) transformation (Cochrans 0.2008, ns). The location x habitat x species interaction was non-significant (P >0.25) and was thus pooled with the residual to increase the power of the test for the treatment x species interaction.

Chapter three

Table 3.1 Analysis of variance of the movement of P. vulgata and P. depressa into uncolonised Fucus patches at Crackington Haven and Trevone. Variances were homogenous following fourth root transformation (Cochran's test C=0.4504, ns). Location x species interaction was non-significant (P >0.25) and was thus pooled with the residual to increase the power of the test for the factor species. Table 3.2 Analysis of variance of distance that limpets were found from their home scar for three treatments: Fucus removal (FR), Fucus patch (FP) and open rock (OR) at Crackington Haven in 2002 and 2003. Variances were homogenous following Ln (X + 1) transformation (Cochran's test C=0.1141, ns). Due to unequal levels of mortality n=12 limpets were used per treatment. Table 3.3 Analysis of variance of distance that limpets were found from their home scar for three treatments: Fucus removal (FR),. Fucus patch (FP) and open rock (OR) at Trevone and Crackington Haven in 2003. Variances were homogenous following Ln (X + 1) transformation (Cochran's test C=0.1496, ns). Location x habitat x species interaction was non-significant (P >0.25) and was pooled with the residual to increase the power of the test for treatment x species interaction. Due to unequal levels of mortality n=1 7 limpets per treatment. Table 3.4 Analysis of variance of limpet mortality in three treatments: Fucus removal (FR), Fucus patch (FP) and open rock (OR) at Crackington Haven in 2002 and 2003. Cochran's test C=0.1948, ns. Year xtreatment x species interaction was non-significant (P>0.25) and was pooled with the residual to increase the power of the test for the treatment x species interaction. Table 3.5 Analysis of variance of limpet mortality in three treatments: Fucus removal (FR), Fucus patch (FP) and open rock (OR) at Trevone and Crackington Haven in 2003. Cochran's test C=0.1932, ns. Location x treatment x species interaction was non-significant (P>0.25) and was pooled with the residual to increase the power of the test for the treatment x species interaction.

Chapter Four

Table 4.1 Analysis of variance of limpet grazing activity at Andurn Point and Kingsand, southwest Britain for three treatments: P. vulgata only, P. depressa only and both species present (Cochran's test C=0.0734, ns). Significant results are indicated > or <. The direction of non-significant trends are also shown - or t- to help illustrate overall patterns in the data. Table 4.2 Analysis of variance of percentage Fucus cover for three treatments: P. vulgata only, P. depressa only and both species present sixteen months after the initiation of the experiment. Variances were homogenous following fourth root transformation (Cochran's test C=0.4369, ns). Table 4.3 Gonad index of limpets from three treatments: P. depressa only, P. vulgata only and both species present as well as individuals collected randomly across the shore at the conclusion of the experiment in March 2005 from Kingsand and Andurn Point, southwest Britain. Table 4.4 Pearson's correlations between limpet grazing activity in three treatment enclosures: P. vulgata only, P. depressa only and both species present with temperature. To control for Type I error rates resulting from multiple tests, significance levels were corrected using the sequential Bonferroni procedure. Spearmans rank correlation was performed on the proportion of the limpet population in advanced states of gonad development and grazing activity. Data from the two locations vii have been combined for P. vulgafa only and both species together treatments as patterns of grazing activity were similar. * P<0.05.

Chapter Five

Table 5.1 Analysis of variance of the number of Semibalanus to recruit to settlement plates (with barnacle settlement factor and without barnacle settlement factor) for three densities of attached adult con-specifics; high density (5cm2), low density (1cm 2) and no adult conspecifics for a) Trevone (Cochran's test C=0.361, ns) b) Looe (Cochran's test C= 0.372, ns) and c) Newquay (Cochran's test C= 0.556, ns). Variances were homogenous for all locations following Ln (x + 1) transformation. -º _ direction of non-significant trend. Table 5.2 Analysis of variance for the number of Chthamalus to recruit to four treatments; cleared 2) rock, control, high density 0+ Semibalanus (approx 5cm2), low density 0+ Semibalanus (1cm for a) Cellar Beach and Looe (Variances were homogenous following square root (x + 1) transformation; Cochran's test C= 0.3545, ns) and b) Crackington Haven (Cochran's test C= 0.5029, ns).

Appendix One

Table I Analysis of variancefor the abundanceof limpetsfound in three treatments:P. vulgata only, P. depressaonly and both species togetherfor Kingsandand Andurn Point. Varianceswere homogenousfollowing square root (x + 1) transformation(Cochran's test C= 0.4662, ns).

VIII Acknowledgments

First and foremost I have to thank Harry Coover who was inducted into the National Inventors

Hall of Fame in 2004 for his invention 'Superglue' (cyanoacrylate). If it hadn't been for Harry I would not have been able to complete this thesis as I think it was used in every experiment I conducted.

Knew I should have bought shares in 2002.

More importantly I would like to thank my supervisors Prof Steve Hawkins and Dr Richard

Thompson for their help and support during the last three and a half years. All at the MBA and

MBERC have offered great support, in particular, I would like to thank Rebecca Leaper and Nova

Mieszkowska from the MarClim team and the rest in the Ecology and biodiversity group. I would also like to thank Ali Board, Jackie Hill, Neil Hutchinson, Rebecca Leaper and Molly Jacobs who all had the dubious 'pleasure' of proof reading a chapter of this thesis. A number of people helped me in the field over the course of my work and your help was invaluable.

Thanks also to Alice Coupland, if I hadn't meet you on the 'clippity cloppity' path I would never have reached the position where I am writing these acknowledgements. Thanks also to 'The Nowhere' and all those who can be found within her, you have provided friendship, debate and a chance to escape 'real' life for a little while. Special thanks to Ali Board who has helped pick me up, entertain me and encourage me, cheers matel

Last but not least I would like to thank limpet number 13 who proved the numberreally is unlucky by getting stuck to my arm with superglueand dying. r

ix Author's Declaration

At no time during the registrationfor the degreeof Doctor of Philosophyhas the author been registered for any other Universityaward without prior agreementof the GraduateCommittee.

This study was financed with the aid of a joint studentship from the Plymouth Environmental Research Centre, University of Plymouth and the Marine Biological Association of the UK. Additional funding was provided by the MarClim consortium (Countryside Council for Wales; The Crown Estates; Department for Environment, Food and Rural Affairs; English Nature; Environment Agency; Joint Nature Conservation Committee; Scottish Executive; Scottish Natural Heritage; States of Jersey and Worldwide Fund for Nature) and a National Environmental Research Council CASE studentship.

Relevant scientific seminars and conferences were regularly attended at which work was often presented; external institutions were visited for consultation purposes and papers prepared for publication.

Publications:

Sims, D.W., Wearmouth, V. J., Southall, E.J., Hill, J., Moore, P., Rawlinson, K., Hutchinson, N., Budd, G. C., Metcalfe, J. D. and Morritt, D. (in press). Hunt warm, rest cool: Bioenergetic efficiency underlying diel vertical migration of a benthic shark. Journal of Animal Ecology

Moore, P., Hawkins, S. J., Thompson, R.C. (in review). The role of biological habitat amelioration in altering the relative responses of congeneric species to climate change. Journal of Animal Ecology

Presentationsand ConferencesAttended:

Visiting lecture at Friday Harbor Marine Laboratories, University of Washington, WA, USA (29thApril 2005). 'Interactions between limpets and canopy forming macroalgae in a warming climate'. (21-23 July 'Climate change and aquatic systems past, present and future', University of Plymouth, UK 2004). Paper presentation 'Ecological complexity places climate envelope in the shade'.

'Postgraduate marine biology workshop', University of Wales, Bangor, UK (2-4 April 2004). Paper presentation 'Ecological complexity places climate envelope in the shade'.

'Citadel Hill seminar series', Marine Biological Association of the UK (March 2003). Paper presentation 'The role of Fucus patches in influencing the distribution of a northern and southern species of limpet'.

'Temperate reefs symposium 2003', Canterbury, New Zealand (13-17 January 2003). Poster presentation 'Climate change and species interactions: the role of Fucus patch in influencing the distribution of a northern and southern species of limpet'.

Word count of main body of thesis: 33 761

fflll X...... d..

Date.. q& 9 ý:.: % ......

X Chapter One

Chapter One

General Introduction

1.1 Introduction

There is now little debate that anthropogenic forcing of the Earth's climate is occurring

(Houghton et al. 1996, Oreskes 2004). Global air temperatures have risen by 0.6°C ± 0.2°C in the

last one hundred years (Albritton et al. 2001) and further increases of between 2°C to 3°C are

predicted by 2100 (Hulme et al. 2002). For the management and conservation of natural habitats it

is therefore important to understand how individuals, populations and assemblages will respond to

changes in their environment (Chapin III et al. 2000, Hughes 2000). It is generally thought that

species will respond to climatic warming by shifting their ranges poleward (e. g. Graham and Grimm

1990, Fields et al. 1993, Southward et al. 1995, Jeffree and Jeffree 1996, Parmesan 1996, Sagarin

et al. 1999, Hiscock et al. 2004), but this will depend on the dispersal capability of species and the

availability of suitable habitats (Parmesan 1996, Walther et al. 2002, Zacherl et al. 2003, Hiscock et

al. 2004). Changes in community structure and ecosystem functioning are also likely to occur

(Brown et al. 1997, Harrington et al. 1999, Chapin III et al. 2000, McCarty 2001), as organisms will respond to climate change in different ways and on different temporal and spatial scales depending on their physiology, behaviour and ecology (Graham and Grimm 1990, Harrington et al. 1999).

Utilising the rocky intertidal as a test system, my thesis examines the role of biotic interactions in modifying the outcome of species responses to climate change. Using limpets, barnacles and Fucus as test organisms, observational studies and manipulative experiments have been used to show how species interactions may modulate the strength and direction of interactions (i. e. competition or facilitation) during a period of climatic warming. In this introductory chapter, the history, predictions and causes of anthropogenic climate change are briefly summarised. Examples from terrestrial and marine realms have been used to demonstrate the responses of species to both natural and anthropogenically forced climate change. Finally the role in rocky shores and particular the shores around southwest Britain can play in assessing and predicting species responses to climate change is explained before presenting the overall aim and specific objectives of the chapters comprising this thesis. Chapter One

1.2 Anthropogenic climate change

Climate is influenced by many factors working at a variety of temporal scales including multi decadal and decadal patterns such as changes In the inclination of the moons orbit and changes in solar flux measured by sun spot activity (Stevens and North 1996). On shorter time scales there can be changes in weather systems reflected by the North Atlantic Oscillation Index

(NAOI) (Hurrell and VanLoon 1997, Visbeck et al. 2001) and El Nino Southern Oscillation (ENSO)

(Trenberth and Hurrell 1994). Since the late 1980s, the climate in northwest Europe seems to have become decoupled from these natural cycles, and many climatologists have linked this to anthropogenic forcing of the climate via increased levels of greenhouse gases (Stouffer et at. 1994,

Meehl et at. 2005), such as, carbon dioxide, methane, nitrous oxide and fluorocarbons (Hulme and

Jenkins 1998, Albritton et al. 2001, Hulme et al. 2002).

In the last one hundred years, average global air temperatures have risen by 0.6°C

±0.02°C, with a corresponding increase in surface seawater temperatures of 0.6°C (Albritton et at.

2001). Air temperatures monitored for Central England since 1659 show a 0.7°C rise in temperature since the seventeenth century; 0.5°C occurring during the twentieth century, with warming being greatest in winter months (Hulme and Jenkins 1998, Hulme et al. 2002). Mean surface seawater temperaturesfor the western English Channel (50-51° N, 1-2° W) have also risen by 0.7°C in the period 1861-1997 (Hawkins et al. 2003). It would appear that globally there has been less of a rise in maximum temperatures; instead minimum temperatures seem to have risen resulting in a narrowing of temperature ranges (Easterling et al. 1997).

Fears of the consequencesof climatic warming, particularly the economic costs, have motivated governments to fund research to predict future climate change. Global circulation models (GCMs), using varying rates of greenhouse gas emissions, have been used to try and predict climate change over long and short time scales. Although individual GCMs use different parameters and assumptions in their predictions, all show an increase in global average sea and temperatures'(1.4 air to 5.8°C by the year 2100), sea levels and extreme weather events such as flooding. Precipitation rates, storm frequency and evapotranspiration are also predicted to change (Hulme Jenkins and 1998, Albritton et al. 2001, Hulme et al. 2002).

Within Britain, the United Kingdom Climate Impacts Programme (UKCIP) has created models using natural climate variations as well as a range of future greenhouse gas emissionsto

2 Chapter One

try to predict the climate over the next one hundred years (Hulme and Jenkins 1998, Albritton et at.

2001, Hulme et at. 2002). Using the HadCM3 model, four climate scenarios relating to low,

medium-low, medium-high and high levels of global warming based on predicted future levels of

anthropogenic forcing of the climate have been derived. Natural forcing factors such as volcanoes,

solar variability as well as sulphur dioxide concentrations released as aerosols, which may have a

modest cooling effect, have not been considered. The climate change scenarios have been

calculated for the 2020s, 2050s and 2080s. The GCM scenarios for Britain predict an average

annual air temperature increase of between 2 to 3.5°C by the year 2080 (Hulme and Jenkins

1998), although there are variations between regions with average temperature increases of

between 1.2 to 4.5°C predicted for southwest Britain (Southwest Climate Change Impacts

Partnership [SWCCIP], 2003).

Predictions for changes in precipitation show large regional variations (Hulme and Jenkins

1998). For southwest Britain predictions show seasonal differences, with summers becoming 15 to

30% drier by the year 2050 and winters becoming 5 to 15% wetter during the same period

(SWCCIP, 2003). It is also likely that there will be an increase in the intensity of precipitation

events resulting in increased chances of flooding (Easterling et al. 2000). Sea levels are also

predicted to rise through the expansion of warmer ocean water with a global average sea level rise of between 7 and 36cm by 2050 (Hulme and Jenkins 1998). In southwest Britain storm frequencies are predicted to increase in number during winter months, with deeper depressions occurring (SWCCIP, 2003).

1.3 The biological consequences of climate change

Species distributions in both terrestrial and aquatic realms are influenced by abiotic and biotic factors (Begon et al. 1996 for review). On a broad-scale, physical factors, such as temperature, play a key role in limiting the geographic distribution of many species (Hutchins 1947,

Andrewartha and Birch 1954, Breeman 1988, Bolton and Anderson 1990, Clarke 1996, Pearson and Dawson 2003). At smaller spatial scales, a variety of other factors including topography, land use and biotic interactions explain the realised niche that species occupy (Pearson and Dawson 2003). In intertidal environments it has been widely accepted that species upper limits are set by factors physical linked to emersion stresses such as temperature and desiccation (Connell 1972), whilst lower limits are set by biotic interactions such as competition (Connell 1961 b) and predation Chapter One

(Paine 1966). However, evidence has since been accumulatedto suggest that species upper limits can also be set by biological interactions (Underwoodand Jernakoff 1981, Hawkins and Hartnoll

1985, Bertness and Leonard 1997, Boaventuraet al. 2002). The influence of latitudinal and temporal changes in climate are superimposedabove these local gradients.

As a consequence of climate change, organisms are predicted to follow one of four responses:

1. They may move elsewhere, with the generally agreed theory that this will be a poleward

and/or upward altitudinal movement (Graham and Grimm 1990, Fields et al. 1993,

Southward et al. 1995, Clarke 1996, Parmesan 1996, Sagarin et al. 1999, Hiscock et al.

2004). Evidence already suggests that this has occurred in a wide range of taxonomic

groups during the twentieth century (Easterling et al. 2000, Hughes 2000, McCarty 2001,

Walther et al. 2002, Parmesan and Yohe 2003), including taxa from oceanic (Roemmich

and McGowan 1995, Beaugrand et al. 2002) and coastal waters (Southward et al. 1995,

Holbrook et al. 1997, Sagarin et al. 1999). However, these range shifts may not be linear

with habitat microclimate ameliorating or exacerbating environmental conditions resulting in

refuge populations being maintained in certain areas (Helmuth et al. 2002). In the case of

intertidal organisms time of low water is likely to have an important effect on how species

respond to increased temperatures (Helmuth 1998, Helmuth and Hofmann 2001, Helmuth

et al. 2002). For example, in locations such as southwest Britain, where low water springs

occur during midday (1100 to 1300), the period of maximum solar insolation, species are

likely to be more affected by increased temperatures than organisms at locations where

spring tides usually occur at times where solar insolation is reduced.

2. Species may cope with the effects of climate change by tolerating the altered

environmentalconditions (Clarke 1996) through physiological and/or behavioural

mechanisms.

3. Species may evolve genetically to suboptimal environmentalconditions (Rodriguez-Trelles

and Rodriguez 1998), although evolutionary change or adaptation will be dependent on the

species generation time, which will be slower for longer lived species (Jeffree and Jeffree 1996) such as many intertidal organisms. As a result, it has been argued that migration than rather adaptation is the usual response of species to temperature change (Huntley 1991).

4 Chapter One

4. Species unable to move or adapt to changes in climatic conditions may become locally

extinct (Fields et al. 1993, Clarke 1996).

Species coexistence is dependent on a host of physical and biologicalfactors, resulting in species showing individualistic responses to climate change (Graham and Grimm 1990, Fields et al. 1993). As a consequence, new species associations are likely to form, leading to new assemblages and changes in community dynamics and ecosystem functioning (Brown et al. 1997,

Chapin III et al. 2000, McCarty 2001). In terrestrial systems the connectivity of suitable habitat patches will influence species ability to respond to climatic warming. In more open marine systems species such as planktonic organisms (Roemmich and McGowan 1995, Beaugrand et al. 2002) and fish (Genner et al. 2004) may be able to respond to climate change relatively quickly. Many intertidal organisms have a planktonic larval stage and therefore may also respond to climate change quite quickly. Range expansion or retraction of intertidal species will, however, still be dependent on the availability of suitable habitat (Walther et al. 2002, Hiscock et al. 2004), which can also be affected by anthropogenic modification of coastal habitats. For example, some species of intertidal epibiota have expanded their range along the English Channel coast with species settling on artificial structures such as sea defences and piers (Herbert et al. 2003, Hawkins, unpub). Anthropogenic habitat modification, such as habitat loss, may also act as a barrier to species being able to respond to climate change (Chapin III et al. 2000).

For many marine species, it is believed that increased water temperaturewill be the dominant factor affecting changes in species distribution and abundance, particularlyduring the next 50 to 100 years (Hiscock et al. 2001). In addition to increased sea temperature, intertidal organisms will be exposed to increasedtemperature when the tide is out. As air temperatureon land will increase more rapidly than surface seawater temperature (Hulme et al. 2002), increases in air temperature could affect intertidal organisms more than changes in seawater temperature.

1.4 Evidence of species responses to climatic fluctuations

During the 1950s, marine scientists became increasingly aware of fluctuations in the distribution and switches in the relative abundance of some pelagic and intertidal species with northern and southern biogeographic distributions (Southward and Crisp 1956, Cushing 1961,

Southward 1967, Russell et al. 1971, Cushing and Dickson 1976, Colebrook 1979). It was that suggested changes in climatic conditions, particularly temperature, resulted in the alteration in 5 Chapter One

the relative abundance of these species (Southward and Crisp 1956, Cushing 1961, Southward

1967, Russell et al. 1971, Cushing and Dickson 1976, Colebrook 1979). During the early 1990s, when awareness of anthropogenic forcing of the climate became apparent, descriptive analyses of the likely response of organisms to rapid climate change, based on knowledge of the species

physiology and ecology, started to appear in the literature (Graham and Grimm 1990, Fields et at.

1993). At the same time, evidence of changes in species abundance, distributions (Southward

1991, Barry et al. 1995, Roemmich and McGowan 1995, Southward et al. 1995, Parmesan 1996,

Holbrook et al. 1997) and phenology (Beebee 1995, Crick et at. 1997, Crick and Sparks 1999) as a result of anthropogenic climate change also started to emerge. As a result, monitoring programmes that had been stopped were restarted and historic sites were resampled in order to compare present day patterns with previous data (Southward et al. 1995, Parmesan 1996, Sagarin et al. 1999). Since the mid 1990s, attempts have been made to predict species range shifts using mathematical models (Jeffree and Jeffree 1994, Jeffree and Jeffree 1996, Pearson and Dawson

2003). More recently, experimental studies have investigated the role that biotic interactions

(Brown et al. 1997, Davies et al. 1998a, Petchey et al. 1999) and habitat microclimate (Helmuth and Hofmann 2001, Helmuth et al. 2002) may play in modulating species responses to climate change.

The pioneering work of Southward and Crisp in the 1950s and 1960s linked climate with the distribution and abundance of several marine organisms (Southward and Crisp 1954a,

Southward and Crisp 1956, Crisp and Southward 1958, Southward 1963,1967). Much of their work concentrated on the fluctuations in the distribution and relative abundance of two barnacle species: Semibalanus balanoides (previously Balanus balanoides) and Chthamalus stellatus (now separated into two species: C. stellatus and C. montagui Southward 1976). During warm periods, numbers of the northern/boreal species, S. balanoides, declined and in many cases became restricted to refugia where environmental conditions were favourable (Southward and Crisp 1954a,

Southward 1967,1991, Southward et al. 1995). At the same time the southern/lusitanian

Chthamalus spp. increased in abundance and was able to colonise areas on the shore where it had been previously out-competed by S. balanoides (Connell 1961 b, Southward 1967,1991,

Southward et al. 1995). In the 1960s, when cooler climatic conditions began to dominate, the relative abundance of these species switched (Southward 1967, Southward 1980, Southward et al.

2005). This in switch the relative abundance of these two genera in response to changed climatic 6 Chapter One

conditions has also been observed in species found on the NE coast of the United States (Wethey

1983, Wethey 1984). In addition to barnacles and other intertidal species, fish and planktonic

organisms have also been shown to alter their distributions and relative abundance with changes in

climate (Southward and Crisp 1954, Southward and Crisp 1956, Crisp and Southward 1958,

Southward 1963,1967, Russell et al. 1971, Southward 1980, Southward et al. 1995, Beaugrand et

al. 2002, Genner et al. 2004, Southward et al. 2005).

The Marine BiologicalAssociation of the United Kingdom (MBA) is the custodian of a

number of these long-term data sets and recently many of these long-term studies have been

restarted (Hawkins et al. 2003, Genner et al. 2004, Hiscock et al. 2004, Kendall et al. 2004, Sims et

al. 2004, Southward et al. 2005). Analysis of these data sets have shown correlations between changing climatic conditions and fluctuations in the abundance and range of boreal and lusitanian species of planktonic organisms, demersal fish and intertidal organisms (Southward et al. 1995,

Hawkins et al. 2003, Genner et al. 2004, Hiscock et al. 2004, Kendall et al. 2004, Sims et al. 2004,

Southward et al. 2005).

The studies outlined above have been useful in demonstratingthat species ranges and abundances have changed in response to climate conditions. However, they do not enable accurate prediction of actual range changes that are likely to occur as a consequence of anthropogenic forcing of the climate. Therefore, in order to anticipate species range shifts, mathematical models have been developed (Scott and Poynter 1991, Jeffree and Jeffree 1994,

Jeffree and Jeffree 1996, Pearson and Dawson 2003). Often termed the 'climate envelope' approach, the main technique used, maps a single climatic variable, usually temperature, in

'climate space' (the area in which a species can survive because the environment is suitable). If the climatic space alters then it is predicted that the geographical distribution of species found within that climatic space will alter accordingly (Jeffree and Jeffree 1996, Davies et al. 1998b,

Pearson and Dawson 2003, Huntley et al. 2004, Thomas et al. 2004). When these approaches have been tested against empirical data they have been shown to be broadly accurate, particularly large at spatial scales (Pearson and Dawson 2003). However the 'climate envelope' approach to tends predict species fundamental (or potential niche) because factors such as biotic interactions, which influence a species realised niche, are not included (Pearson and Dawson 2003).

7 Chapter One

Recent microcosm studies which have Included biotic interactions In making predictions of

species range shifts have shown that when more than one species was examined, responses were

different to single species treatments (Davies et al. 1998b, Petchey et al. 1999, Fox and Morin

2001). The potential for biotic interactions to alter with changes in climatic conditions has been further emphasised by a long-term study of an arid ecosystem In North America. Increased

precipitation led to changes in the vegetation present, resulting in cascading effects on consumers

and their predators (Brown et al. 1997). The switch from grassland to dense woody shrubs as a

result of increased precipitation resulted in a number of rodent and ant species of open arid

grassland declining in number, presumably as a result of changes in habitat as well as the wetting of seed stores (Brown et al. 1997). The species that showed greatest declines were those that stored seeds (Brown et al. 1997). The loss of these species in turn resulted in a drop in predator species through loss of prey and an increase in species associated with increased shrub cover which benefited from reduced competition and increased cover which reduced predation (Brown et al. 1997). The change in vegetation led to local extinctions and reductions in diversity in some

species, either as a direct result of the change in climatic conditions or as a direct and/or indirect effect of changes in biotic interactions (Brown et al. 1997). The change in the community structure of this ecosystem could not have been predicted based purely on the effects of abiotic factors on the species assemblage (Brown et al. 1997). Furthermore, the complexity of change in the outcome of interactions between species could not have been predicted from simple laboratory experiments (Brown et al. 1997). Due to the complex nature of species interactions, it has been argued that understanding community scale processes is fundamental in order to predict the responses of natural systems to climate change (Callaway et al. 2002).

The use of small scale manipulative experiments may provide information on how biotic interactions alter with changes in the abiotic and biotic environment (e. g. Bertness et al. 1999,

Leonard 2000, Bertness and Ewanchuk 2002). Recent experiments in New England, USA, have shown that both shading and crowding could have negative effects on barnacle recruitment and survival at cooler northern sites, while having positive effects at warmer southern sites (Bertness et 1999, Leonard al. 2000). Thermal buffering by shading and the presence of conspecifics was being suggested as responsible for the increased survival of species at these southern sites

(Bertness et al. 1999, Leonard 2000). This switch from competitive to facilitative interactions occurred after a change of just 1 to 2°C. The results of these two studies indicate that the local 8 Chapter One

persistenceof some species will depend on intra- and interspecificinteractions alleviating the harsher conditions experienced as a result of a warming climate. In terrestrial environments,a switch from competitiveto facilitative interactionshas been shown to alter with changing environmentalconditions across a broad geographicalarea (Callaway et al. 2002). In general positive interactionsoccurred where environmentalconditions were harsher (Callaway et al. 2002).

If the strength and sign of biotic interactionsalter with changing environmentalconditions, it is likely that there will be broad scale implicationsfor assemblagestructure and dynamics. Changes in species assemblagesas a consequence of climate change are likely to result in changes in ecosystem functioning, especially as some species strongly affect ecosystem processes by directly mediating energy fluxes or by altering the abiotic environment (Chapin III et al. 2000).

Manipulative experiments that alter environmental conditions and also biotic interactions should provide insights into the likely impacts of climatic warming on biotic interactions. When results from such manipulations are used in conjunction with predictions made using the'climate envelope' approach, biologically realistic predictions of species range shifts may be possible.

1.5 Rationale of thesis

Rocky intertidal shores provide a unique system for undertaking manipulativeexperiments to understand community ecology because of their accessibility,visibility, substrate stability, attached or near sedentary species and the comparativelylow cost in carrying out experiments

(Lewis 1996, Bertness and Leonard 1997, Underwood2000). As a result there is a long history of the study of organisms on rocky shores and the physiology, behaviour and ecology of the species present are well understood (Newell 1979, Barnes et al. 1993, Little and Kitching 1996, Lobban and

Harrison 1997, Raffaelli and Hawkins 1999, Underwood2000 for reviews). In particular rocky shores have been used to illustrate the importance of biotic interactions in regulating assemblage structure (Jones 1948, Connell 1961b, a, Paine 1966, Dayton 1971, Connell 1972, Menge 1978,

Paine 1979, Lubchenco and Gaines 1981, Menge and Sutherland 1987, Bertness and Callaway

1994, Menge 1995, Bertness and Leonard 1997). Grazers, such as limpets and snails, can influence the abundance and distribution of intertidal organisms both directly and indirectly (see Underwood 1979, Branch 1981, Hawkins and

Hartnoll 1983). As a result they can have an important role in regulating community structure (Southward Southward and 1978, Branch 1981, Hawkins 1981, Lubchenco and Gaines 1981, Chapter One

Hawkins 1983, Hawkins and Hartnoll 1983, Lubchenco 1983, Hartnoll and Hawkins 1985, Hawkins

et al. 1992, Harley 2002, Paine 2002). This is particularly true on moderately exposed shores of

the NE Atlantic where interactions between limpets, macroalgae and barnacles have an Important

role in the structure and dynamics of rocky shore communities (Jones 1948, Burrows and Lodge

1950, Southward and Southward 1978, Hawkins 1981, Hawkins 1983, Hawkins and Hartnoll 1983,

Hartnoll and Hawkins 1985, Hawkins and Southward 1992, Johnson et al. 1997, Burrows and

Hawkins 1998, Jenkins et al. 2001). As a result of the interaction between these groups of

organisms, on many moderately exposed shores of the NE Atlantic there is a mosaic of algal

patches which often result in increased diversity (Thompson et al. 1996, Jenkins et al. 1999) and

can determine whether shores are net producers of primary production or a net importers of

primary production (Hawkins et al. 1992). Therefore, changes in the limpet-Fucus-barnacle

interaction could have broad-scale implications for rocky shore community dynamics and

ecosystem functioning.

The British Isles straddles two major marine biogeographiczones, with both warm

lusitanian and cool boreal biota (Forbes 1853, Lewis 1964), resulting in the co-existence of

ecologically comparable species with northern and southern centres of distribution. As a consequence, many marine species present are at either the northern or southern edge of their biogeographic range (Southward and Crisp 1954, Crisp and Southward 1958, Southward 1963,

Southward et al. 1995). The abundance of some of these species has been shown to fluctuate with changes in climatic conditions (see Southward et al. 2005 for a recent review). As populations close to their range edges are likely to be more sensitive to changes in climate than those at the centre of their range (Lewis 1996), southwest Britain provides an ideal region in which to investigate how species respond to climate change and in particular, the way in which biotic interactions may modify these responses.

The overall aim of my work was to use species of limpets and barnacles with northern and southern biogeographic ranges to investigate how biotic interactions, such as interactions with algal canopies, may modulate species responses to climate change. More specifically I investigated the distribution spatial of the northern/boreal limpet, Pate/la vulgata and the southern/lusitanian limpet P. depressa in relation to Fucus (Chapter 2) and then investigated potential causes of this spatial (Chapter 3). Canopy pattern forming macroalgae are predicted to become restricted to more latitudes northerly or refuges buffered from environmental extremes on temperate shores 10 Chapter One

(Breeman 1990, Barry et at. 1995, Beardall et al. 1998, Coelho et at. 2000). Therefore,

consequencesof Fucus loss on the behaviour and mortality of P. depressa and P. vulgata was also investigated(Chapter 3).

Due to similarities in the ecology of P. depressa and P. vulgata, it has been assumed that

they both have similar roles in controlling the abundance of macroalgae in the intertidal (e. g.

Thompson et at. 2000, Jenkins et al. 2001, Boaventura et at. 2002, Jenkins et at. 2005). The

behaviour of these species has, however, been shown to differ (Chapters 2,3 and 4). Therefore,

differences in the grazing activity of these species were explored further (Chapter 4). More

specifically, the relative roles of P. vulgata and P. depressa in controlling macroalgal abundance on

shores of southwest Britain was investigated with particular emphasis on the effects of time of

reproduction and the effect of temperature (sea and air) on limpet grazing activity.

The balance of species with northern and southern biogeographic distributions is expected

to alter with climatic warming (Southward et at. 1995). Using two barnacle taxa, I investigated the

effects of intra- and interspecific competition on population structure under predicted warming

scenarios for southwest Britain. The effect of adult Semibalanus balanoides abundance on its

gregarious settling cyprids was investigated to see if the presence of con-specifics affected

recruitment success. Subsequently, the effect of S. balanoldes recruitment success on the

recruitment success of the competing barnacle species Chthamalus spp. was investigated

(Chapter 5).

The general discussion (Chapter 6) synthesises the outcomes of this work in order to

inform predictionson the likely effects of increased climatic warming on rocky shore community dynamics in southwest Britain. In particular, the interaction of limpets, macroalgaeand barnacles are discussed in relation to how species responses to climate change may be modified by biotic interactions. Conceptual models are proposed to explain how the relationship between limpets, barnacles and Fucus alters as a consequenceof changes in the relative abundance of the dominant limpet and barnacle species. Finally, the Importanceof incorporatingfactors such as biotic interactions into predictions of species range shifts is discussed with reference to the use of the 'climate envelope' approach.

11 ChapterTwo

Chapter Two

The association between two limpet species and patches of fucoid

macroalgae

2.1 Introduction

Grazers have an important role in controlling the quantity and spatial distribution of vegetation in marine (Lubchenco and Gaines 1981, Underwood and Jernakoff 1981, Hawkins and

Hartnoll 1983b, Hartnoll and Hawkins 1985, Hawkins et al. 1992), terrestrial (Hulme 1994) and freshwater systems (see Feminella et al. 1989 and references within). Macroalgal abundance in the NE Atlantic is predominantly controlled by the grazing activity of patellid limpets (Jones 1948,

Burrows and Lodge 1950, Southward and Southward 1978, Hawkins 1981, Hawkins and Hartnoll

1983b, Hawkins and Hartnoll 1985, Hawkins et al. 1992, Burrows and Hawkins 1998, Johnson and

Hawkins 1998), although grazing by trochids and littorinids can also be important (Hawkins and

Hartnoll 1983b). At mid-shore levels in the NE Atlantic the key grazers are Patella vulgata, a northern/boreal limpet distributed from northern Norway to southern Portugal (Fischer-Piette 1948,

Bowman and Lewis 1986, Guerra and Gaudencio 1986, Southward et al. 1995, Jenkins et al. 2001) and P. depressa, a southern/lusitanian limpet distributed from Senegal in West Africa to north

Wales, UK (Fischer-Piette 1948, Bowman and Lewis 1986, Southward et al. 1995). The geographic distributions of these two species overlap between southern Portugal and north Wales,

UK. At present P. vulgata is the numerically dominant limpet species on most shores in Britain, while from northern Spain to the range limit of P. vulgata the numerically dominant limpet species is

P. depressa (Ballantine 1961, Jenkins et al. 2001, Boaventura et al. 2002c).

In Britain the relative abundance of these two species has been shown to fluctuate, broadly correlatingwith changes in climatic conditions (Southward et at. 1995, Kendall et al. 2004,

Hawkins, unpub). During the 1950s, a warmer climatic period, the relative abundanceof P. depressa increased, becoming common and occasionally the dominant limpet on some southwest shores (Crisp and Southward 1958). This pattern reversed during the cool period from the 1960s to the early 1980s (Southward et al. 1995, Kendall et al. 2004, Hawkins, unpub. ). As temperatures have risen since the mid-1980s the proportion of P. depressa to P. vulgata has again increased on some shores (Southward et al. 1995). As global warming causes temperatures to rise further it is anticipated that the biogeographic ranges of Patella vulgata and Patella depressa will alter, with a

12 Chapter Two

general pole-ward shift in their ranges (Southwardet al. 1995). It Is therefore, likely that the dominant limpet on rocky shores of southwest Britain will gradually change from P. vulgata to P. depressa, and there is evidence that this is already occurring on certain shores In north Cornwall

(Hawkins, unpub.).

The patellid limpets, and especially P. vulgata, have been well studied including research on their morphology and physiology (e. g. Orton et al. 1956, Orton and Southward 1961, Bowman and Lewis 1977, Bowman 1981, Branch 1981, Bowman and Lewis 1986, Guerra and Gaudencio

1986, Hawkins et at. 1989, Morais et at. 2003), behaviour (e. g. Hartnoll and Wright 1977, Branch

1981, Chelazzi 1990, Chelazzi et al. 1990, Della Santina et al. 1994) and ecology (e.g. Underwood

1979, Branch 1981, Guerra and Gaudencio 1986, Hawkins et al. 1992, Johnson et al. 1997,

Burrows and Hawkins 1998, Delany et al. 1998, Jenkins and Hartnoll 2001, Jenkins et al. 2001,

Boaventura et at. 2002b, Boaventura et at. 2003). P. vulgata and P. depressa are generalist grazers consuming microbial films, which are made up of a matrix of mucus, bacteria, microalgae and macroagal propagules (Hawkins et at. 1989). It is generally thought that prosobranch limpets are unspecialised grazers feeding upon whatever they encounter that is small enough to consume on the rock surface (Hawkins and Hartnoll 1983b), including, in the case of P. vulgata, barnacle plates, calcareous algae and small gastropods (Hawkins et at. 1989, Hill and Hawkins 1991).

Both P. vulgata and P. depressa occupy a 'home scar' to which they return between foraging excursions (Hartnoll and Wright 1977). Such behaviour ensures limpets maintain a close fit with the substratum which reduces emersion stress at low tide, the risk of dislodgementthrough wave action at high tide and the risk of predation at either high or low tide (Hartnoll and Wright

1977, Hawkins and Hartnoll 1983b, Chelazzi 1990). Both species remain loyal to these home scars unless disturbed, for example, by changes in environmentalconditions (Jones 1948, Burrows and Lodge 1950, Hawkins and Hartnoll 1983a).

Grazing by limpets is thought to regulate the abundance of macroalgae,such as Fucus spp., by grazing their propagules in the microbial film (Jones 1948, Southward and Southward

1978, Hawkins and Hartnoll 1983b, Hartnoll and Hawkins 1985). There is also some evidence to suggest that Patella spp. can also control macroalgal abundance by consuming mature stages of macroalgae (Hawkins pers. comm.). Thus, it is thought that both P. vulgata and P. depressa have a key role in controlling the abundance of macroalgae on shores of the NE Atlantic (Southwardand Southward 1978, Hawkins et al. 1992, Boaventura et al. 2002a, Jenkins et al. 2005). 13 Chapter Two

Canopy forming macroalgae, such as the fucoids, are conspicuousand important

members of the community on all but the most exposed rocky shores of the NE Atlantic (Lewis

1964, Stephensonand Stephenson 1972). In aquatic systems macroalgalcanopies (Menge 1978,

Hicks 1980, Bertness et al. 1999), freshwater macrophytes(Petry et al. 2003, Taniguchi et al.

2003) and seagrass beds (Heck and Wetstone 1977) are all recognised as having key roles in

structuring benthic communities by modifying the physical environment. Intertidal environments

are physically harsh and as a consequencespecies that reduce environmentalstress, such as fucoid algae, can provide an important habitat for a variety of species (Hawkins et al. 1992,

Thompson et al. 1996, Bertness and Leonard 1997, Bertness et at. 1999, Leonard 2000), including limpets, anemones, dogwhelks, littorinids, isopods and amphipods. Fucus patches appear to provide an important habitat for P. vulgata juveniles in particular (Southward and Southward 1978,

Hawkins 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992), because they still require habitat amelioration once they leave 'nursery' pools (Bowman 1981). It is thought that P. vulgata and other intertidal species may preferentially aggregate beneath Fucus patches because they ameliorate temperature extremes (Burrows and Lodge 1950, Hawkins and Hartnoll 1983b, Hawkins and Hartnoll 1983a, Hartnoll and Hawkins 1985, Johnson et al. 1997, Burrows and Hawkins 1998,

Johnson et al. 1998), increase relative humidity (Thompson et al. 1996), and possibly because higher concentrations of microalgal food are found beneath them (Thompson et al. 2004). As a result of the microhabitat provided by Fucus patches they often have enhanced diversity beneath them relative to adjacent open rock (Thompson et at. 1996, Jenkins et al. 1999). Therefore P. vulgata has an important role in controlling the abundance of macroalgae on the shore, at the same time as macroalgae providing an important habitat for P. vulgata and other intertidal species.

A common feature on moderately-exposedshores of the NE Atlantic is the presence of a mosaic of variously aged macroalgal patches (Southward 1953, Lewis 1964). In part this is a result of P. vulgata preferentially aggregating beneath Fucus patches, leading to areas of reduced grazing away from these limpet aggregations and hence higher survival of Fucus propagules

('Fucus escapes') (Hartnoll and Hawkins 1985, Burrows and Hawkins 1998, Johnson et al. 1998).

The interaction between barnacles and limpets is also important in allowing Fucus escapes to occur as P. vulgata is a less effective grazer across barnacle covered rock allowing Fucus propagules that settle amongst barnacles to escape grazing (Jones 1948, Burrows and Lodge 1950, Southward and Crisp 1956, Hawkins 1981, Hawkins 1983, Hawkins and Hartnoll 1983b, 14 Chapter Two

Hawkins et al. 1992). Thus P. vulgata not only controls the amount of macroalgae on the shore, but also plays an important role In determining the spatial structure and dynamics of rocky shore communities.

Most work on the role of limpets in regulating rocky shore community structure in the NE

Atlantic has been carried out on shores where P. vulgata is the sole limpet species present, mainly on the Isle of Man (see Hawkins et al. 1992 for review). In studies which Included both P. vulgata and P. depressa (Southward and Southward 1978, Thompson et al. 2000, Jenkins et al. 2001,

Boaventura et al. 2002a, Jenkins et al. 2005) it has been assumed based on their similar feeding behaviour, physiology and ecology that both species have similar roles in controlling the abundance of Fucus and therefore structuring rocky shore communities. However, little has been done to quantify the spatial distribution of P. depressa in relation to Fucus or to assess the specific contribution of P. depressa to rocky shore community dynamics. Initial observations suggest that

P. depressa does not preferentially aggregate beneath Fucus patches (Roberts 2002, Hawkins pers obs. ) and following the Torrey Canyon oil spill, P. depressa was much quicker to recolonise barnacle covered rock than adjacent Fucus dominated areas. It has therefore been suggested that

P. depressa does not prosper beneath Fucus patches (Hawkins and Southward 1992).

If these two species of limpet do exhibit different spatial distributions in relation to Fucus, then they may also play different roles in structuring rocky intertidal communities. If the relative abundance of the two limpet species changes, in response to climatic warming, then there could be broadscale consequencesfor rocky shore community dynamics. The overall aim of this chapter was to investigate the interaction between Fucus patches and both P. vulgata and P. depressa on shores where both species co-exist. Initially I compared the occurrence of P. depressa beneath

Fucus patches and on open rock at a broad spatial scale (100's of kilometres) before making detailed finer scale investigationsof the spatial distribution and population structure of P. depressa and P. vulgata on selected shores on the north coast of Cornwall.

15 Chapter Two

2.2 Materials and methods

2.2.1 The distribution of Patella depressa In relation to Fucus patches

To test the hypothesis that, at a broad spatial scale, a greater proportion of P. depressa to

P. vulgata would be found on open rock compared to beneath Fucus patches, the proportion of P. depressa to P. vulgata was compared between haphazardly selected Fucus patches (clumps of

Fucus plants between 0.09-0.25m2 in size) and adjacent areas of open rock (areas of rock not covered or swept by macroalgae and less than 60% barnacle cover) of approximately the same size at five locations on the south and southwest coast of Britain (Fig 2.1). For each habitat (plus or minus Fucus) the relative abundance of each species of limpet was recorded from three 0.25 x

0.25m quadrats. The proportion of P. depressa to P. vulgata inside and outside Fucus patches was compared using a two factor ANOVA (WinGmav5) with the factor location (five levels) considered random and the factor habitat (two levels, plus or minus Fucus) considered fixed.

2.2.2 The spatial distribution and population structure of P. vulgata and P. depressa in

relation to Fucus patches

Focused finer scale surveys were then undertaken, at shores on the north coast of

Cornwall, to look at differences in the spatial distribution and population structure of P. vulgata and

P. depressa at locations where both species co-exist. Differences in the distribution of P. depressa and P. vulgata in relation to Fucus were compared to test the hypothesis that a greater relative abundance of P. vulgata should be found beneath Fucus patches compared to open rock and a greater relative abundance of P. depressa should be found on open rock compared to beneath

Fucus patches. The abundance of P. vulgata and P. depressa were recorded in Fucus patches and on areas of open rock using ten haphazardly placed 0.5 x 0.5m quadrats for each species- habitat combination at each of two, locations, Trevone (500 55'N, 40 98'W) and Crackington Haven

(50° 74'N, 41164W) on the north coast of Cornwall, UK (Fig 2.1). The effect of macroalgal canopy on the relative abundance of P. vulgata and P. depressa was compared using a three factor

ANOVA with the factor location (two levels) considered random and the factors habitat (two levels, plus or minus Fucus) and species (two levels) both considered fixed. In order to increase the factors power of of interest post-hoc pooling was utilised to remove non-significant terms (p>0.25; Winer 1971, Underwood 1997). For all ANOVAs heterogeneity of variance was tested using

16 Chapter Two

Cochran's test and where appropriate data were transformed. Student-NewmanKeuls (SNK) post- hoc tests were carried out on significant data (p < 0.05).

To test the hypothesis that the proportion of P. depressa to P. vulgata should decrease with increasing Fucus cover 0.5 x 0.5m quadrats were, randomly placed at two locations, Trevone

(n=120) and Newquay (n=80; 50° 27'N, 5° 05'W) on the north coast of Cornwall (Fig 2.1). The percentage cover of the area covered by Fucus plants was quantified using the point intercept method and the relative abundance of each species was recorded. Data were Arcsine transformed and analysed using model one regression. To compare for generality between the two shores the slopes of the two regressions were analysed using analysis of covariance (ANCOVA).

To investigatethe level of aggregation of P. depressa and P. vulgata over a larger spatial area the relative abundances of P. depressa and P. vulgata were quantified for a 9m2 area of mixed Fucus patches and open rock at Trevone, on the north coast of Cornwall, UK. It was hypothesised that P. vulgata should be more contiguous beneath Fucus patches than on open rock. Conversely, P. depressa should avoid Fucus and be more randomly distributed on open rock. Using coordinates from a1 m2 quadrat separated into 0.03m2 grid squares the presence of

Fucus was noted and any limpets within the grid squares were identified and counted. The association of the two limpet species with Fucus and open rock habitats was analysed using a two- way contingency x2 test. The distance of each limpet to their nearest conspecific neighbour was calculated for both species in each of the two habitats.

To compare the population structure of limpets found on open rock and beneath Fucus patches size frequency distributionswere calculated. Limpet lengths for both species from the two habitats were collected from five locations: Kimmeridge,Corbyn's Head, Brixham, Salcombe and

Trevone on the south and southwest coast of Britain (Fig 2.1). For each habitat shell length of each species of limpet was recorded from three 0.25 x 0.25 m quadrats. Due to the scarcity of P. depressa beneath Fucus patches this species was excluded from formal statistical analysis.

Cumulative size frequencies of P. vulgata from each habitat (beneath Fucus patches and on open rock) were compared using a two sample Kolmogarov-Smirnoftest for each location.

17 Chapter Two

Crackingon Haven -"Osmington -t ' Trevone ý., Corbynrsýý *ICimmeridge Head Newquay Brixham 50km Salcombe

Fig 2.1. Locationof study sites on south and southwestcoast of Britain.

2.3 Results

2.3.1 The distribution of Patella depressa In relation to Fucus patches

Across a broad geographical scale (100's of kilometres) the proportion of P. depressa to P. vulgata was consistently higher on open rock and either low or zero beneath Fucus patches (Fig

2.2). The effect was significant with post-hoc SNK tests indicating that across all five locations higher proportions of P. depressa to P. vu!gata were found on open rock compared to beneath

depressa Fucus patches (F 1,484.02, P <0.01; Table 2.1). At one location, Corbyn's Head, P. was slightly although not significantly more abundant beneath Fucus patches than at any of the other locations (Fig 2.2). This may be a result of the macroalagae on this shore being juvenile Fucus plants (frond length of less then 0.01 m Knight and Parke 1950) at the time of sampling. Previous work carried out on the Isle of Man has shown that P. vulgata does not migrate into newly settled

Fucus patches, but waits until the patches are older (Hawkins and Hartnoll 1983a).

The proportion of P. depressa to P. vulgata differed across locations irrespectiveof the presence of Fucus, with the proportion of P. depressa to P. vulgata generally decreasing eastwards

01; along the English Channel coast (Fig 2.2). The significant effect of location (F 4,204.76, P

Table 2.1) was unsurprising considering the geographical area covered in the survey. This included areas near the range edge of P. depressa, where the proportion of P. depressa to P. vu/gata was low, to areas in the middle of the range of P. depressa in Britain, where the proportion of P. depressa to P. vu/gata neared 0.5 (see Crisp and Southward 1958, Kendall et al. 2004). It is perhaps interesting to note that at Trevone and Torbay where P. depressa was most common, this species was also found in small proportions beneath Fucus patches (Fig 2.2). In contrast, further

18 ChapterTwo

east along the English Channel coast where the abundance of P. depressawas not so high, no P. depressa individualswere found beneath Fucus patches.

Table.2.1. Analysis of variance of the proportionof P. depressato P. vulgatafound beneath Fucus patches and on open rock at five locations around the south and southwestcoast of Britain. Varianceswere homogenousfollowing arcsine transformation(Cochran's test 0.36; ns).

Source SS DF MS F P F versus Location (Lo) 4044.52 4 1011.13 4.76 <0.01 Residual Habitat (Ha) 17832.65 1 17832.65 84.02 <0.01 Lo x Ha Location x Habitat 1378.53 4 344.63 1.62 0.21 Residual Residual 4244.72 20 212.24 Total 27490.42 29 SNK tests Habitat Location Open rock > Fucus patch Trevone = Corbyn's Head > Brixham = Osmington = Kimmerid ge

2.3.2 The spatial distribution and population structure of P. vulgata and P. depressa in

relation to Fucus patches

At a smaller spatial scale, differenceswere found in the spatial distribution of P. vulgata and P. depressa in relation to Fucus patches. The abundance of P. vulgata was significantly

higher beneath Fucus patches than on open rock. In contrast, the abundance of P. depressa was Table 2.2; significantly higher on open rock than beneath Fucus patches (F 3,73133.87; P <0.01; Fig 2.3).

There was a clear negative relationship in the proportion of P. depressa to P. vulgata with

increasing Fucus cover (P<0.01; Fig 2.4) at Newquay and Trevone. There was no significant for difference in the ANCOVA comparing the differences between the slopes of the two regressions

in the of P. depressa the two locations (F 1,1960.13; P>0.05), indicating that the reduction proportion

to P. vulgata with increasing Fucus cover was consistent between locations.

19 ChapterTwo

Table 2.2. Analysis of variance of the abundanceof P. vulgata and P. depressa beneathFucus and on open rock at CrackingtonHaven and Trevone. Varianceswere homogenousfollowing Ln (X + 1) transformation (Cochrans0.2008, ns). The location x habitat x species interactionwas non-significant(P >0.25) and was thus pooledwith the residualto increasethe power of the test for the treatment x species interaction.

Source SS DF MS FPF versus Location (Lo) 3.84 1 3.84 15.54 <0.01 1-Pooled Habitat (Ha) 9.70 1 9.69 99.57 0.06 Location x Habitat Species (Sp) 30.46 1 30.46 504.05 0.02 Location x Species Location x Habitat 0.10 1 0.09 0.39 0.53 1-Pooled Location x Species 0.06 1 0.06 0.24 0.62 1-Pooled Habitat x Species 33.07 1 33.07 133.87 <0.01 1-Pooled 0.04 1 0.04 0.15 0.69 1-Pooled Lo x Ha x Sp Residual 17.99 72 0.25 Total 95.25 79 1-Pooled 18.03 73 0.25 SNK tests Habitat x Species interaction P. vulgata Open rock < Fucus patch P. depressa Fucus patch

(236 Over a 9m2 area, greater numbers of P. vulgata were found beneath Fucus patches 27% individuals) compared to on open rock (96 individuals) even though Fucus made up only of the available habitat. In contrast, more P. depressa were found on open rock (41 individuals) with

only thirteen P. depressa individuals found beneath Fucus. There was a statistically significant

association between habitat and the two limpet species with P. vulgata in particular, associated with Fucus (chi squared stat = 42.77,1 d. f., P<0.01). When distance to nearest neighbour was have calculated, 86% of P. vulgata individuals located beneath Fucus patches were found to

conspecifics in either the same 0.03m2 grid square or the adjacent one (Fig 2.5a). On open rock 59% the number of individuals found in either the same grid square or the adjacent one reduced to

(Fig 2.5b). Fucus was the less extensive of the two habitats and perhaps made patterns of

aggregation more evident for P. vulgata. In contrast, P. depressa individuals were more evenly

dispersed on open rock with only 31 % of individuals found within 5cm of a conspecific and with

over 43% of P. depressa individuals found 20cm or more from a conspecific (Fig 2.5d). As only

thirteen P. depressa individuals were found beneath Fucus it is hard to interpret this species level

of aggregation beneath Fucus patches.

There were no significant differences between the size frequency distributionsof P. vulgata

individuals found beneath Fucus patches and on open rock at Salcombe, Brixham, Corbyn's Head

or Kimmeridge (Fig 2.6b-e), indicating that the population structure on open rock and beneath

Fucus patches was similar. There were however, differences in the size frequency distributions of

P. vulgata between locations. At Salcombe there were generally higher proportions of small 20 ChapterTwo

individualsthan at Brixham, Corbyn's Head and Kimmeridge,which were made up of a range of size classes (Fig 2.6). At Trevone there was a significant size frequency difference between P. vulgata individuals from Fucus patches and those from open rock habitats (P<0.05; Fig 2.6a).

Smaller P. vulgata were found on open rock compared to beneath Fucus patches, with no limpets larger than 19mm being found on open rock (Fig 2.6a).

The size frequency distribution of P. depressa differed between the five locations. At

Kimmeridge there were no P. depressa individuals smaller than 16mm, indicating that there had (Fig been no recent recruitment at this shore near the eastern range limit in Britain of P. depressa

2.7). The size frequency distributions of P. depressa at Salcombe, Brixham, Corbyn's Head and

Trevone were all similar with a range of juvenile and adult limpets, although there was a higher percentage of smaller P. depressa individuals at Trevone compared to the other three locations

(Fig 2.7).

21 ChapterTwo

** M. I. 1.0

0.s 0 0.6 ® Fucus patch 4 0.4 Q Open rock

0.2 0 c 0.0 0 t 0 CL 2 CL. 0yß' ý. ccý Goýýý

five Figure 2.2. Proportion of P. depressa to P. vulgata found beneath Fucus patches and on open rock at locations around the coast of south and southwest Britain. Error bars ±1 SE. `" P<0.01.

** 30 44 25 N,n Ö 20 V C 0 Fucus patch 15 c Q Open rock J2 10 wd a5 E J 0 P. vulgafa P. depressa

This Figure 2.3. The effect of macroalgal cover on the relative abundance of P. vulgata and P. depressa. been effect was consistent at both Crackington Haven and Trevone therefore data from both locations have combined. Error bars ±I SE. ** P<0.01.

22 ChapterTwo

M. I.ea 100 I y=-0.37x+37 a) Newquay CL R2= 0.56 80 P <0.01

60

40 " " " f ý c f ý 20 ý ý o " " 0 CL 0 0 0 20 40 60 80 100 D. Percentage cover Fucus

ß 100 + 42 b) Trevone y= -0.35x R2=0.60 80 0 P<0.01 Nf 60 o. 40 """ 0"" r- 20 0

ä. 0 0 IL 0 20 40 60 80 100 Percentage cover Fucus

Figure 2.4. The proportionof P. depressato P. vulgatawith increasingcover of Fucus at a) Newquayand b) Trevone.

23 Chapter Two

P. vulgata

a) Beneath Fucus (n=236) b) Open rock (n=96) 1.0 1.0

0.8 0.e 12 g 0.6 E 0.e

0.4 0.4

0.2 0.2

0.0 0.0 0-5 5.10 10-15 15-20 20-25 25-30 30-35 35-40 40.45 45.50 50. 0.5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50+

Distance from nearest neighbour (cm) Distance from nearest neighbour (cm)

P. depressa

c) Beneath Fucus (n=13) d) Open rock (n=41) 1.0 1.0

08 0.$

aE 06 E_ 0.0

0.4 0.4

0.2 02

0.0 0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 60+ o.o as 5-10 10.15 15-20 20-25 25-30 90.96 35-10 40-45 45-50 50. (cm) Distance from nearest neighbour (cm) Distance from nearest neighbour

Figure 2.5. The distance limpets were found from their nearest conspecificneighbour at Trevone.

24 Chapter Two

a) Trevone b) Salcombe 0 100 100 V-r-. /n=36 -0000, 80 80 4-60 n= 12 60 0) 40 40 G) " M 20 . P<0.01 20 1/-lf :1 0 17 21 25 29 33 37 41 <5 9 13 17 21 25 29 33 37 41 <5 9 13

Brixham Corbyn's Head C) d) 100 n22 too ------n=11 90 _J so " c n25 '--- .= Bo- so . n7 a) N 'N 40 40- G) " 20 20 P>0.10 - P<0.10 Eýo" 0 29 33 37 41 <5 9 13 17 21 25 29 33 37 41 <5 9 13 17 21 25

ö Kimmeridge eý c 100 a) = , 80 "'

N 60 n=51, '' n=64 Cl) 40 > 20. " P>O. 10

<5 9 13 17 21 25 29 33 37 41 U Limpet shell length (mm)

Fig 2.6. Cumulativesize frequenciesof P. vulgata Inside and outside Fucus patches at five locationsaround line. P- the south and southwestcoast of Britain. Open rock - bold line, Fucus patches- dotted values are results from two sample kolmogarov-Smirnoftests to compare size frequencydistributions of P. vulgata beneath Fucus patches and on open rock.

25 ChapterTwo

100 --. - 80 "'"/ý ýý , "ýý, Imo, 60

ir+u

0 <5 9 13 17 21 25- 29 33 37 41 Limpet shell length (mm)

Trewne (n=26) ---- Salcombe (n=93) --"---" Brixham (n=50) -"--- Corbyn's Head (n=32) ----- Kimmeridge (n=18)

Fig 2.7. Cumulativesize frequency distributionof P. depressaon open rock at five locationson the south and southwestcoast of Britain.

26 ChapterTwo

2.4 Discussion

The results presented here perhaps most importantly show that P. vu/gata and P. depressa exhibit different spatial distributions on the shore in relation to Fucus patches and to a lesser extent in relation to open rock habitats. P. vulgata preferentially aggregates beneath Fucus patches at all locations studied, supporting previous work (Jones 1948, Hawkins and Hartnoll 1983a, Burrows been and Hawkins 1998). In contrast, across a broad spatial scale (100s km), P. depressa has shown to rarely create home scars beneath Fucus patches. Results also indicate that P. vulgata 86% appears to have a more aggregated distribution than P. depressa irrespective of habitat with of P. vulgata individuals found beneath Fucus patches and 59% of P. vulgata individuals on open rock being located within 10cm of a conspecific. In contrast, just 31 % of P. depressa found on open rock were found at a similar distance from a conspecific. Hence it would appear that not only do these two limpet species behave differently in relation to Fucus patches, but they also have a different spatial distribution on open rock.

Both species of limpet are able to tolerate temperaturesup to around 43°C (P. vulgata

42.80C;P. depressa 43.3°C; Evans 1948) so it is unlikely that differences in the spatial distribution It on the shore of either species is as a result of differences in the lethal effects of temperature.

could be that sub-lethal stresses experiencedby P. vulgata are reduced by aggregating beneath

Fucus patches. Initial observations indicate that the gonads of P. vulgata develop quicker beneath

Fucus patches than on open rock (Moore, pers. ob.) and P. vulgata individualsfound beneath (17.1 Fucus patches are known to experience greater annual growth rates beneath Fucus mm yr

than on barnacle covered rock (14.7mm yr-1) (Fretter and Graham 1976).

The present study provides the first quantitative observations showing that P. depressa does does not preferentially aggregate beneath Fucus patches. It is not clear why P. depressa not

preferentially aggregate beneath Fucus patches, but a number of hypotheses can be suggested.

Fucus is the dominant space occupying species at northern latitudes, but becomes less dominant

at lower latitudes being replaced by smaller turf forming algae (Ballantine 1961). Fucus

vesiculosus is the dominant mid-shore fucoid on moderately exposed shores in northern latitudes,

but is rare on open coasts becoming restricted to estuaries from the Bay of Biscay to its southern

range limit in Morocco (Luning 1990, Murias Dos Santos 2000, Boaventura et al. 2002c, Alcock

2003). In contrast, P. depressa is more commonly found on moderately exposed to exposed

shores where fucoids are absent over much of its range (Fischer-Piette and Gaillard 1959, Hawkins 27 ChapterTwo

et al. 1992, Alcock 2003). Hence, the lack of a preferencefor Fucus patches by P. depressamay

be a direct result of the absence of Fucus over much of its range. The biogeographicrange of P.

depressa extends to Senegal in West Africa with this species reaching its northern range limit in

Britain. Therefore P. depressa is unlikely to experience thermal stresses in Britain that might

require it to seek habitat ameliorationfrom canopy forming algae. This theory requires testing as

there may be latitudinal clines in the tolerance of P. depressa populationsto high temperatures.

On shores of SW Britain, P. depressa is a summer breeder with peak spawning times coinciding with maximum air temperatures and possibly increased wave action (Orton and

Southward 1961, Fretter and Graham 1976). Towards the southern range limits of P. depressa gonad activity drops with increasing temperatures (Guerra and Gaudencio 1986), which may reflect the adverse effects of excessive warmth near this species southern range limits (Lewis et al. 1982).

On shores in southwest Britain higher numbers of P. depressa are found on warmer southern facing rock surfaces than on cooler northern facing surfaces (Hawkins, unpub). This may indicate that P. depressa near its northern range limit seek habitats that increase the amount of heat absorbed, either for gonad development or for some other physiological reason, such as grazing activity or digestion.

Both species of limpet settle into permanentlywet places such as pools (Bowman 1981,

Delany et al. 1998), with P. vulgata emerging out of pools at the end of their first year. P. vulgata individuals still require protection from temperature and desiccation stress and are therefore restricted to damp microhabitats such as crevices, beneath Fucus patches, amongst barnacles and mussels (Bowman 1981). Thus whilst Patella spp. control the establishment of fucoid stands, once established fucoids provide ideal 'nursery' grounds for P. vulgata (Southward and Southward

1978). P. depressa does not emerge from 'nursery' pools until the end of their second or third year

(Bowman 1981) by which time they may already be able to tolerate the environmental conditions experienced on open rock. Higher concentrations of microalgal food are found beneath Fucus patches than on open rock (Thompson et al. 2004), and although P. depressa may not require the habitat amelioration provided by Fucus patches it would be reasonable to expect P. depressa to home create scars beneath Fucus patches to exploit this food resource. It may be that P. depressa avoids preferentially aggregating beneath Fucus patches to reduce inter- and intraspecific competition, as lower overall limpet densities are found on open rock compared to beneath Fucus

28 ChapterTwo

patches. Further experiments to examine the behaviour of the two limpet species when exposed to uncolonisedFucus patches will test this hypothesis(Chapter 3).

The lack of significant differences in the size frequency distributions of P. vulgata populations found beneath Fucus patches and open rock was surprising considering previous studies, which have found that juvenile P. vulgata require habitat amelioration once they emerge from 'nursery' pools (Bowman 1981), resulting in a greater proportion of smaller size classes beneath Fucus patches than on open rock (Hawkins 1983, Burrows and Hawkins 1998). Since that in these studies were carried out at locations further north than the present work it is possible the warmer southwest of Britain, all size classes of P. vulgata may need to preferentially aggregate to under Fucus patches. The high proportion of larger P. vulgata found beneath Fucus compared open rock at Trevone may be a result of larger individuals competitively excluding smaller individuals from Fucus patches in areas where it is the preferred habitat. Intraspecific competition has been shown to occur in both P. vulgata (Boaventura et al. 2002b) and P. depressa

(Boaventura et al. 2003), with juveniles or smaller individuals more likely to lose out (Lewis and

Bowman 1975). Although increased temperatures may explain the different size frequency distribution of P. vulgata beneath Fucus patches and on open rock in this study, it is also worth noting that previous studies (Hawkins 1983, Hartnoll and Hawkins 1985, Johnson et al. 1997) were that in carried out on shores where the rock was smoother. On rock which is more rugose, such as the current study, smaller limpets may find some habitat amelioration in crevices on open rock.

29 ChapterThree

Chapter Three

Behaviour of a northern and southern species of limpet in relation to Fucus

patches

3.1 Introduction

Canopy forming algae are conspicuous members of the community on all but the most exposed rocky shores of the North Atlantic (Lewis 1964, Stephenson and Stephenson 1972) and provide an important habitat for a variety of other species (Hawkins et al. 1992, Thompson et al. in 1996, Bertness and Leonard 1997, Bertness et at. 1999, Leonard 2000). This often results

increased diversity beneath canopy forming algae compared to areas of adjacent open rock the (Thompson et al. 1996, Jenkins et at. 1999b). Therefore this key habitat has a large impact on

distribution of species in the intertidal and in turn for rocky shore community structure and

dynamics.

Fucus vesiculosus,the main canopy forming macroalgae at mid-tide on moderately

exposed shores in the NE Atlantic becomes less abundant with decreasing latitude (Ballantine

1961) and becomes rare and restricted to very sheltered and estuarine shores from southern

France to its biogeographicrange limit in Morocco (Luning 1990, Veiga et al. 1998, Calvo et al.

1999, Murias Dos Santos 2000, Boaventuraet al. 2002c). At these lower latitudes grazers and

barnacles become the dominant organisms (Ballantine 1961) and the large canopy forming algae

are replaced by functionally different low growing turf forming algae (Ballantine 1961, Murias Dos

Santos 2000, Boaventura et al. 2002a, Boaventura et al. 2002c). In the British Isles the balance

between grazers and canopy forming fucoid algae is delicately poised (Southward and Southward

1978, Hawkins 1983).

In chapter two I showed that where P. vulgata and P. depressa co-exist on moderately

exposed shores in southwest Britain, they have differing spatial distributions in relation to Fucus.

P. vulgata preferentially aggregates beneath Fucus patches, perhaps as a result of Fucus patches

ameliorating temperature extremes (Burrows and Lodge 1950, Hawkins and Hartnoll 1983b,

Hawkins and Hartnoll 1983a, Hartnoll and Hawkins 1985, Johnson et al. 1997, Burrows and

Hawkins 1998, Johnson et al. 1998), increasing relative humidity (Thompson et al. 1996), and

30 Chapter Three

having higher concentrationsof microalgal food beneath them (Thompsonet at. 2004). In contrast,

P. depressa does not preferentially aggregate beneath Fucus.

Although P. depressa may not need the habitat amelioration provided by Fucus patches, there is evidence to suggest that there is a higher concentration of microalgal food beneath Fucus

patches than on open rock (Thompson et at. 2004). It could therefore be suggested that P.

depressa should create home scars beneath Fucus patches to take advantage of this increased food resource, particularly during summer months when there is a reduced abundance of microalgal food (Thompson et at. 2000, Jenkins et at. 2001) and some evidence to suggest that this food source is limited (Thompson et al. 2000). It may be that P. depressa does not preferentially aggregate beneath Fucus patches because it is competitively excluded by P. vulgata. Competition between individuals or different species occurs when there is an adverse effect on a species or individual trying to acquire a common resource (Birch 1957). In this case amelioration from environmental extremes or increased food resources. Hence, due to similarities in the diet and radula structure of P. depressa and P. vulgata (Hawkins et at. 1989) it is likely that these species will experience some interspecific competition for resources in Britain as they do in Portugal

(Boaventura et at. 2002b). It could be that P. depressa creates its home scars on open rock, where there are lower limpet densities, to avoid or reduce interspecific competition with P. vulgata for resources and/or space. Similar explanations have been suggested for why P. vulgata emigrated out of 'nursery' pools where the density of P. vulgata and P. ulyssiponensis was high into habitats outside of rockpools where limpet density was lower (Delany et al. 1998).

Although canopy-formingmacroalgae provide an important habitat for a variety of intertidal species (Hawkins et al. 1992, Thompson et al. 1996, Bertness and Leonard 1997, Leonard 2000), including P. vulgata, anemones, dogwhelks and littorinids, little research has been carried out on algal responses to climate change. It is thought that temperature and light are probably the most important factors In limiting the biogeographicdistributions of many algal species (Breeman 1988,

Bolton and Anderson 1990, Lobban and Harrison 1997), with small rises in temperature leading to major changes in algal distribution (Breeman 1990, Beardall et al. 1998). Barry et al. (1995) found that a sea temperature rise of 2.2°C over a sixty year period was correlated with an increased dominance of low growing turf forming algae at the expense of larger fucoid species. It is believed increased that temperatures as a result of global warming will lead to an extension of the summer sterile period experienced by fucoid algae and increased storminess will reduce the fertilisation 31 ChapterThree

success and gamete release of some species (Coelho et al. 2000). It is also predicted that increases in photosynthetic active radiation (PAR) and ultra violet light will be lethal to algae at critical times of development (Franklin and Forster 1997). As a consequence of Increases in sea and air temperature and PAR as well as increased storminess it is likely that fucoids will become restricted to more northerly latitudes and sheltered locations with global warming (Ballantine 1961,

Franklin and Forster 1997, Sagarin et al. 1999, Coelho et al. 2000). There is some evidence to suggest that this northwards range movement has already started with some cold-temperate canopy forming species, such as Laminaria spp. (Breeman 1990, Murias Dos Santos 2000),

Himanthalia elongata and Pelvetia canaliculata (Murias Dos Santos 2000), already becoming scarcer at their southern range limits.

If as predicted under climate change scenarios canopy forming macroalgae becomes restricted to more northerly latitudes, it Is likely that this loss will affect P. vu/gata and P. depressa in different ways. If P. vulgata preferentially aggregates beneath Fucus patches for the habitat amelioration provided, it is likely that the loss of Fucus will result in P. vulgata relocating to environments where environmental extremes are ameliorated. At its southern range limit, where

Fucus patches are rare (Ballantine 1961, Murias Dos Santos 2000, Boaventura et al. 2002c) and heat and desiccation stress greater, P. vulgata is limited to damp microhabitats and the low shore

(Guerra and Gaudencio 1986, Lewis 1986, Murias Dos Santos 2000). As a result of the interaction between P. vulgata and Fucus patches it is likely that the loss of either species will modify how the other species responds to climatic warming.

The pverall aim of this chapter was firstly to investigate the reasons for the differential behaviour of P. depressa and P. vulgata in relation to Fucus patches and secondly the consequencesfor these two limpet species should Fucus patches be lost due to climatic warming.

P. vulgata has been shown to relocate home scars to beneath new Fucus patches when they become available (Burrows and Lodge 1950, Hawkins and Hartnoll 1983b, Hartnoll and

Hawkins 1985, Burrows and Hawkins 1998, Roberts 2002), therefore uncolonised Fucus patches were made available to both limpet species. It was hypothesised that if P. depressa creates home scars on open rock solely to avoid interspecfic competition, then if uncolonised Fucus patches became available P. depressa should relocate home scars to beneath Fucus patches in the same proportions as P. vulgata.

32 Chapter Three

Manipulativefield experiments were used to investigatethe response of both P. vulgata and

P. depressa to the loss of Fucus patches. Responseswere quantified in terms of mortality and relocation of home scars. I hypothesisedthat the loss of Fucus would have a much greater effect on P. vulgata compared to P. depressa, by effectively reducing a high quality microhabitat.

3.2 Materials and methods

3.2.1 Behaviour of a northern and southern species of limpet In relation to Fucus patches

To investigate the movement of P. vulgata and P. depressa Into uncolonised Fucus patches experiments were run from March to August 2004 at two locations, Trevone and

Crackington Haven on the north coast of Cornwall, UK (Fig 3.1). Both shores are moderately exposed with areas of open rock, barnacle covered rock and a mosaic of Fucus patches. At both locations the proportion of P. depressa to P. vulgata is approximately equal. To create Fucus patches live individual Fucus plants were removed from the substrate with a scalpel, removing the holdfast and the frond, then attached to PVC plates (10 x 10cm) using cable ties where they continued to survive and grow. Enough individual plants were attached to each plate to create a

Fucus patch of approximately 0.25m2. Fifteen Fucus plates were attached to the rock substrate via stainless steel screws into rawlplugs at mid-shore level on each of the two shores. All limpets located beneath the attached Fucus patches were removed. Fucus plates were examined every two weeks and any limpets that had relocated home scars to beneath the Fucus patches were identified and double tagged with micro-numbers glued to the shell with cyanoacrylate.

A two factor ANOVA was performed to compare the movement of P. vulgata and P. depressa into the uncolonised Fucus patches, with the factor location (two levels) considered random and the factor species (two levels) considered fixed. Heterogeneityof variance was checked using Cochran's test with the data subsequentlyfourth root transformed. In order to increase the power of factors of interest post-hoc pooling was utilised to remove non-significant terms (p>0.25; Winer 1971, Underwood 1997). Student-NewmanKeuls (SNK) post-hoc tests were carried out on significant data (p < 0.05).

33 Chapter Three

3.2.2 Limpet behaviour following the experimental loss of Fucus

Manipulative experiments to investigate the effect of the loss of Fucus on the behaviour of

P. vulgata and P. depressa were carried out at Trevone and Crackington Haven, on the north coast of Cornwall, UK (Fig 3.1). Experiments were conducted between June and September in both

2002 and 2003. At mid-shore level 20 areas with Fucus vesiculosus (hereafter referred to as

Fucus) patches of approximately 900cm2 in size were randomly allocated to two treatments: Fucus present (unmanipulated control) or Fucus removal. In the Fucus removal treatments Fucus canopy, including holdfast, were removed from the substrate with a scalpel. Ten areas of open rock (control) were also selected, creating three treatments. Responses of P. vulgata and P. depressa were monitored independently in each of five replicate plots for each of the three treatments. Five P. vulgata or P. depressa were randomly selected within each experimental plot; these individuals were double tagged with micro-numbers glued to the shell with cyanoacrylate and shell length measured with callipers.

The response of limpets to the removal of Fucus patches was recorded in two ways: distance moved to a new home scar or mortality. To quantify relocation to new home scars the position of each limpet was recorded using coordinates from a1 m2 grid separated into 0.03m2grid squares. Limpet positions prior to treatment manipulationswere taken to be the limpets initial home scars. The position of each individualwas then determined every fourteen days during low water when the limpets were not active (i.e. on their home scars). In cases where limpets had relocated from their home scar, the habitat to which they had moved was noted. Where limpets were missing from plots, a fifteen-minute search of the surroundingarea (approximately9m2) was made and if the limpets were not found they were assumed to have died.

In 2002 a large number of limpets from all treatments died at Trevone, probably as a consequenceof natural sand scour in some of the experimental areas. Therefore the data was unsuitable for formal statistical analysis and analysis in 2002 was only carried out on the data collected for Crackington Haven. In 2003 problems with sand scour did not occur and comparisons were made between Trevone and Crackington Haven.

To examine for temporal consistency the differences between the three treatments were compared with a three factor ANOVA for the two dependant variables, movement of limpets from their home scars and mortality, at Crackington Haven in 2002 and 2003. The factors treatment (three levels) and species (two levels) were considered fixed and the factor year (two levels) was 34 Chapter Three

considered random. To examine for consistency of spatial patterns data for Trevone and

Crackington Haven in 2003 were compared with a three factor ANOVA with the factors treatment

(three levels) and species (two levels) considered fixed and the factor location (two levels) considered random. The number of individual limpets available for analysis of distance moved to a new home scar was unequal due to differential mortality between treatments, and so data were randomly removed from treatments to create a balanced design (Underwood 1997). This resulted in n=12 limpets per treatment being used for data from Crackington Haven in 2002 and 2003 and n=1 7 limpets per treatment being used for data from Crackington Haven and Trevone in 2003. For all analyses heterogeneity of variance was tested using Cochran's test and where appropriate data were either natural log (x + 1) transformed or fourth root transformed. In order to increase the power of factors of interest post-hoc pooling was utilised to remove non-significant terms (p>0.25;

Winer 1971, Underwood 1997). Student-Newman Keuls (SNK) post-hoc tests were carried out on significant data (p < 0.05).

To investigate the relationship between limpet size and the distance moved in finding a new home scar correlation analysis was performed for each species/treatment combination. As the data differed significantly from a normal distribution following transformation Spearman's rank-order correlation was used.

N Crackingtýn Haven

Trevone ,.. -

01. - rinkm

Fig 3.1. Location of study sites on the north coast of Cornwall, UK.

35 ChapterThree

3.3 Results

3.3.1 Behaviour of a northern and southern species of limpet In relation to Fucus patches

At Crackington Haven up to 26 P. vulgata individualswere found to have relocated home scars to beneath a single 0.25m2Fucus patch. In contrast, a maximum of two P. depressa individualswere found to have relocated their home scars to beneath a single Fucus patch. At

Trevone a maximum of twelve P. vulgata individualswere found to have relocated their home scars beneath a single uncolonised Fucus patch, compared to a total of just two P. depressa individuals found to have relocated their home scars to beneath any of the fifteen Fucus patches. Significantly higher numbers of P. vulgata relocated their home scars to beneath Fucus patches comparedto P. depressa at both locations (F 1,5760.85, P<0.01; Table 3.1; Fig 3.2).

Table 3.1. Analysis of variance of the movement of P. vulgata and P. depressa Into uncolonised Fucus patches at Crackington Haven and Trevone. Variances were homogenous following fourth root transformation (Cochran's test C=0.4504, ns). Location x species Interaction was non-significant (P >0.25) and was thus pooled with the residual to increase the power of the test for the factor species.

Source SS DF MS F P F versus Location 1.12 1 1.12 3.64 0.06 1-pooled Species 18.75 1 18.75 60.85 <0.01 1-pooled Location x Species 0.20 1 0.20 0.64 0.43 1-pooled Residual 17.37 56 0.31 Total 37.44 59 1-pool 17.57 57 0.31 SNK test Species P. vulgata > P. depressa

36 ChapterThree

0 12

Cyrn

pt 10 E"

*ýO'dy 8 0. N P. vulgata Ö O. 0 0 P. depressa dd 4 .o0 co äc 2 ýv o ä0

Crackington Haven Trevone

As Figure 3.2. Mean number of limpets relocating their home scars to beneath uncolonised Fucus patches. been results were consistent at both Crackington Haven and Trevone the data from both locations have combined. Error barst 1SE. ** P<0.01.

37 Chapter Three

3.3.2 Limpet behaviour following loss of Fucus patches

Movement

P. vulgata relocated their homes scars more often than P. depressa in response to the experimentalremoval of Fucus patches, moving further from their home scars than in all other treatments. This behaviour was consistent at Crackington Haven in 2002 and 2003 (F 2,14323.58, Table 3.3; Fig 3.3b). P<0.05; Table 3.2; Fig 3.3a) and across locations in 2003 (F 2,1943.4, P<0.05; Fucus The maximum distance an individual P. vulgata moved to relocate their home scar following removal was over two metres away from their original home scar. Of the P. vulgata that moved following Fucus removal 45% from Crackington Haven in 2002 and 2003 and 25% at Trevone and

Crackington Haven in 2003 were'found to have relocated their home scars to beneath another

Fucus patch. In contrast, P. depressa stayed loyal to their home scars in Fucus removal, unmanipulated Fucus patches and open rock treatments at both locations in 2003. However, when comparing the two years at Crackington Haven, P. depressa did move a small but significant distance from their home scars in the unmanipulated Fucus patch (8.5cm ± 1.6cm) and Fucus removal (5.6cm ± 1.4cm) treatments compared to open rock treatments (Table 3.2; Fig 3.3a).

However, all individuals remained within the area of the Fucus patch in which they were initially tagged.

in There was a significant positive correlation between limpet size and distance moved relocating home scars'for P. vulgata in Fucus removal treatments at Crackington Haven in 2002

(r8= 0.53; n =15; P<0.05), with larger limpets moving greater distances in relocating their home scars. Larger P. vulgata individuals in Fucus removal treatments also moved greater distances in relocating home scars at Crackington Haven in 2003, although this proved not to be significant,

(r5=0.28; n =23; P =0.28). There were no significant correlations between limpet size and distance moved for P. vulgata at Trevone or in Fucus patch and open rock treatments at Crackington

Haven. There were no correlations between distance moved in relocating home scars and length for P. depressa in any of the treatments at Trevone or Crackington Haven.

Mortality

P. vulgata experienced higher mortality (approx 40%) in Fucus removal treatments than in unmanipulated Fucus patches (approx 20%) and on open rock (approx 24%). P. vulgata experienced higher mortality than P. depressa in Fucus removal treatments, with P. depressa

38 ChapterThree

experiencingthe same level of mortality irrespectiveof treatment. These patterns were evident for Fig 3.4a) each species at Crackington Haven in 2002 and 2003 (F 2,504.42, P<0.05; Table 3.4; and at both locations in 2003 (F 2,503.75, P<0.05; Table 3.5; Fig 3.4b).

Table 3.2. Analysis of variance of distancethat limpetswere found from their home scar for three treatments: Fucus removal (FR), Fucus patch (FP) and open rock (OR) at CrackingtonHaven in 2002 and 2003. Varianceswere homogenousfollowing Ln (X + 1) transformation(Cochran's test C=0.1141,ns). Due to unequal levels of mortality n=12 limpets were used per treatment.

Source SS DF MS F P F versus Year (Ye) 3.04 1 3.04 0.67 0.42 Residual Treatment (Tr) 35.46 2 17.73 4.44 0.18 Ye x Tr Species (Sp) 6.00 1 6.00 332.59 0.03 Ye x Sp Year x Treatment 7.98 2 3.99 0.88 0.42 Residual Year x Species 0.02 1 0.02 0.00 0.95 Residual Sp Treatment x Species 17.07 22 8.54 23.58 0.04 Ye x Tr x Ye x Tr x Sp 0.72 2 0.036 0.08 0.92 Residual Residual 601.00 132 4.55 Total 671.32 143 SNK tests Treatment x Species Species x Treatment P. vulgata FR>FP=OR Fucus removal P. vulgata > P. depressa P.depressa FR=FP>OR Fucus patch P. vulgata = P. depressa Open rock P. vulgata = P. depressa

Table 3.3. Analysis of variance of distancethat limpetswere found from their home scar for three treatments: in 2003. Fucus removal (FR), Fucus patch (FP) and open rock (OR) at Trevone and CrackingtonHaven Location Varianceswere homogenousfollowing Ln (X + 1) transformation(Cochran's test C=0.1496,ns). x to increasethe habitat x species interactionwas non-significant(P >0.25) and was pooled with the residual limpets power of the test for treatment x species interaction. Due to unequal levels of mortality n=17 per treatment.

Source SS DF MS F P F versus Location (Lo) 9.45 1 9.45 8.68 0.00 1-Pooled Treatment (Tr) 4.33 2 2.17 0.89 0.53 Lo X Tr Species (Sp) 7.39 1 7.39 35.55 0.11 Lo x Sp Location x Treatment 4.86 2 2.43 2.23 0.11 1-Pooled Location x Species 0.21 1 0.21 0.19 0.66 1-Pooled Treatment x Species 7.39 2 3.7 3.4 0.04 1-Pooled Lo x Tr x Sp 0.73 2 0.37 0.34 0.72 1-Pooled Residual -210.5 192 1.1 Total 244.86 203 1-Pooled 211.23 194 1.09 SNK Tests Treatment x Species Species x Treatment P. vulgata FR>FP=OR Fucus removal P. vulgata > P. depressa P. depressa FR=FP=OR Fucus patch P. vulgata = P. depressa Open rock P. vulgata = P. depressa

39 ChapterThree

Table 3.4. Analysis of variance of limpet mortality in three treatments:Fucus removal (FR), Fucus patch (FP) and open rock (OR) at CrackingtonHaven in 2002 and 2003. Cochran'stest C=0.1948,ns. Year x treatment x species interactionwas non-significant(P>0.25) and was pooledwith the residualto increasethe powerof the test for the treatmentx species interaction.

Source SS DF MS F P F versus Year (Ye) 4.27 1 4.27 6.70 0.01 1-pooled Treatment (Tr) 4.43 2 2.22 19.00 0.05 Ye x Tr Species (Sp) 1.07 1 1.07 4.00 0.30 Ye x Sp Year x Treatment 0.23 2 0.12 0.18 0.83 1-pooled Year x Species 0.27 1 0.27 0.42 0.52 1-pooled Treatment x Species 5.63 2 2.82 4.42 0.02 1-pooled Ye x Tr x Sp 1.03 2 0.52 0.81 0.45 1-pooled Residual 30.80 48 0.64 Total 47.73 59 1-pooled 31.83 50 0.64 SNK tests Treatment x Species interaction Species x Treatment Interaction P. vulgata FR>FP=OR Fucus removal P. vulgata > P. depressa P. P. depressa P. depressa FR=FP=OR " Fucus patch vulgata = Open rock P. vulgata = P. depressa

(FP) Table 3.5. Analysis of varianceof limpet mortalityin three treatments:Fucus removal (FR), Fucus patch and open rock (OR) at Trevone and CrackingtonHaven in 2003. Cochran'stest C=0.1932,ns. Location x the treatment x species interactionwas non-significant(P>0.25) and was pooled with the residualto increase power of the test for the treatment x species interaction.

Source SS DF MS F P F versus Location (Lo) 1.67 1 1.67 2.30 0.13 1-pooled Treatment (Tr) 3.23 2 1.62 1.59 0.39 Lo x Tr Species (Sp) 3.27 1 3.27 49.00 0.09 Lo x Sp Location x Treatment 2.03 2 1.02 1.40 0.26 1-pooled Location x Species 0.07 1 0.07 0.09 0.77 1-pooled Treatment x Species 5.43 2 2.72 3.75 0.03 1-pooled Lo x Tr x Sp 1.03 2 0.52 0.71 0.50 1-pooled Residual 35.20 48 0.73 Total 51.93 59 1-pooled 36.23 50 0.72 SNK tests Treatment x Species interaction Species x Treatment interaction P. vulgata FR>FP=OR Fucus removal P. vulgata > P. depressa P. depressa FR=FP=OR Fucus patch P. vulgata = P. depressa Open rock P. vulgata = P. depressa

40 Chapter Three

a)

T ** T 250

200 E E 150 vd ß 100 N 0 50

0 P. vulgata P. depressa

Fucus removal p Fucus patch (unmanipulated) p Open rock

b)

**

180 r- 160 140 E 120 100 c 80 ea N 60 40 20 0 P. vulgata P. depressa

®Fucus removal 0Fucus patch (unmanipulated) [3 Open rock

Fucus Figure 3.3. Mean distance moved from home scar for P. depressa and P. vulgata for three treatments: removal (FR), Fucus patch (FR) and Open rock (OR). Error bars ±I SE. a) Means for Treatment x species for interaction at Crackington Haven for 2002 and 2003 as pattern was consistent among years. b) Means Treatment x species interaction at Trevone and Crackington Haven in 2003 as pattern was consistent among sites. ** P<0.01; ns=not significant.

41 ChapterThree

a) CD

EE. ßL4 °.'

ßE 01

E0 'CL P. vulgata P. depressa m..

Fucus removal p Fucus patch (unmanipulated) p Open rock

b) .

m U Ec. I

'ý E2 o'n Ec1

Eü0 P. depressa CI P. vulgata I- Fucus removal p Fucus patch (unmanipulated) Q Open rock

Figure 3.4. Limpet mortality In three treatments:Fucus removal (FR), Fucus patch (FP) and Open rock (OR). a) Means for Treatment x species interactionat CrackingtonHaven for 2002 and 2003 as pattern was in consistentamong years. b) Means for Treatment x species Interactionat Trevone and CrackingtonHaven 2003 as pattern was consistentamong sites. ** P<0.01; ns = not significant

42 ChapterThree

3.4 Discussion

3.4.1 Behaviour of a northern and southern species of limpet In relation to Fucus patches

Large numbers of P. vulgata individuals relocated home scars to beneath newly available

Fucus patches. These results support previous work on P. vulgata, which have shown that this species will relocate their home scars to beneath Fucus patches when they become available

(Hawkins 1983, Hartnoll and Hawkins 1985, Roberts 2002). In contrast, very few P. depressa individuals relocated their home scars to beneath uncolonised Fucus patches with a total of only two being found under any of the fifteen Fucus patches at Trevone. If as hypothesised, P. depressa creates home scars on open rock, where limpet densities are lower, in order to reduce interspecific competition with P. vulgata, it would be expected that P. depressa would relocate their home scars to beneath the uncolonised Fucus patches at the same rate as P. vulgata. These results would suggest P. depressa avoids Fucus patches even when the patches are uncolonised by con-specifics or P. vulgata. Therefore P. depressa seems to avoid creating home scars beneath Fucus patches for reasons other than competition with P. vulgata.

In chapter two I gave a number of suggestionsto explain why P. depressa does not preferentially aggregate beneath Fucus patches including, that this southern/lusitanian species may simply not need the habitat amelioration provided by Fucus patches, or that P. depressa actually requires the increased temperatures experienced on open rock compared to beneath

Fucus patches for physiological reasons. Significantly more P. depressa are found on warmer been south facing rock than warmer north facing rock (Hawkins, unpub. ). Alternatively, it has also in suggested that Fucus may have a negative effect on P. depressa as this species was slow Torrey recolonising Fucus covered rock compared to scoured rock without Fucus following the It Canyon oil spill in 1967 (Southward and Southward 1978, Hawkins and Southward 1992). may

be that the chemical defences that Fucus possesses such as phlorotanins and galactolipids, which

are readily extruded and abundant in Fucus exudates (Lobban and Harrison 1997), and are known

to restrict herbivore grazing, may deter P. depressa individuals from creating home scars beneath

Fucus patches (Deal et al. 2003, Jormalainen et al. 2003). The reasons for P. depressa not

aggregating beneath Fucus patches cannot be explained by data presented here, but it would appear that P. depressa does not relocate home scars to beneath Fucus patches even when they

are uncolonised by P. vulgata.

43 ChapterThree

Experimentalremoval of Fucus led to increased P. vulgata mortality probably as a

consequence of greater thermal stress and desiccation associated with the removal of this

important habitat. Temperatures beneath Fucus patches in the summer can be up to 5°C cooler

than on open rock, while the relative humidity beneath Fucus patches can be up to 20% higher

than on open rock (Wilson and Hawkins, unpub). Similar subtle changes In temperature have been

shown to influence the survival of other intertidal species such as barnacles (Bertness et al. 1999,

Leonard 2000). Removal of Fucus also resulted in increased relocation of P. vulgata to new home

scars with individuals moving to new Fucus patches. Hence, the local persistence of P. vulgata is

enhanced presumably by the habitat amelioration Fucus patches provide. In contrast, removal of

Fucus appears to have little effect on the behaviour or mortality of the southern/Iusitanian

congeneric species, P. depressa.

3.4.2 Biotic interactions modify species responses to climate change

The direct ameliorationof environmentalconditions by organisms such as canopy forming

algae will become more important as warming increases, and species of limpets, littorinids,whelks

and anemones will increasinglybecome restrictedto habitats where climatic extremes are buffered.

Habitat amelioration also increases the persistenceof other marine species including the northern

barnacle, Semibalanusbalanoides, whose survivorship was increased by shading from macroalgae

at hotter southern sites, but not at cooler northern sites (Bertness et al. 1999, Leonard 2000). Furthermore high-level salt marsh plants experience facilitative effects of neighbouring plants at

hotter southern sites but not at cooler northern sites (Bertness and Ewanchuk 2002). Therefore if warming continues as predicted, the local persistence of many species may become more

dependent on the presence of others and shifts in species distributions may be magnified where

biologically generated habitats, such as algal canopies, are lost.

Under 'climate envelope' predictionsthe abundance of P. depressa at sites in Britain

should increase with temperature (Fig 3.4a) while the abundance of P. vulgata and Fucus should

decrease as temperature rises (Fig 3.4a & b). However, when the interaction with Fucus is

incorporated a stepped response in the relative abundance of these limpet species is likely to occur

at the edge of the distributional boundary of Fucus (Fig 3.4c). The abundance of P. vulgata decreasing sharply as a consequence of increased environmental stress resulting from the loss of habitat amelioration from Fucus. in contrast, the abundance of P. depressa is likely to rapidly 44 Chapter Three

increase as it fills niches vacated by P. vulgata. This switch in the relative abundance of P. vulgata and P. depressa is apparent when the proportionof the two limpet species is compared between shores in south-west Britain (mean proportionof P. depressa 0.30; 50° 19'N, 4° 07'W; Jenkins et at.

2001), where Fucus can be abundant, and shores in northern Spain (mean proportion of P. depressa 0.81; 43 °34'N, 6° 11'W; Jenkins et al. 2001), where Fucus becomes rare on open coasts

(Fischer-Piette1955, Ballantine 1961, Murias Dos Santos 2000, Boaventuraet al. 2002c).

Therefore synergistic indirect effects may modify the direct effects of climate change through positive and/or negative interactionsand feedback. For example, the rate at which species with northern distributions are replaced by those with southern distributionsmay be modulated by changes in the abundance of habitat forming species which amelioratethe environment. The extent and quality of habitat will decrease for some species and increase for others. Hence, without the consideration of biotic interactions, predictions of species range shifts are likely to be inaccurate.

If canopy forming algal abundancereduces on southwest shores, perhaps as a result of climate change, it is likely that P. vulgata will become restricted to damp microhabitats, as currently occurs at its range edge in southern Portugal (Guerra and Gaudencio 1986). Survival of younger more vulnerable stages will also be reduced in the absence of Fucus canopy. In contrast, loss of

Fucus may increase the relative abundance of P. depressa. Similar increases in the abundance of species with southern biogeographic ranges in areas vacated by species with northern biogeographic ranges has been predicted for other marine and terrestrial species (Sagarin et al.

1999, Berry et al. 2002). Alternatively if Fucus is able to tolerate the effects of climate change, as a result of the environmental amelioration provided by Fucus patches, P. vulgata may be able to persist where otherwise it would not.

3.4.2.2 Community level responses to local changes in grazer distribution

The grazing behaviour of P. vulgata contributes to the conspicuous mosaic of variously aged macroalgal patches common on many moderately exposed shores of the NE Atlantic (Lewis

1964, Hawkins 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992), which in turn influencesthe community structure of rocky shores as a higher diversity of species is found beneath Fucus patches compared to areas of adjacent emergent rock (Thompson et al. 1996, Jenkins et at.

1999a). P. vulgata plays a key role in structuring rocky shore communities because of its 45 ChapterThree

aggregative behaviour,which can lead to a patchy distribution of grazing Intensity enabling Fucus

escapes to occur (Hartnoll and Hawkins 1985). My data shows that P. depressa does not

preferentiallyaggregate beneath Fucus patches, resulting In a more even distribution of grazing activity and probably an associated reduction in the probability of Fucus escapes, compoundingthe reduction in Fucus canopy caused by climatic warming. Hence changes in the spatial distribution of grazers, as a response to climate change, could have broad-scale implicationsfor community dynamics on rocky shores. Differencesin the grazing activity of P. depressa and P. vulgate will be investigatedfurther (chapter 4) to see if both species have a similar role in controlling the abundance of macroalgae on the shore.

These results not only show that habitat amelioratingspecies can have an important role in the local persistence of some species. More importantly they show that, in response to climate change, the relative balance of species with northern and southern biogeographic distributions may be amplified by changes in the abundance of habitat ameliorating species such as fucoid algae. I also demonstrate that species, which are thought to be very similar in terms of their physiology and ecology can have subtle differences in their behaviour, which may fundamentally alter community dynamics in benthic habitats. Without taking into consideration interactions such as these, predictions of species range shifts, across both temporal and spatial scales, are likely to be inaccurate.

46 ChapterThree

a)

limpet abundance / ' / Ü I I C Co I

C I M Q 00 I 00 00 0000 00

Direct effects - b) P. vulgata P. depressa 'climate

ýýu+t je. nk, in d., ni r. envelope'

U (U ß

C)

Interactive effect of Fucus on limpet abundance .' - 00 Ua) . C Indirect ß 00 .- effects - ß 00 C i - . biotic Q r interactions "I "

Increasingtemperature

Fig 3.4. Changesin the abundanceof a) two species of limpet (P. vulgata a northern/borealspecies and P. depressa a southern/lusitanianspecies) and b) Fucus vesiculosuswith increasingtemperature c) the interactiveeffect of Fucus vesiculosusand temperatureon the abundanceof P. vulgata and P. depressa.

47 Chapter Four

Chapter 4

The effects of temperature and gonad ripening on the grazing activity of P.

vulgata and P. depressa

4.1 Introduction

P. On shores where both P. vulgata and P. depressa co-exist it has been assumed that depressa exerts a similar role in controlling macroalgal abundance because of similarities between these two species radula structure (Hawkins et al. 1989), ecology and behaviour (Thompson et al.

2000, Jenkins et al. 2001, Boaventura et al. 2002a, Boaventura et al. 2002b). In previous chapters

have shown that P. depressa and P. vulgata have different spatial distributions in the intertidal and this is driven by their behaviour in relation to Fucus (Chapter 2). P. vulgata preferentially aggregates beneath Fucus patches while P. depressa avoids them (Chapter 2). In addition the loss of Fucus results in mortality and relocation of home scars for P. vulgata, but has little effect on

P. depressa (Chapter 3). The relative abundance of these two species is expected to alter with P. global warming. The abundance of P. depressa is predicted to increase and the abundance of vulgata is expected to decrease (Southward et al. 1995). If the proportion of P. depressa increases, for example, in response to climatic warming, the chance of Fucus escapes occurring This are likely to be reduced due to the more even spatial distribution of limpet grazing activity. will have broad-scale implications for rocky shore community structure in the NE Atlantic.

Both limpet species are generalist grazers consuming microbial films made up of bacteria, diatoms, cyanobacteria, protozoa and juvenile stages of macroalgae embedded in a muco- has polysaccharide matrix (Wahl 1989, Hill and Hawkins 1991). Gut content analysis of P. vulgata, shown that in addition to elements of the microbial film, including juvenile stages of macroalgae, large items such as barnacle plates, calcareous algae and small gastropods were also present finer (Hawkins et al. 1989). The diet of P. depressa while very similar to P. vulgata; had a much gut matrix and contained no large items (Hawkins et al. 1989). However, many other factors also influence the feeding activity of marine invertebrates including body size, reproductive cycle, the

presence of feeding or digestive rhythms as well as environmental factors such as temperature, tidal level and exposure (Orton et al. 1956, Newell et al. 1971, Boyden and Zeldis 1979, Newell

1979, Jenkins et al. 2001). In gastropod molluscs, radula activity has been used as an index of feeding intensity with increases in temperature being linked to increased rasping rates for Littorina

48 Chapter Four

littorea (Newell et al. 1971) and the patellid limpet Cellana ornata (Boyden and Zeldis 1979).

Although adult P. vulgata and P. depressa are able to tolerate temperatures ranging from <0°C to

42°C (Evans 1948), there are likely to be physiological differences between the species enabling one species to graze more effectively than the other at different temperatures or during different stages of their reproductive cycle. In limpets for example, there Is evidence that In the later stages of gonad development feeding may be reduced or even cease as pressure is exerted on the surrounding tissue by the developing gonad (Orton et al. 1956). Jenkins et al. (2001) demonstrated that P. vulgata reduces its grazing activity in late summer/early autumn just prior to spawning followed by an increase in grazing directly after spawning. Following the peak spawning period grazing increased by about 25% based on estimates obtained using Jenkins et al. (2001).

The reproductivecycle of P. vulgata and P. depressa has been well studied in Britain and

Portugal (Orton et al. 1956, Orton and Southward 1961, Bowman and Lewis 1977,1986, Guerra and Gaudencio 1986). P. vulgata is an autumn/winter breeder in Britain although the time of peak spawning differs across a latitudinal gradient (Orton et al. 1956, Bowman and Lewis 1977,1986).

In northern Scotland the main spawning period is August/September (Bowman and Lewis 1986), in southwest Britain it is November (Orton et al. 1956, Bowman and Lewis 1986) and in northern

Portugal between November and January (Guerra and Gaudencio 1986). It is presumed that the latitudinal gradient in peak spawning activity is related to climatic variables, which result in populations releasing their gametes so that recruits settle on the shore during favourable conditions

(Bowman and Lewis 1986). P. vulgata has a single reproductively active period, although there may be more than one spawning event (Bowman and Lewis 1986).

P. depressa spawns from May to November with peak spawning occurring in July/August in southwest Britain (Ballantine 1961b, Orton and Southward 1961, Bowman and Lewis 1986). A latitudinal gradient in reproductiveactivity of P. depressa has been demonstratedwith populations in northern Portugal spawning in late autumn and late spring followed by resting periods. In southern Portugal there is a long summer resting period followed by development and an extended period of reproductive activity over the winter months (Guerra and Gaudenclo 1986). It is thought that the long resting period during the summer months in southern Portugal is a response to the adverse temperatures experienced in southern Portugal during the summer. P. depressa may spawn a number of times with re-ripening of the gonads between spawning events (Bowman and Lewis 1986, Guerra and Gaudencio 1986), although re-ripening was not found to be common in 49 Chapter Four

southwest Britain (Bowman and Lewis 1986, Roberts 2002). Although, redevelopmentof gonads may not be common towards the northern range limits of P. depressa, this species has full gonads four months prior to P. vulgata (Ballantine 1961b, Roberts 2002). With acceleratingclimatic warming it is likely that redevelopment of P. depressa gonads will become more common In southwest Britain. A drop in grazing activity as a result of gonad development has so far only been demonstrated in P. vulgata. If P. depressa shows a similar decline in grazing just prior to spawning, then these two species of limpet could have different effects on the chance of Fucus escapes because P. depressa grazing activity will decline at different times to P. vulgata. The reduction in grazing activity may occur more often if there is re-ripening of P. depressa gonads.

Because the timing of reproduction is so different for these two species, the impact of grazing cycles on juvenile Fucus and other macroalagae is likely to vary depending on which species Is the numerically dominant grazer. Thus, any changes in the proportion of these species may have unanticipated secondary effects on the distribution of macroalgae and community composition as a whole in the rocky intertidal of the NE Atlantic.

Studies of limpet grazing activity, using wax discs inserted Into the shore (Thompsonet al.

1997, Jenkins et al. 2001), have shown that there is a temporal pattern in grazing activity on NE

Atlantic shores (Thompson et al. 2000, Jenkins and Hartnoll 2001). In Britain low levels of grazing occurred in late winter (January/February)and increased over the rest of the year to a summer peak (Thompson et at. 2000, Jenkins et al. 2001). In southern Portugal the grazing intensity of limpets increased through the summer to a peak in late autumn, followed by a gradual decline in grazing activity into early spring (Jenkins et al. 2001). In general there was higher grazing activity throughout the year with decreasing latitude, with mean monthly grazing of wax discs in the Isle of

Man of 26% compared to 42% in southwest Britain and 66% in southern Portugal (Jenkins et at.

2001). It has been suggested that an increase in grazing activity with decreasing latitude may be a function of the increasing abundance and diversity of grazers with decreasing latitude (Ballantine

1961a, Hawkins et al. 1992, Jenkins et al. 2001), rather than an increase in grazing activity related to temperature (Jenkins and Hartnoll 2001). Although these studies quantified the grazing behaviour of limpets and other grazers over broad temporal and spatial scales, no attempt has been made to separate the grazing activity of P. vulgata and P. depressa populations,and the role of each in preventing Fucus escapes is unknown.

50 Chapter Four

The aim of this chapter was to separately quantify the grazing activity of P. depressa and

P. vulgata. Differences in the two limpet species grazing activity were compared to each other over a twenty month period and also related to physical and biological parameters such as temperature and gonad cycle. Specifically I examinedthe hypotheses that there would be differences in the grazing activity of P. vulgata and P. depressaas a result of:

1. Differences in their time of reproduction;

2. Differences in their responses to temperature.

Although I anticipated that there would be differences in the temporal pattern of grazing activity between the two limpet species I hypothesisedthat P. vulgata and P. depressa would both have a role in controlling the abundance of macroalgae on the shore.

4.2 Materials and methods

Temporal variation in the grazing activity of P. vulgata and P. depressa in relation to gonad cycle and temperaturewas compared from August 2003 to March 2005 at two locations: Kingsand

(50° 33'N, 40 19W) and Andurn Point (50° 32'N, 4° 12'W), on the southwest coast of Britain (Fig

4.1). Both locations are moderately exposedwith areas of open and barnacle covered rock. Three treatmentswere established in circular enclosures (50cm diameter) at mid-shore height on free draining open rock, uncolonised by macroalgae:P. vulgata only, P. depressa only and both species present. Enclosureswere approximately5cm high and made of plastic coated garden mesh

(13mm mesh size) fastened to the substratum with rawlplugs and stainless steel screws. Five replicates were randomly assigned to each treatment.

In each enclosure an equal biomass of limpets (1.38g ± 0.09), was created using a predeterminedempirical relationship between length and biomass of individual limpets for shores on which the experiment was run. There was no significant difference in the abundance of limpets in the three treatments enclosures at either Andurn Point or Kingsand (appendix 1).

Limpet grazing activity was measured as the percentage area scraped on discs of dental wax embedded in pre-drilled holes on the shore. These discs become scratched by the limpets radula during foraging excursions providing an index of grazing intensity (Thompson et al. 1997). Eighteen holes were drilled in each enclosure with nine holes randomly selected for wax discs to be deployed for two weeks of every month, thereby ensuring independence with time (Thompson 1997). Grazing et al. intensity was calculated by examining wax discs in the laboratory using a 51 Chapter Four

binocular microscope,following dusting With printer toner to highlight the grazing marks. Grazing activity on each disc was calculated for each array by placing a circular piece of acetate with twenty-five randomly placed holes onto the wax discs to calculate percentagearea of wax grazed.

Within each enclosure the length of all limpets were measured and double tagged with micronumbersattached to the shell with cyanoacrylate. Any missing limpets were replacedwith individualsof the same size and species from outside the enclosures.

The reproductive cycle of the two limpet species was described by collecting fifty individuals of each species (15 to 35mm size range), every two weeks from outside treatment enclosures at the two locations. Limpets were dissected in the laboratory and gonad development assessed according to the classification scheme of Orton et at. (1956). The proportion of limpets in advanced states of gonad development (stages 4 and 5 in Orton et al. 1956 classification scheme) was then calculated for both P. vulgata and P. depressa for each of the two locations.

Temperature data loggers (Thermochron®Button DS1921G)were deployed on each of the two shores to take air and sea temperaturesat fifteen minute intervals. Temperature readings were separated with the aid of tide timetables into emersed temperature readings, to provide air temperaturesand immersed temperaturereadings, to provide seawater temperatures. Maximum and minimum monthly sea and air temperatureswere calculated for each location.

On each sampling occasion any individual macroalgal germlings present were counted for each enclosure. Following the growth of Fucus vesiculosus (hereafter Fucus) in some experimental enclosures the percentage cover of Fucus was calculated for each enclosure using the point intercept method. To compare the growth of Fucus within the experimentalenclosures with the rest of the shore the percentagecover of Fucus was estimated at mid-shore height using twenty haphazard quadrats also using the point intercept method.

At the completion of the experiment all limpets within the enclosures were removed and returned to the laboratory where they were measured to compare the growth rate among treatments. Limpets were dissected and the state of gonad development recorded to investigate differences between limpets that had been enclosed and those from outside enclosures. The limpet gonad index (Orton et al. 1956) was calculated as few limpets of either species had gonads in advanced states of development and therefore using proportion of limpets in advanced states of development was not applicable.

52 Chapter Four

Differences in grazing activity between the three treatments were compared for each time period using Friedman's matched set test. The effect of limpet grazing intensity on the probability of Fucus escapes was compared using a two factor ANOVA with location (two levels) considered a random factor and treatment (three levels) considered fixed. Due to the Increased number of replicates (n=20) used to quantify the percentage cover of Fucus across the whole shore a student- t test was performed to compare shore Fucus cover with that found within P. depressa only treatments (n=5) at both locations. As a result of the growth of Fucus in some treatments the effects of Fudus presence on grazing intensity was analysed using a four factor nested ANOVA with location (two levels) and date (four levels) nested within presence of Fucus considered random factors and treatment (three levels) and presence of Fucus (two levels) considered fixed factors. For all ANOVAs heterogeneity of variance was tested using Cochran's test and where appropriate data were transformed. Student-Newman Keuls (SNK) post-hoc tests were carried out on significant data (p < 0.05).

To compare the growth of limpets in the three treatments over the course of the experiment, Ford-Walfordplots were created and regressions calculated for P. depressa only treatments, P. vulgata only treatments, P. depressa in both species treatments and P. vulgata in both species treatments. ANCOVA was used to compare the slopes of the regression lines to see if there were differences in the growth of the two species among the three treatments.

To investigatethe relationship between temperature and grazing activity and grazing activity and the proportion of limpets in advanced gonad states correlation analysis was performed.

Pearson's correlation was used to investigate the relationship between grazing activity and temperature. Spearman's rank-order correlation was performed to investigate the relationship between the proportion of the limpet population with gonads in advanced states of development and grazing activity as the data did not meet the assumptions of Pearson's correlation analysis following transformation. To control for Type I error rates in multiple testing situations significance levels were corrected using the sequential Bonferroni procedure (Quinn and Keough 2002).

53 Four

M N

1'ý' ''1 ýf 1 i.

Kingsand " Andurn Point 1ý ýi 50km

Fig 4.1. Location of study sites in southwest Britain

4.3 Results 4.3.1 Grazing activity

P. vulgata grazed more than P. depressa over the course of the experiment at both locations (Fig 4.2). Grazing activity was, however, greater at Kingsand in treatments with both species present and P. vulgata only treatments (Fig 4.2). Although there were differences in the area of wax disc scraped among treatments the pattern of grazing was consistent at Andurn Point for the duration of the experiment and was consistent at Kingsand until June 2004 (Fig 4.2). At

Kingsand after June 2004 treatments with P. vulgata only and with both species present continued to follow the same pattern of grazing activity. In contrast, P. depressa only treatments did not show the same pattern and grazing remained generally low (< 20% of the area of wax disc scraped; Fig

4.2). For all treatments, at both locations, the greatest amount of grazing occurred in late autumn and winter (Fig 4.2).

At both Kingsand and Andurn Point grazing was significantly less in P. depressa only treatments than Individuals in P. vulgata only treatments (F 2,1223.92, P<0.01; Table 4.1; Fig 4.3). in P. depressa only treatments grazed significantly less than the other two treatments (P. vulgata only and both species together) on 14 out of 20 sampling occasions at Kingsand. These differences were apparent from January 2004 to April 2004 and from June 2004 to February 2005.

At Andurn Point grazing by limpets in P. vulgata only and both species present treatments was more than P. depressa over the course of the experiment, although this difference was only significant for 4 out of 20 sampling occasions. Individuals in P. depressa only treatments grazed

54 Chapter Four

significantly less than those in the other two treatments from February 2004 to April 2004 and in

March 2005. During December 2003 and 2004 and June 2004 grazing in P. depressa only treatments was less than in the other two treatments, but insignificant at the 0.05 level (P<0.07).

During the course of the experiment there was no statistical difference between the amount of grazing that occurred in treatments with both species present and treatments containing P. vulgata

These therefore that P. depressa only (F 2,1223.92, P<0.01; Table 4.1; Fig 4.3). results show grazes less than P. vulgata over time and that this pattern was consistent between locations.

for Table 4.1 Analysis of variance of limpet grazing activity at Andurn Point and Kingsand, southwest Britain three treatments: P. vulgata only, P. depressa only and both species present (Cochran's test C=0.0734, ns). help Significant results are indicated > or <. The direction of non-significant trends are also shown -+ or t- to illustrate overall patterns in the data.

Source SS DF MS F P F versus Location (Lo) 5275.00 1 5275.00 14.72 <0.01 Lo x Da[Fu] Treatment jr) 23595.35 2 11797.67 0.00 NO TEST Fucus present/absent(Fu) 14901.88 1 14901.88 0.00 NO TEST Date [Fu] (Da[Fu]) 4323.77 6 720.63 2.01 0.21 Lo x Da[Fu] Lo x Tr 5984.01 2 2992.00 42.39 <0.01 Lo x Tr x Da[Fu] Lo x Fu 420.19 1 420.19 1.17 0.32 Lox Da[Fu] Lo x Da[Fu] 2149.83 6 358.30 2.77 <0.05 Residual Tr x Fu 8453.26 2 4226.63 0.00 NO TEST Tr x Da[Fu] 1894.11 12 157.84 2.24 0.09 Lox Tr x Da[Fu] Lo x Tr x Fu 3377.44 2 1688.72 23.92 <0.01 Lo X Tr x Da[Fu] Lo x Tr x Da[Fu] 847.07 12 70.59 0.55 0.88 Residual Residual 24828.64 192 129.32 Total 96050.55 239 SNK tests Treatment (Location x Fucus presence) Andurn Fucus absent P. vulgata > P. depressa P. vulgata -º Both- P. depressa Fucus present P. vulgata = Both > P. depressa Kingsand Fucus absent P. vulgata = Both > P. depressa Fucus present P. vulgata = Both > P. depressa

Fucus (Location x Treatment) Andurn Both Fucus absent< Fucus present P. depressa Fucus absent< Fucus present P. vulgata Fucus absent< Fucus present Kingsand Both Fucus absent< Fucus present P. depressa Fucus absent> Fucus present P. vulgata Fucus absent< Fucus present

Location (Treatment x Fucus presence) Both Fucus absent and present Andurn < Kingsand P. vulgata Fucus absent and present Andurn < Kingsand P. depressa Fucus absent Andurn -º Kingsand Fucus absent Andurn > Kingsand

55 Chapter Four

4.3.2 Fucus growth

By September 2003 juvenile Fucus was present in all of the P. depressa only treatments at both locations, with some treatments where both species were present also having juvenile Fucus within the enclosures (Fig 4.4a & c). The Fucus in these enclosures continued to grow throughout the remainder of the experiment in some of the enclosures at Kingsand covering up to 90% of the enclosures. Statistical analysis of differences amongst treatments in the percentage cover of Table 4.2; Fig Fucus revealed a significant location x treatment interaction (F 2,245.57, P<0.01;

4.5). At both locations the percentage cover of Fucus in P. depressa only treatments was greater than the percentage cover of Fucus in either of the other two treatments. The percentage cover of to Fucus in P. depressa only treatments was, however, significantly higher at Kingsand compared

Andurn Point. There was no difference between the percentage cover of Fucus in P. vulgata only the treatments and treatments with both species present. The percentage cover of Fucus across shore as a whole at Kingsand (mean 1.48% ± 0.54%) was considerably lower than the percentage df 23; cover of Fucus in the P. depressa only enclosures (mean 74% ± 5.10%; t= 27.76, P<0.01, = 2.3%) Fig 4.5). The same trend was also evident at Andurn Point (whole shore mean 4.87% ± compared to within P. depressa only enclosures (mean 14.4% ± 6.77%). However, this proved not to be statistically significant (t = 1.69, P=0.10, df = 23; Fig 4.5).

The increase in the percentagecover of Fucus in P. depressa only treatments, particularly While the at Kingsand, may have had a large impact on the grazing behaviour of this species. percentage cover of algae was low the grazing behaviour of P. depressa at both locations although less than in the other two treatments did follow the same general temporal pattern of grazing in intensity. Once the percentage cover of Fucus reached over 40% at Kingsand, grazing activity these enclosures decreased and remained low for the remainder of the experiment (< 20% of wax disc area grazed; Fig 4.2). Statistical analysis of the effect of Fucus on the mean percentage area

of wax disc scraped revealed a significant location x treatment x Fucus interaction (F 2,1223.92,

P<0.01; Table 4.1). At Kingsand, P. depressa only treatments grazed significantly less once Fucus

cover in these treatments reached 40% (April 2004). In contrast, limpets in P. vulgata only

treatments and treatments with both species present at both locations and P. depressa only

treatments at Andurn Point grazed significantly more after April 2004. It would therefore appear

that Fucus cover inhibits the grazing activity of P. depressa once it reaches a certain density.

56 Chapter Four

Table 4.2 Analysis of variance of percentage Fucus cover for three treatments: P. vulgata only, P. depressa only and both species present, sixteen months after the initiation of the experiment. Variances were homogenous following fourth root transformation (Cochran's test C=0.4369, ns). Source SS DF MS FPF versus Location (Lo) 1.58 1 1.58 9.76 <0.01 Residual Treatment (Tr) 32.32 2 16.16 17.92 0.05 Lo x Tr Location x Treatment 1.80 2 0.90 5.57 <0.01 Residual Residual 3.88 24 0.16 Total 39.59 29 SNK tests Location x Treatment interaction Treatment x Location interaction P. vulgata Andurn = Kingsand Andurn P. depressa > P. vulgata = Both P. depressa Andurn < Kingsand Kingsand P. depressa > P. vulgata = Both Both Andurn = Kingsnad

4.3.3 Reproductive cycles

The time, length and proportion of limpets in advanced gonad states differed between species. P. vulgata was reproductively active from October to the beginning of January with a peak in spawning activity in November (Fig 4.6). During the spawning period the proportion of being fully individuals with gonads in advanced states was low with less than 20% of the population reproductively active during 2003/2004 and less than 10% during 2004/2005 (Fig 4.6).

P. depressa had gonads in advancedstates of developmentfrom the beginning of the experiment in August 2003 to the beginning of October 2003. This was followed by a resting from period from November 2003 to the end of February'2004. P. depressa began to spawn again the P. the end of February to August in 2004. There were two distinct peaks in the proportion of in June 2004, depressa population with gonads in advanced states one during April and a second that then a third more reduced peak occurred at Andurn Point at the end of July 2004, suggesting 80% there were several spawning events followed by gonad re-ripening. At both locations over of the September in population had gonads in advanced states of development at the beginning of during 2003 and between 60-80% of the population had gonads in advanced states of development the April and June peaks in 2004. At Andurn Point a small proportion, less than 10% of the population, had gonads in advanced states of development from the beginning of February 2005

(Fig 4.6).

The state of gonad development between the single species enclosures and enclosures with both species present were similar. This was also the case when gonad development between

limpets from enclosures and those randomly selected from outside enclosures were compared at

the end of the experiment (Table 4.3).

57 Chapter Four

Table 4.3 Gonad index of limpets from three treatments: P. depressa only, P. vulgata only and both species present as well as individuals collected randomly across the shore at the conclusion of the experiment in March 2005 from Kingsand and Andurn Point, southwest Britain.

Single species Both species Outside treatments treatments enclosures Kingsand P. vulgata 1.25 n=68 1.21 n=42 1.06 n=49 P. depressa 2.52 n=61 2.79 n=34 2.38 n=50

Andurn Point P. vulgata 1.81 n=36 1.61 n=25 1.74 n=52 P. depressa 2.30 n=54 2.96 n=23 2.40 n=54

4.3.4 Limpet growth

There was little relationship between growth increment and length with regressions for all the treatments having low r2 values and the regression lines for P. depressa only and P. vulgata from both species together treatments not being statistically significant (Fig 4.7). ANCOVA showed

no significant differences between any of the regression lines (F 3,3430.78, P=0.50). There were therefore no differences in the growth rate of either P. vulgata or P. depressa in the various

treatments.

4.3.5 Temperature, Grazing and Reproduction

To investigatethe effects of temperatureon limpet grazing activity, grazing and temperature data were combined at the two locations for P. vulgata only treatments and treatments with both species present, due to the similarities in grazing intensity between locations. For P.

depressa only treatments the two locations were kept separate due to large differences in grazing

patterns. There were no significant correlations between any of the temperature parameters and

grazing activity in P. vulgata only, P. depressa only and both species together treatments (Table

4.4).

The proportion of the population in advanced gonad states had a significant negative effect P Table 4.4). on grazing activity in P. depressa only treatments at Andurn Point (r$ = -0.44, <0.05;

In contrast, P. depressa at Kingsand seemed to have a weak positive relationship between the

proportion of the population with gonads in advanced states and grazing activity (Table 4.4).

However, this may relate to this species low grazing activity from July to the completion of the

experiment, which may be linked to the increased percentage cover of Fucus in the P. depressa

only treatments (see page 56). At both locations, P. depressa scraped less area of wax discs

58 Chapter Four

during April and June 2004, correspondingwith peaks in this species spawning activity (Fig 4.2 &

4.6).

There was no significant correlation between the proportion of the population with advanced gonad states and grazing activity for P. vu/gata only treatments at either location (Table

4.4). This lack of correlation between grazing activity and spawning may be a result of the small percentage of the P. vulgata population that were spawning at any one time. Therefore it is unlikely that grazing was reduced in P. vu/gata only treatments because developing gonads were exerting pressure on other organs.

Table 4.4. Pearson's correlations between limpet grazing activity in three treatment enclosures: P. vulgata only, P. depressa only and both species present with temperature. To control for Type I error rates resulting from multiple tests, significance levels were corrected using the sequential Bonferroni procedure. Spearmans rank correlation was performed on the proportion of the limpet population in advanced states of gonad development and grazing activity. Data from the two locations have been combined for P. vulgata only and both species together treatments as patterns of grazing activity were similar. * P<0.05.

Max Min Max Air Min Air Gonads sea temp sea temp P. vu!gata 0.04 (n=37) Grazing (n=32) -0.12 0.25 -0.27 0.14 P-value 0.58 0.17 0.14 0.49 0.83

P. depressa (n=20) Andurn Grazing (n=15) 0.05 0.43 -0.13 0.32 -0.44 P-value 0.92 0.12 0.67 0.26 0.05* 0.38 (n=17) Kingsandgrazing (n=17) -0.12 -0.10 -0.28 -0.16 P-value 0.69 0.74 0.29 0.56 0.14

Both Grazing (n=32) -0.10 0.29 -0.28 0.16 P-value 0.63 0.11 0.15 0.41

59 Chapter Four

a) Kingsand

x 100 co ö 80 ca aý 'v dß 60 ý'. Qf

40 d N

CL 20 "x---l 0ý ' IE "*., K. "i"I..... I. 0 AS0NDJFMAMJJAS0NDJF Month

b) Andurn Point

100 4-0 ß 80 a aý ß °1 00 60

U' 40 ax 20

5d

ASONDJFMAMJJAS0NDJF Month --ý- P. vulgata -- -a---P. depressa -- -i- -" Both

Fig 4.2. Mean percentagearea of wax discs scraped for three treatments:P. vulgata only, P. depressaonly and both species present for a) Kingsandand b) Andurn Point. Error bars ±1 SE. Arrow indicatesthe time when Fucus cover reached40% in P. depressaonly treatments at Kingsand.

60 Chapter Four

® Both o P. vulgata 0 P. deprassa 0 LO 100 r **"I .10 80 0 60

CI) 40

20 CL CO ß d before after before after

Andurn Point Kingsand

P. depressa and Fig 4.3. Mean percentagearea of wax disc scraped in three treatments:P. vu/gataonly, only P. depressa both species present at Andurn Point and Kingsandbefore Fucus cover reached40% in only Error bars SE. * P<0.05, enclosuresat Kingsandand after Fucus cover reached40% cover at Kingsand. ±1 ** P<0.01, ns = not significant.

61 Chapter Four

Kingsand Andurn Point

a) P. depressa only

ý ,,ý

ýý: ,;

b) P. vulgata only

at Fig 4.4. Fucus in three treatments: a) P. depressa only, b) P. vulgata only, c) both species present growth in Kingsand and Andurn Point. Pictures taken 15 months after initiation of experiment. Wax discs placed pre- drilled holes to quantify limpet grazing activity are also visible in the treatment enclosures.

62 Chapter Four

Andurn Point Kingsand **ý 100 y V LL 80 v- 0 a. > O 60 V

D1 40 w C

CL 20 C ca m 20 Sh Pd Pv Bo Sh Pd Pv Bo

Fig 4.5. Mean percentage cover of Fucus across the whole shore (Sh) and in three treatments: P. depressa only (Pd), P. vulgate only (Pv) and both species present (Bo) for Kingsand and Andurn Point. Error bars t ISE. ** P<0.01, ns = not significant.

63 Chapter Four

a) Kingsand

V u C ß . 0.8 CyCo 0.6 aý Eý w. c 0.4 o0 cM :I0.2 0 a o0 a ASONDJFMAMJJASONDJF Month

b) Andurn Point

-" o" - P. vulaata V1 dV c 0.8 ß N 43 0.6 Cid ýº E öö0.4 cm c 0 0.2 öa 2 CL 0 ASONDJFMAMJJASONDJF Month

Fig 4.6. The proportionof P. vulgata and P. depressain advancedgonad states (gonad stages 4&5 after Orton et al. 1956)for a) Kingsandand b) Andurn Point.

64 Chapter Four

14

12

vv 10 v° Both Pv 00 v E vv E v" vm 8 " .ý c Pv ng vi CD - ýv f clý v v; E 6 ý ý Both Pd ýý. c ""

0 t9 2

ö vp Qv 0 0

10 15 20 25 30 35

Initial Length (mm)

" P. vulgata only; y=-0.2182x+8.8855; r2=0.17; P<0.01; n=110 o P. depressa only; y=-0.1058x+5.2291; r2=0.03; P>0.05; n=117 f P. depressa both; y=-0.169x+7.0367; r2=0.08; P<0.05; n=60 o P. vulgata both; y=-0.254x+11.328: r2=0.05; P>0.05; n=64

Fig 4.7. Ford-Walford plots of growth increment plotted against initial shell length for P. depressa and P. vulgata from three treatments: P. depressa only; P. vulgata only and both species present. P. vulgata from treatments with both species present (Both Pv); P. vulgata only treatments (Pv); P. depressa from treatments with both species present (Both Pd) and P. depressa only treatments (Pd). ANCOVA indicates no significant difference between the slopes of the regressions for each species-treatment plot (F 3,3430.78; P>0.05).

65 Chapter Four

4.4 Discussion

4.4.1 Comparisons of the grazing activity of two limpet species with different

biogeographic ranges

Previous experiments have quantified the grazing intensity of limpets in Britain (Thompson et al. 2000, Jenkins et al. 2001) and across a broader European scale (Jenkins et al. 2001). These studies however, did not separate the relative importance of grazing activity on shores where there was more than one species present (e. g. southwest Britain Thompson et al. 2000, Jenkins et al.

2001 and Spain Jenkins et al. 2001). My current work has separated the grazing of P. vulgata and

P. depressa and shown that there were differences in their grazing activity. Although not always statistically significant, P. vulgata grazed more than P. depressa at both locations on all sampling occasions. This trend was particularly apparent during April 2004 and June 2004 when P. vulgata scraped significantly more than P. depressa, coinciding with peaks in the proportion of the P. depressa population with gonads in advanced states of development.

Interestinglythere was generally no difference in the percentagearea of wax disc scraped in treatments with both species present and treatments of P. vulgata only at both locations and over the course of the experiment. This could indicate one of two things, either P. depressa grazed more in treatments where P. vulgata was also present or that P. vulgata in treatments where both species were present grazed a greater area of wax disc than in single species treatments and therefore compensated for P. depressa not grazing as much. Previous work has shown that food may be a limiting resource for grazers, particularly in the summer (Thompson et al. 2000). If P. vu/gata in treatments with both species present did increase their grazing intensity then it is likely that food may be a limiting resource not only in the summer, but throughout the year on shores in southwest Britain. Unfortunately the relative grazing intensity of the two species in treatments with both species present cannot be separated, therefore it is impossible to confirm this hypothesis from the data collected here.

The grazing activity of limpets has been shown to follow a seasonal pattern with the lowest amount of grazing, in Britain, occurring in late winter (January and February),followed by a steady rise in grazing activity to a peak in early summer (Thompson et al. 2000, Jenkins et al. 2001).

Although there were differences in the amount of grazing occurring between treatments, the pattern of grazing was very similar between treatments at both locations through the course of the experiment. The exception to this was in P. depressa only treatments at Kingsand from June 2004 66 Chapter Four

to the end of the experiment, where grazing activity became consistently low, perhaps due to the

effect of Fucus growth, which is discussed more fully below. In contrast to previous work the

seasonal pattern of grazing activity in this work was lowest highest in late autumn, a pattern more similar to the grazing of limpets in Portugal than in southwest Britain (Jenkins et al. 2001).

The temporal pattern of grazing shown by limpets in Britain has been shown to positively correlate with sea (Thompson et al. 2000, Jenkins et al. 2001) and air temperature (Thompsonet al. 2000). However, Jenkins et al. (2001) found no significant relationship between limpet grazing activity and sea temperature in northern Spain. My results also showed no correlation between grazing activity and either sea or air temperature indicating that factors other than temperaturehad more of an influence on the grazing activity of limpets.

4.4.2 Phenology and climate change

Investigationsinto the breeding cycle of P. depressa in southwest Britain in the late 1940s and late 1950s revealed that the gonad of P. depressa began to fill in April and by the middle of

May reached a gonad state at which spawning could occur (Ballantine 1961 b, Orton and

Southward 1961). Spawning continued until July, with most of the population having 'spent' gonads by August/September (Ballantine 1961b, Orton and Southward 1961). More recent work, carried out in 1997 showed a similar pattern of gonad development by P. depressa (Roberts 2002).

The proportion of limpets in advanced states of gonad development (ie able to spawn) increased from April to peak in June to August (Roberts 2002). In the present study P. depressa had gonads in advanced states of development from the end of February to September. Although sampling during winter and early spring in the 1940s was carried out sporadically, only a single individual was found with a gonad in an advanced state of development in February 1946 and none were found in March 1949 on the north coast of Cornwall in southwest Britain (Orton and Southward

1961). On shores close to those used in the current work there were no P. depressa individuals in advanced states of gonad development until April 1997 (Roberts 2002). It would therefore appear that the period in which P. depressa is spawning has extended considerably compared to previous studies, with spawning occurring over a period of six months.

P. depressa can experience gonad redevelopmentresulting in multiple spawning peaks (Bowman and Lewis 1986, Guerra and Gaudencio 1986). Although gonad redevelopmentis known to occur in populations from southwest Britain, it is thought to be rare (Bowman and Lewis 67 Chapter Four

1986, Roberts 2002) and in recent work this phenomenon was not observed (Roberts 2002). In the present work there was evidence of gonad redevelopment and re-spawning during 2004, with two distinctive spawning peaks at Kingsand and three distinctive spawning peaks at Andurn Point.

There was some evidence that redevelopment may also have occurred in 2003 due to a dramatic in rise in the proportion of the population with gonads in advanced states of development

September compared to August. Although P. depressa continues to spawn during the summer months, lengthening of the reproductive season and the multiple redevelopment of gonads and re- P. spawning that occurred during 2003 and 2004 was more similar to the reproductive cycle of depressa further south in its range (Guerra and Gaudencio 1986).

For P. vulgata the proportion of the population in advanced states of gonad development the reached a peak during late November in both 2003 and 2004, with spawning continuing until P. beginning of January. My results are similar to those obtained from previous studies on vulgata 1986, gonad cycles in southwest Britain (Orton et al. 1956, Ballantine 1961b, Bowman and Lewis

Roberts 2002). The difference between my work and previous studies was the much lower percentage of the P. vu/gata population in advanced states of gonad development. In the current in study less than 20% of the population in 2003/2004 and less than 10% of the population in 2004/2005 reached gonad states where spawning could occur. In contrast studies carried out the 1940s found between 80 and 100% of the population had gonads in advanced states of 1956). In development on the north coast of Cornwall, southwest Britain (Orton et al. northern 1980s. During Portugal the reproductive cycle of P. vu/gata was followed over three years in the this time P. vu/gata experienced two years where 100% of the population reached advanced states in of gonad development and one year where only 40% of the population had gonads advanced states of development (Guerra and Gaudencio 1986). It would therefore appear at the southern range limit of P. vulgata, where environmental conditions are harsher, P. vulgata occasionally in experiences poor reproductive years. As a result of increased warming P. vulgata populations

southwest Britain may also be experiencing adverse environmental conditions which are affecting

their reproductive success.

Towards the south of its range P. vulgata becomes more restricted to sheltered shores and

estuaries where environmentalconditions are ameliorated (Ballantine 1961a, Murias Dos Santos

2000, Boaventura et al. 2002b, Hawkins, unpub). If climatic warming increases as predicted it is

likely that P. vulgata will become more restricted to sheltered shores in southwest Britain. Recently 68 Chapter Four

the reproductiveoutput of P. vulgata has been shown to be greater on sheltered shores compared to exposed shores as a result of a greater abundance of larger individuals being found on sheltered shores (Le Quesne, pers. comm.). Populationsof P. vulgata that continue to survive on more exposed shores are likely to experience more reproductive failures, resulting in fragmentation of reproductively viable populations and a reliance on the recruitment of juveniles from sheltered populations.

If, in response to increased climatic warming, the length of time that P. depressa is reproductivelyactive increases with continued multiple spawning events and the number of years where P. vulgata experiences reproductivefailure increases it is likely that the relative abundance of these two limpet species will alter faster than anticipatedfrom adult physiologicalperformance.

4.4.3 Impacts of differences in limpet grazing activity for community structure

In the British Isles there is a dynamic balance between the relative abundance of grazers and fucoid algae (Southward and Southward 1978, Hawkins and Hartnoll 1983, Hawkins et al.

1992). Limpets and particularly P. vulgata have been shown to control the abundance of macroalgae on shores of the NE Atlantic through their grazing behaviour (Jones 1948, Burrows and Lodge 1950, Southward and Southward 1978, Hawkins and Hartnoll 1983, Hawkins et al.

1992, Boaventura et al. 2002a, Jenkins et al. 2005). Due to similarities in the feeding mechanisms and diet of P. vulgata and P. depressa (Hawkins et al. 1989) it has been assumed that they have the same role in controlling macroalgal abundance on rocky shores of the NE Atlantic. This assumption is not consistent with the results from this work where Fucus vesiculosus germlings began to appear in September 2003 in P. depressa enclosures, but not in P. vulgata enclosures. It is thought that the reduction in grazing observed in P. depressa during the peak of its reproductive season in the spring and summer allowed juvenile Fucus settling at that time (Knight and Parke

1950) to survive long enough to reach a size where they escaped limpet grazing. P. vulgata, which does not reproduce during the time of Fucus recruitment prevented the establishment and growth of Fucus juveniles in treatments where it was present either on its own or in combination with P. depressa. Thus, despite similar grazing preferences P. depressa appears to be far less effective than P. vulgata at controlling the abundance of macroalgae on shores in southwest Britain.

The percentage cover of Fucus was significantly more in P. depressa only treatments at Kingsand (60-90% cover) compared to Andurn Point (maximum of 40% cover). Recent 69 Chapter Four

manipulative experiments, which experimentally removed limpets to reduce grazing pressure, resulted in macroalgal growth on shores in southwest Britain (Jenkins et al. 2005). However, this response was weak and variable on the scale of patches within shores (1Os to 100s of metres) and location (kilometres) compared to higher latitudes where Fucus developed quickly and dominated all limpet exclusion plots (Jenkins et al. 2005). This variability could be a result of differences in the supply of macroalgal propagules (Arrontes 2002), mortality of germlings in unfavourable microclimates (Brawley and Johnson 1991) or dislodgement of propagules by wave action (Vadas et at. 1990). Of these hypotheses Jenkins et at. (2005) proposed that the variability in macroalgal growth between locations was probably a result of differences in the supply of propagules, as the dense adult Fucus stands are often rare on shores in southwest Britain compared to shores on

Isle of Man. The two locations in the present study had little Fucus cover outside of P. depressa only enclosures with less than 2% at Kingsand and less than 5% cover at Andurn Point, suggesting from that the supply of algal propagules to the two locations would be variable and reliant on supply other locations.

The reduction in grazing activity in P. depressa only treatments at Kingsand coincidedwith the percentage cover of Fucus in treatment enclosures reaching over 40%. Fucus in this treatment continued to grow resulting in canopy cover of between 60 - 90% by March 2005, with grazing remaining low (< 20% of wax disc area scraped) for the remainder of the experiment. This P. reduction in grazing activity for over 8 months of the experiment probably explains why smaller depressa individuals grew less than individualsfrom other treatments. It has been suggested that

P. depressa does not prosper beneath Fucus (Hawkins and Southward 1992) and these results does would appear to support that hypothesis. Although the proximate reason for why P. depressa not prosper beneath Fucus cannot be explained, this data provides further evidence that P. depressa does not prosper beneath Fucus patches, resulting in this species preferentiallycreating home scars on open rock.

On coasts of western Europe there is a latitudinal trend in the balance between fucoid algae and limpet/barnacle dominated shores (Ballantine 1961a). Fucoids are more abundant at northern latitudes, but are replaced by an increased abundance and diversity of grazers at more southerly latitudes (Ballantine 1961 a). This switch from fucoid dominated shores to grazer and barnacle dominated shores has been suggested to be a result of changes in the effectiveness of grazers and the ability of fucoids to grow with decreasing latitude (Ballantine 1961 a, Hawkins and 70 Chapter Four

Hartnoll 1983, Hawkins et al. 1992). Fucoid algae become Increasingly restricted to sheltered and estuarine shores from northern Spain, yet P. depressa is the dominant limpet grazer on northern

Spanish shores where the only other limpet species is P. vulgata (Jenkins et al. 2001). It would therefore appear that fucoid algae may become restricted to sheltered and estuarine environments either purely because of environmental conditions; or because P. depressa is a more effective grazer further south in its range than in southwest Britain. However, there Is an increase in the number of other grazers, including topshells, crabs and fish, present on Spanish shores.

Therefore, at more southern latitudes individual grazers may not be as effective as grazer species at higher latitudes, but the increase in the diversity and abundance of grazers may be able to control macroalgal abundance. To distinguish between these hypotheses further manipulations of grazer diversity and abundance would have to undertaken across a broader latitudinal area.

In conclusion this chapter has shown that P. depressa grazes less than P. vulgata over time at different shores in southwest Britain. This resulted in Fucus settling and growing in P. depressa treatments, which had a negative effect on the growth of juvenile P. depressa. These results indicate that P. depressa has less of a role in controlling the abundance of macroalgaeon shores in southwest Britain than P. vulgata. If, as predicted, the relative abundance of P. vulgata and P. depressa alters in response to increasedclimatic warming the role of limpets in controlling macroalgal abundance is likely to alter.

71 Chapter Five

Chapter Five

Intra- and interspecific interactions during settlement and recruitment In a

northern and southern species of barnacle

5.1 Introduction

The study of plant and animal communities on rocky shores has been invaluable in understanding the interplay between physical factors and biotic interactions in structuring natural assemblages (Connell 1961a, Dayton 1971, Connell 1972). Barnacles have proved a particularly tractable model for such studies, including pioneering experimental work on intra and interspecific factors, competition (Connell 1961 b, a). On the shores of southwest Britain the effects of physical especially temperature, on the interactions between and the relative abundance of species of barnacles has been studied for nearly seventy years (Moore and Kitching 1939, Southward and

Crisp 1954, Southward and Orton 1954, Southward and Crisp 1956, Southward 1967,1991,

Southward et al. 1995). The main space-occupying barnacles on shores of southwest Britain are

Semibalanus balanoides, formally Ba/anus balanoides, a northern/boreal species with a distribution from the Arctic Circle to the Iberian Peninsula (Fischer-Piette and Prenant 1956,1957, Barnes

1958) and Chthamalus montagui and Chthamalus stellatus, two south ern/Iusiantian species, with a Crisp biogeographic range from the Shetland Isles, Scotland to northern Africa (Southward 1967, et al. 1981, Kendall et al. 1985, Lewis 1986, Southward et al. 1995). C. montagui and C. stellatus were formerly considered one species Chthamalus stellatus (Southward 1976). On some shores these three species are joined, albeit in smaller numbers by the introduced Australasian barnacle

Elminus modestus, although this species is predominantly limited to more sheltered and estuarine conditions (Crisp 1958, Southward 1991).

Semibalanusbalanoides (hereafter referred to as Semibalanus)and Chthamalusspp.

(hereafter referred to as Chthamalus) have both been shown to respond to changes in the climate with shifts in their relative abundance (Southward and Crisp 1956, Southward 1967,1991,

Southward et al. 1995). The relative abundance of Chthamalus and Semibalanus in southwest

Britain were first examined at a selection of sites in 1934 (Moore 1936). Repeat surveys of these sites, undertaken in the 1950s, showed that there had been a significant increase in both the abundance and distribution of Chthamalus with a corresponding decrease in the abundance of

Semibalanus (Southward and Crisp 1954, Southward and Crisp 1956). This switch in relative

72 Chapter Five

abundance was also evident in Scotland near the northern range limit of Chthamalus (Southward and Crisp 1956), and indicated that the change in the relative abundance of the two taxa was broad-scale. The increase in relative abundance of Chthamalus and the associated decline of

Semibalanus was attributed to a warming of 0.25°C during the 1950s compared to the 1930s taxa (Southward 1960). It was hypothesised that the switch in the relative abundance of these two was broadly related to changes in climatic conditions (Southward and Crisp 1956). to Further work has shown that the relative abundance of these species has continued increased fluctuate with fluctuations in air and sea surface temperature. For example Chthamalus in in abundance during the 1950s, a climatically warm period, while Semibalanus increased Southward abundance during the cool period of the 1960s and 70s (Southward and Crisp 1954, in and Crisp 1956, Southward 1967,1991). Since the mid 1980s Chthamalus has increased fluctuations in abundance with a corresponding decrease in Semibalanus (Southward 1991). The in the relative abundance of these two barnacle taxa have been shown to follow oscillations sea and air temperatures with a two-year phase lag (Southward 1967), with sea temperature the two accounting for 40% of the variation (Southward 1991). The two-year lag in the response of (Southward barnacle taxa has been attributed to the time it takes for individuals to reach maturity

1991, Southward et al. 1995). Therefore, temperature may only be operating on the maturation larvae and reproduction of individuals that produce the settling larvae rather than on the in themselves or on later juvenile stages (Southward 1967). Initially the fluctuations the relative 11 abundance of the two barnacle taxa appeared to correlate to changes in the 10 to year sun spot

cycle (Southward 1991) reflecting the flux of solar energy to the earth, resulting in warmer and

cooler climatic periods. However, it has now been shown that switches in the relative abundance

of the two taxa have deviated from the sun spot cycle and show considerable year-to-year

variation, combined with an overall upward trend in the relative abundance of Chthamalus in (Southward 1991, Southward et al. 1995). Southward et al. (1995) suggested that the switches

the relative abundance of these barnacle species no longer correlated with the sunspot cycle

because anthropogenic climate change was having a greater effect on their population dynamics.

The current view is that a dynamic equilibrium exists between the two taxa with both

temperature and other environmental factors influencing their relative abundances. These factors

act to alter the strength of competitive interactions of each taxa, primarily for space (Southward

1967,1991, Southward et al. 1995). Semibalanus is generally thought to be the dominant 73 Chapter Five

competitor over Chthamalus (Connell 1961 b). The rapid growth rate of Semibalanus juveniles results in them undercutting and overgrowing Chthamalus Individuals (Connell 1961b, Wethey

1983, Wethey 1984b), with Chthamalus individuals up to two years old being smothered by settling

Semibalanus cyprids and spat (Connell 1961b). Hence, on Semibalanus dominated shores

Chthamalus is restricted to high shore levels where Semibabalus suffers desiccation-related mortality (Connell 1961b, Wethey 1983, Wethey 1984b). However, these studies were undertaken at the northern range limit for Chthamalus where it suffers much higher physical stress. In particular the work by Connell (1961b) was carried out on shores with exceptionally high levels of

Semibalanus recruitment, as a result of high larval retention, compared to most shores in the

British Isles (Hawkins and Hartnoll 1982, Hansson et al. 2003), therefore competition for space would be expected to be particularly strong. On shores of southwest Britain, where Semibalanus is near its southern range limit, this species is still generally thought to be the dominant competitor during cooler climatic conditions out competing Chthamalus, and restricting this species to the high shore (Southward 1967). During warmer climatic conditions, however, Chthamalus dominates the intertidal, presumably by occupying free space which has become available as a result of the reduction in the number of settlers and survivors of Semibalanus (Southward and Crisp 1956,

Southward 1967,1991, Southward et al. 1995). Interspecific competition is therefore reduced because the abundance of Semibalanus is so low. Unpublished analysis of the southwest barnacle data set has indicated that the negative correlation of Semibalanus to temperature is due to sensitivity of settling early stages to high spring temperatures (Hawkins, Southward, Burrows &

Leaper, unpub). The positive relationship between temperature and Chthamalus is probably an indirect effect, primarily a result of release from competition from the competitively superior

Semibalanus. Thus space can be occupied, possibly assisted by greater numbers of more successful broods of Chthamalus in warmer years (Burrows et al. 1992, O'Riordan et al. 2001).

As early as 1954 Semibalanusand Chthamaluswere proposed as indicator organisms to investigate species responses to changes in the general environment as both taxa were likely to be sensitive to even small changes in-temperature (Southward and Crisp 1954). As anthropogenic warming of the climate accelerates, monitoring fluctuations in the abundance of these two taxa has been proposed as a cost-effective index of changes in other shallow water marine biota (Southward et al. 1995). With increased warming it is likely that the abundance of Semibalanus will decline either as a result of temperature related mortality or as a result of a reduction in the number 74 Chapter Five

of larvae being able to settle due to a reduced adult population following an increase In the number years where Semibalanus settlement 'fails' (Svensson et al. 2005). At very low densities Allee effects will make this situation worse (Kent et al. 2003). In contrast, the abundance of Chthamalus is expected to increase as a result of reduced competition with Semibabalnusallowing more individualsto survive to reproduce. Increasedtemperature will also enable Chthamalusto produce more broods of larvae.

In addition to responding to changing climatic conditions in contrasting ways the timing of reproduction and recruitment of juveniles differs in these these two taxa. Semibalanus reproduces once a year with the cyprids settling on southwest shores from early March to mid April through to the end of May with peak settlement at the mid to the end of April (Southward 1976, Jenkins et al.

2000). In contrast, Chthamalus cyprids settle from July to September (Southward 1976, Burrows

1988). The earlier settlement by Semibalanus is likely to pre-empt space for Chthamalus settlers later in the year (Southward et al. 1995).

The aims of this chapter were to experimentally simulate climatically driven fluctuations in the population structure of these taxa to inform predictions on the likely outcomes of warming in southwest Britain. Firstly I examinedthe effects of Semibalanusadult abundance on the settlement success of Semibalanuscyprids and spat (section 5.1.1). Secondly I investigatedthe effects of good and poor recruitment of Semibalanuson the settlement success of Chthamalus cyprids and spat to test the idea that interspecific competition between these taxa is a limiting factor for Chthamalusrecruitment and survival on shores in Britain (section 5.1.2).

5.1.1 The effect of adult abundance of Semibalanus balanoides, on the recruitment

success of its gregarious settling young

Barnacle cyprids are known to explore their potential settlement surface for a period of time prior to settling, if the site is unsuitable they are able to detach themselves from the substrate and explore further for a more suitable site (Knight-Jones 1953, Le Tourneux and Bourget 1988).

Cyprid settlement behaviour may be influenced by both physical and biological cues such as available space (Jeffery 2000), shore microtopography (Crisp and Barnes 1954, Wethey 1984a,

Yule and Walker 1984, Wethey 1986, Anderson and Underwood 1994), microbial films (Le Tourneux and Bourget 1988, Keough and Raimondi 1995, Thompson et al. 1998), the presence of juvenile adult or con-specifics (Knight-Jones 1953, Barnett and Crisp 1979, Miron et al. 1996, 75 Chapter Five

Thompson et al. 1998) or cues from con-specifics (Knight-Jones1953, Crisp 1961, Crisp and

Meadows 1963, Larman and Gabbott 1975, Rittschof et al. 1984, Yule and Walker 1984, Hills et al.

1998).

Recent work has shown that where adult barnacle abundance is relatively low and juvenile

settlers are gregarious, relying on cues from con-specifics, there may be a reduction in the

recruitment of juvenile barnacles (Jeffery 2000). As a consequence of this behaviour, low adult

numbers can have important implications for both the population structure and dynamics of the

species involved (Kent et al. 2003). Semibalanus cyprids are known to settle gregariously with

con-specifics (Knight-Jones 1953, Barnett and Crisp 1979, Thompson et al. 1998). Therefore

during climatically warm periods when the relative abundance of Semibalanus would be expected

to decrease (Southward and Crisp 1956, Southward 1967,1991, Southward et al. 1995), the

chance of good settlement occurring in subsequent years may be effected, irrespective of the

numbers of cyprids present in the plankton.

I investigated the intraspecfic' interaction between Semibalanus adults and cyprids to

monitor how changes in the abundance of adult Semibalanus individuals affect the recruitment

success of juvenile Semibalanus. Using settlement plates, adult Semibalanus were manipulated to

simulate warm (bad recruitment) and cool (good recruitment) periods. I hypothesised that high

numbers of adult Semibalanus would encourage higher recruitment of Semibalanus juveniles

compared to settlement plates with low numbers of adult Semibalanus. The presence of barnacle

cues have been shown to encourage barnacle settlement success (Knight-Jones 1953, Crisp 1961,

Crisp and Meadows 1963, Larman and Gabbott 1975, Rittschof et al. 1984, Yule and Walker 1984,

Hills et al. 1998), therefore to maximise treatment effects and to investigate if settlement factor did

enhance Semibalanus recruitment, extracts of Semibalanus (barnacle settlement factor) were

added to half the settlement plates.

5.1.2 The effects of poor 0+ Semibalanus recruitment on the subsequent recruitment

success of Chthamalus spat

The earlier settlement of Semibalanuscould be expected to pre-empt space required by

the later settling Chthamalus (Moore 1936, Moore and Kitching 1939), especially during years with

favourable conditions for Semibalanussettlement and recruitment. Assuming the dynamic

equilibrium hypothesis proposed by Southward and co-workers, which suggested that a decrease 76 Chapter Five

in the relative abundance of one species may be needed before the relative abundance of the other species could increase (Southward 1967,1991, Southward et al. 1995), the abundance of

Semibalanus would have to decrease before the abundance of Chthamalus could increase. This pattern was evident during the cool period in the 1960s when a drop in the relative abundance of

Chthamalus was observed with a corresponding increase in the relative abundance of in Semibalanus the following spring (Southward 1967). Chthamalus should be able to increase abundance when above average temperatures in the summer and autumn, encourage high

Chthamalus recruitment (Southward and Crisp 1956, Kendall et al. 1985), increase Semibalanus juvenile mortality and sometimes kill adult Semiba/anus on the upper shore, creating increased space for Chthamalus recruits (Connell 1961 b, Foster 1971, Kendall et al. 1985). With repeated lower poor settlement and/or'survival of Semibalanus, Chthamalus should become established at

shore levels where it is unaffected by interspecific competition due to the reduced abundance of

Semibalanus (Kendall et al. 1985).

I experimentally manipulated the abundance of Semibalanus 0+ recruits to simulate

conditions representing good and poor Semibalanussettlement years and monitored the

consequencefor recruitment success of Chthamalus. I hypothesisedthat freeing up space through

manipulating Semibalanus0+ numbers to represent poor settlement years should result in

increased recruitment of Chthamaius. In treatments where densities were manipulatedto

represent good Semibalanussettlement years there should be reduced recruitment of Chthamalus.

5.2 Materials and Methods

5.2.1 The effect of adult abundance of Semibalanus balanoides, on the recruitment

success of its gregarious settling young

To assess the role of different densities of con-specifics on the settlement success of

Semibalanus, perspex settlement plates (10 x1 0cm) were covered with clear marine anti-slip tape

(Heskins, Chorley, UK), which acted as a rough surface to encourage cyprid settlement and were

then placed in seawater tanks for a period of 14days to allow a biofilm to grow.

Individual Semibalanuswere carefully removed with a scalpel from rocks collected at St

Ives, North Cornwall (501122'N 50 47'W) and attached to the settlement plates by glueing the outer

edge of the rostrum to the rock with small spots of cyanoacrylate (Thompsonet al. 1998).

Barnacles were attached randomly in the central 5cm x 5cm of the settlement plates to reduce 77 Chapter Five

edge effects. Settlement plates were then placed into seawater tanks for 24 hours to check survival during transplantation (ie barnacles were observed to be feeding normally). In the low density treatments any individuals that had not survived the transplantationwere replaced. High density treatments were only used if 80% of individuals had survived transplantation;if less than

80% of individuals survived, dead barnacleswere replaced.

Seven treatments were subsequentlycreated to formally test the hypothesis that increaseddensity of adult con-specificswould result in increased recruitment of Semibalanusspat:

1. Plate control - Perspex plate with anti-slip tape (1984), the 2. Barnacle settlement factor - following the methods of Rittschof et al except

supernatantwas made up just before deployment and not frozen (Hills et al. 1998;

appendix 2).

3. Low density adult con-specifics -1 cm'2

4. High density adult con-specifics- approximately5 cm"2

5. Low density adult con-specifics(1 cm-2)with barnacle settlement factor

6. High density adult con-specifics(approx 5 cm'2) with barnacle settlement factor

7. Control - natural rock late Five replicates of each treatment were randomly interspersedat four locations during

April and early May 2003. Due to natural differences and fluctuations in the recruitment of

93'W) barnacles two locations were used on the north coast of Cornwall, Trevone (500 55'N, 4° and

Newquay (50° 42'N, 50 10'W) and two were used on the south coast of Devon and Cornwall, Looe

(500 35'N, 40 45'W) and Cellar Beach (50° 31'N 40 06'W; Fig. 5.1) to reduce the chance of no

settlement occurring during the course of the study. Shores on the north coast of Cornwall are

known to experience high recruitment of Semibalanus, whereas on the south coast it is usually

lower (Hawkins, unpub observations). Plates were attached at mid-shore height to smooth

surfaces on north or east facing rocks by screwing the plates with stainless steel screws into

rawlplugs. Rock control treatments were marked by screws in rawlplugs. Due to the large distance

between locations it was impractical to visit locations every day to establish total settlement.

Therefore digital images (2272x1704 pixels) were taken of each treatment at the end of May when

Semibalanus settlement had ceased enabling total recruitment to be quantified. The number of

recruits were counted on enlarged digital images on a computer screen.

78 Chapter Five

Cellar Beach was not included in formal statistical analysis as Semibalanussettlement was very low (0.1 cm2) at this site in 2003 (Moore, pers ob; Hawkins, pers ob), as well as on other shores in the Plymouth area (Hawkins, pers ob). Due to the unequal loss of settlement plates at the remaining locations each location was analysed separately using a two factor ANOVA with barnacle density (three levels: high, low, no adult conspecifics)and barnacle settlement factor (two levels: present or absent) both considered fixed. Heterogeneityof variance was tested using

Cochran's test and where appropriate data were natural log (x + 1) transformed. Student-Newman

Keuls (SNK) post-hoc tests were carried out on significant data (p < 0.05).

5.2.2 The effects of poor 0+ Semibalanus recruitment on the subsequent recruitment

success of Chthamalus spat

Areas of natural rock substrate (5cm x 5cm) were located with 80% barnacle cover and with '0' class Semibalanus cover of between 4-5 cm2 at mid-shore height, where interspecific competition is greatest between the two barnacle taxa, using a random stratified sampling methodology. Areas meeting these requirementswere marked with stainless screws in rawlplugs to at the corners of the plots. Four treatments,each replicated five times, were randomly allocated areas representing conditions of poor Semibalanusrecruitment (1 cm'2),good Semibalanus recruitment (5 cm-2),cleared areas to observe the level of recruitment in the absence of other barnacles and control areas (unmanipulated)were prepared at two locations on the north coast of

Cornwall, Trevone and Crackington Haven (500 68'N, 4° 70'W) and at two locations on the south coast of Devon and Cornwall, Looe and Cellar Beach (Fig. 5.1).

The high and low density treatments were created by removing '0' class individual 2003 to Semibalanus from the rock surface with a scalpel between the 10th-16th July prior day to Chthamalus settlement in late July and August. It was impractical to visit locations every establish total settlement therefore digital images (2272x1704 pixels) were taken of each treatment at the beginning of the settlement season and repeated after settlement had finished in the middle of September enabling total recruitment to be quantified. Chthamalus settlers were counted on enlarged images on a computer screen.

Due to a dense growth of the blue-green algae, Rivularia, obscuring the barnacles in all treatments at Trevone, this location was not included in the analysis. Due to the removal of the screws (vandalism) marking some treatments at Crackington Haven resulting in these treatments

79 Chapter Five

not being relocated, data for Crackington Haven were analysed separately using a single factor

ANOVA with treatment (four levels) considered fixed. Data for Looe and Cellar Beach were compared using a two factor ANOVA with the factor location (two levels) considered random and the factor treatment (four levels) considered fixed. Heterogeneity of variance was tested using

Cochran's test and where appropriate data were transformed. Student-Newman Keuls (SNK) post- hoc tests were carried out on significant data (p < 0.05).

rF' N

Crackington Haven

Trevone

Newquay " Looe " Cellar Beach

50km

Fig 5.1 Location of barnacle study sites in southwest Britain

5.3 Results

5.3.1 The effect of adult abundance of Semibalanus balanoides, on the recruitment

success of its gregarious settling young

It was apparent at all three locations, Newquay, Trevone and Looe, that there was considerably higher recruitment of Semibalanus spat onto natural rock substrate compared to settlement plates, with between six to nine times more Semibalanus recruiting to the natural substrate. Therefore formal statistical analysis was restricted to the experimental plates and caution must be taken extrapolating results to natural conditions.

At all locations higher recruitment of juvenile Semibalanus occurred on settlement plates with high densities (5 cm"Z) of adult con-specifics (Fig 5.2). The lowest recruitment of Semibalanus juveniles occurred on settlement plates where there were no adult con-specifics present. Although all locations showed a similar trend in the effects of adult con-specifics on juvenile recruitment success, statistical analysis revealed differences in the magnitude of this trend. There were also

80 Chapter Five

differences between locations in the effect of barnacle settlementfactor on the number of

Semibalanusspat to recruit to settlement plates. At both Trevone and Looe there was no effect of barnacle settlement factor. In contrast, at Newquay there were differences in the recruitment success of Semibalnus spat on settlement plates with barnacle settlement factor and those without.

At Trevone there was a significant effect of barnacle density (F 2,1818.79, P<0.01; Table

5.1a; Fig 5.2a). Significantly more Semibalanusspat recruited to settlement plates with high densities (5 cm"2)of adult barnacles compared to plates with low densities (1 cm'2) of adult barnacles compared to plates with no adult conspecifics (Table 5.1a; Fig 5.2a).

The settlement of Semibalanus spat at Looe followed the same trend as at Trevone, with high adult barnacle treatments having greater settlement of juveniles than low density treatments.

These in turn both had greater settlement than treatments without con-specific adults. This proved, however, not to be significant (F 2,182.16, P>0.14; Table 5.1b; Fig 5.2b).

At Newquay there was increased recruitmentto plates with high densities of adult barnacles compared to the low density and plates with no adult con-specifics. There was, however, a significant barnacle density x settlement factor interaction (F 2,125.48, P<0.05; Table

5.1c; Fig 5.2c). Post-hoc tests indicatedthere were no significant differences in the number of

Semibalanusspat to recruit to high density and low density treatments without settlement factor, although high density treatments experiencedhigher recruitment than low density treatments

(Table 5.1c; Fig 5.2c). Both high and low density treatments had significantlygreater recruitment than treatments with no adult con-specifics. Significantly more recruitment of Semibalanus juveniles occurred on plates with high densities of adult con-specificscompared to low density and no adult con-specifics in treatments with barnacle settlement factor (Table 5.1c; Fig 5.2c).

Although higher recruitment occurred in the low density treatments compared to treatmentswith no adult con-specifics, this proved not to significant (Table 5.1c; Fig 5.2c).

Although not always statistically significant it would appear that a higher abundance of adult conspecifics resulted in greater recruitment of Semibalanus spat compared to low density treatments and treatments with no conspecifics (Table 5.1; Fig 5.2). The presence of barnacle settlement factor had no effect, or in the case of two treatments at Newquay a negative effect on

Semibalanus recruitment (Table 5.1; Fig 5.2).

81 Chapter Five

Table 5.1. Analysis of variance of the number of Semibalanus to recruit to settlement plates (with barnacle settlement factor and without barnacle settlement factor) for three densities of attached adult con-specifics; high density (5cm'2), low density (1 cm2) and no adult conspecifics for a) Trevone (Cochran's test C=0.361, ns) b) Looe (Cochran's test C= 0.372, ns) and c) Newquay (Cochran's test C= 0.556, ns). Variances were homogenous for all locations following Ln (x + 1) transformation. -+ = direction of non-significant trend. a) Trevone

Source SS DF MS F p F versus Barnacle density (BD) 3.63 2 1.81 18.79 <0.01 Residual Settlement factor (SF) 0.00 1 0.00 0.00 0.93 Residual BD x SF 0.02 2 0.01 0.12 0.89 Residual Residual 1.74 18 0.10 Total 5.37 23 SNK tests High density > low density> no adult con-specifics

b) Looe

Source SS DF MS F P F versus Barnacle density (BD) 1.13 2 0.57 2.16 0.14 Residual Settlementfactor (SF) 0.04 1 0.04 0.17 0.69 Residual BD x SF 0.67 2 0.33 1.28 0.30 Residual Residual 4.73 18 0.26 . Total 6.57 23 Treatment trends High density - low density -º no adult consp ecifics

c) Newquay

Source SS DF MS F P F versus Barnacle density (BD) 4.44 2 2.22 39.79 <0.01 Residual Settlement factor (SF) 0.54 1 0.54 9.67 <0.01 Residual BD x SF 0.61 2 0.31 5.48 <0.05 Residual Residual 0.67 12 0.06 Total 6.26 17 SNK tests Barnacle density x settlement factor Settlement factor x barnacle density No SF SF No SF High -º Low > no adult con-specifics High -º SF SF SF High > Low -º no adult conspecific Low No > No adult conspecifics No SF - SF

82 Chapter Five

a) Trevone

70

60

50

m 40 äE 30

20

10

0 co

N Hgh Q Low Q No adult conspecifics b) Looe

8 70 ns 60 50 (ß N L 40 y0 30 a 20

10 w co 0

® High © Low o No adult conspecifics c) Newquay

m 70 ý- ** -. --' U 60 50 ý0 40 NE LO 30 20 c 10 ;QE0 a No Settlement factor Settlement factor

N Hign p Low Q No adult conspecitics

Fig 5.2. The abundance of Semibalanus to recruit to settlement plates (with barnacle settlement factor and (5cmZ), low without barnacle settlement factor) for three densities of adult con-specifics attached: high density 2) barnacle density (1 cm and no adult conspecifcs for a) Trevone: results were consistent for treatments with for settlement factor and those without, therefore data has been combined b) Looe: results were consistent Newquay. treatments with barnacle settlement factor and those without, therefore data has been combined c) Error bars i ISE. ** P<0.01, ns = not significant.

83 Chapter Five

5.3.2 The effects of poor 0+ Semibalanus recruitment on the subsequent recruitment

success of Chthamalus spat

At both Cellar Beach and Looe there was a significant location x treatment Interaction(F

3,327.31, P<0.01; Table 5.2a; Fig 5.3a) with SNK tests indicating that recruitment of Chthamalus spat showed a similar trend at both locations, although overall settlement was greater in treatments cleared of all barnacles at Looe (Table 5.2a). There was no difference between locations in the abundance of Chthamalus spat to recruit to high density (5 cm'2), low density (1 cm'2) and control treatments (Table 5.2a). At both locations there was significantly higher settlement in the cleared plots compared to the high Semibalanus 0+ density, low Semibalanus 0+ density and control treatments (Table 5.2a; Fig 5.3a). Although not statistically significant on average more

Chthamalus spat recruited to plots where the density of Semibalanus 0+ individuals had been manipulated to represent a poor Semibalanus recruitment years compared to plots which represented good recruitment years for Semibalanus 0+ individuals (Fig 5.3a). At Crackington

Haven there was no significant difference between any of the four treatments (F 3,120.06, P=0.98;

Table 5.2b; Fig 5.3b).

Table 5.2. Analysis of variance for the numberof Chthamalusto recruit to four treatments;cleared rock, control, high density 0+ Semibalanus(approx 5cm 2), low density 0+ Semibalanus(1cm 2) for a) Cellar Beach and Looe. Variances were homogenousfollowing square root (x + 1) transformation(Cochran's test C= 0.3545, ns) and b) CrackingtonHaven (Cochran'stest C= 0.5029, ns). a) Cellar Beach and Looe

Source SS DF MS FPF versus Location (Lo) 83.73 1 83.73 16.84 <0.01 Residual Treatment (Tr) 1327.10 3 439.03 12.08 0.04 Lo x Tr Location x Treatment 109.08 3 36.36 7.31 <0.01 Residual Residual 159.07 32 4.97 Total 1668.98 39 SNK tests Location x Treatment Treatment x Location Cleared rock Looe > Cellar Beach Looe Cleared rock > all other treatments Control Looe = Cellar Beach Cellar Beach Cleared rock > all other treatments High density Looe = Cellar Beach Low density Looe = Cellar Beach b) CrackingtonHaven

Source SS DF MS FPF versus Treatment 2126.83 3 708.94 0.06 0.98 Residual Residual 141554.17 12 11796.18 Total 143681.00 15

84 Chapter Five

a) Cellar Beach and Looe

ä 900 u 800 ns 700 m 600 i1!

00 ns 300 200 100 0 I Cellar Beach Looe

9 Cleared rock a Control Q High density Q low density

b) Crackington Haven

N E 900 800 ns CL 700 600

500 ' 400

(CL300 H 200 r 100 11 r P__ Üý

Cleared rock 0 Control Q High density Q Low density

Fig 5.3. Chthamalus spat recruitment to four treatments: bare rock, control, high density (5cm'2) and low density (1 cm-2) for a) Cellar Beach and Looe and b) Crackington Haven. Error bars ±1 SE. ** P<0.01, ns = not significant.

85 Chapter Five

5.4 Discussion

5.4.1 Limitations of work

Settlement plates were used instead of manipulating Semibalanus individuals on the

in Britain natural substrate due to the low numbers of Semibalanus on most shores southwest

(Moore, pers ob). Semibalanus covered rocks were collected from St Ives on the north coast of localised Cornwall, which has consistently had large populations of Semibalanus due to an area of have been upwelling (Southward 1967). Settlement plates, similar to the ones used in this study, ). shown to be effective for investigating Chthamalus recruitment in Portugal (Hawkins, pers. comm. to Unfortunately in the present work significantly higher numbers of Semibalanus spat recruited natural substrate compared to settlement plates, indicating that settlement plates had a negative effect on the recruitment success of Semibalanus spat. Jenkins (1995) also found that

Semibalanus cyprids behaved differently to artificial settlement surfaces with increased settlement on natural substrates.

A large number of adult Semibalanus died following attachment to settlement plates with significantly higher proportions of barnacles remaining attached to low density treatments in compared to high density treatments at Trevone and Newquay. There was no difference the proportion of adult Semibalanus surviving on the different treatments at Looe. Although both the to high and low density treatments experienced mortality, many dead barnacles were still attached the settlement plates. Barnacles are known to preferentially settle where evidence of previous 1963, Larman barnacle recruitment is present (Knight-Jones 1953, Crisp 1961, Crisp and Meadows high and Gabbott 1975, Rittschof et al. 1984, Yule and Walker 1984, Hills et al. 1998), therefore the levels of mortality that occurred may not have affected the outcome of the experiment as plates of dead Semibalanus were still attached to the settlement plates.

Previous experiments have successfully removed live barnacles from natural rock substrate and attached them to other substrates using the method that was used in the present work (Miron et at. 1996, Thompson et at. 1998). The higher levels of mortality experienced in the compared present work to other studies may be related to a combination of the length of time the experiment ran, the dynamic environment in which the settlement plates were attached and the

species of barnacle that was used. In the case of work by Thompson et at. (1998), Semibalanus was attached to the substrate for a much shorter period of time under laboratory conditions. The

present experiment was carried out on semi-exposed shores for a period of over a month, therefore 86 Chapter Five

the attached barnacles were exposed to a much more dynamic environment where the risk of dislodgement was greater. Field experiments were conducted by Miron et al. (1996) using a different species of barnacle, Balanus crenatus, where no barnacle mortality was recorded (Miron base that et al. 1996). Balanus crenatus has a calcareous base, in contrast to the membranous

Semibalanus possesses. It may be that the cyanoacrylate used to attach the adult Semibalanus It Is was able to leach out across this membranous base resulting in high adult mortality. much leach more unlikely that the glue used to attach Balanus crenatus to the substrate was able to across its calcareous base.

5.4.2 Effects of barnacle settlement factor on Semibalanus recruitment

Previous work has shown that there is increased settlement on panels with barnacle settlement factor compared to treatments that are without barnacle settlement factor (Crisp and

Meadows 1963, Larman and Gabbott 1975, Yule and Walker 1984, Hills et al. 1998). In this study barnacle settlement factor had a negligible or in the case of low density treatments and treatments In with no adult conspeciifcs at Newquay a negative affect on Semibalanus recruitment success. many cases where barnacle settlement factor has been shown to have a positive effect on cyprid (Hills settlement the experimental plates have only been exposed to cyprids for a period of minutes et al. 1998) or hours (Larman and Gabbott 1975). Experiments that have been carried out over a longer temporal scale have shown that the positive effect of barnacle cues is lost after a couple of months (Wethey 1984a, Le Tourneux and Bourget 1988). Evidence from the present work would suggest that the positive stimulus for cyprid settlement provided by barnacle cues might be lost even more quickly, particularly in dynamic intertidal environments such as semi-exposed shores of southwest Britain where the present work was carried out.

5.4.3 The effect of adult abundance of the barnacle, Semibalanus balanoides, on the

recruitment success of its gregarious settling young

At all locations a higher number of Semibalanus spat recruited to settlement plates with high densities (5cm'2) of conspecific adults compared to plates with low densities of (1cm-2)

Semibalanus adults and plates with no adults. At both Trevone and Newquay this effect was significant. At Looe, a similar trend was evident but was not statistically significant, probably as a result of Elminius modestus spat also settling on the settlement plates. Elminius is an introduced 87 Chapter Five

Australasian species of barnacle which produces larvae throughout the year (Rainbow 1984), although reproduction peaks in the summer months (Crisp 1958, Rainbow 1984, Pannacciulli

1995). Large numbers of both adult and juvenile Elminius were present at Looe, In contrast to

Trevone and Newquay where Elminius adults were rare or not found in rock control treatments.

Therefore it is likely that some Elminius spat recruited to the settlement plates at Looe. Due to the nature of the settlement plates used, Elminius spat were hard to distinguish from Semibalanus spat on enlarged digital images. Elminius spat are also less discerning than Semibalanus In their choice of settlement substrate (Hawkins, pers ob), it is therefore likely that the Increased recruitment of spat to control plates was a result of Elminius recruitment rather than Semibalanus recruitment.

The settlement plates from Looe were not collected and brought back to the laboratory until June

2003 when the barnacles present on the plates were counted. There were few if any Semibalanus juveniles on low density settlement plates and settlement plates with no con-specifics present. In contrast, between ten and one hundred Elminius juveniles were found on settlement plates with low density and no adult con-specifics, indicating that the high numbers of spat on low density and settlement plates without con-specifics were probably juvenile Elminius.

These results indicate that Semibalanus cyprid settlement was enhanced by the presence of Semibalanusadults and that when Semibalanusadults were found in higher densities there was increased settlement and recruitmentof Semibalanusjuveniles. These results agree with previous work which has shown that higher settlement occurred where there were the most remnants of adult Semibalanus(Kent et al. 2003). Recent work modelling the effect of low recruitmenton

Semibalanuspopulation dynamics found that four or more years of poor settlement is likely to alter

Semibalanuspopulations drastically (Svensson et al. 2005). As the model did not include the gregarious nature of Semibalanussettlement, it is likely that the model underestimatedthe reduction in the population size of Semibalanuswith an increase in the frequency of low recruitmentyears (Svensson et al. 2005). With continued climatic warming it is likely that the number of low or no recruitmentyears will increase in number and frequency, resulting in

Semibalanuspopulations becoming restricted to refuges or more northerly latitudes.

88 Chapter Five

5.4.4 The effects of poor 0+ Semibalanus recruitment on the subsequent recruitment

success of Chthamalus spat

At two out of the three locations examined in this study, significantly higher recruitmentof

Chthamalusjuveniles occurred in treatments that had been cleared of all adult barnacles suggesting that space for juvenile settlement may have been limited at these locations. At

Crackington Haven there was no difference in the number of Chthamalus juveniles to recruit to the cleared rock, control, high (5cm'2) and low (1cm'2) density treatments. The number of individuals to recruit to the treatments cleared of adults was also lower at Crackington Haven than at Looe and

Cellar Beach. Barnacle settlement success can be highly variable even when comparing locations in close proximity (Jenkins et at. 2000), therefore differences in the amount of recruitment to the three locations is not surprising. Crackington Haven is located on the north coast of Cornwall,

Britain whilst Looe and Cellar Beach are located on the south coast of Cornwall and Devon and the success of barnacle settlement and recruitment is highly dependent on factors such as currents and wind direction (Gaines et al. 1985, Kendall and Bedford 1987). As a consequenceit is likely that barnacle populations on two sides of the same coast may experience differences in recruitment success. As consequence of reduced supply of cyprids at Crackington Haven, resulting in the different levels of recruitment at this location, space may not have been a limiting factor. This may explain the lack of any significant difference in the recruitment success of

Chthamalus to cleared rock, control, high density and low density treatments at Crackington

Haven.

Although the amount of available space in treatments seemed to affect recruitment success none of the replicates had more than 85% barnacle cover following Chthamalus settlement. It would therefore appear that factors other than just the availability of space might be responsible for the increased number of Chthamalus individuals to recruit to treatments where adult barnacles had been removed. Manipulation of adult Chthamalus abundance in the treatments occurred two weeks before Chthamalus cyprids began to settle therefore there might have been some remnants of the barnacles, which acted as a cue to encourage the settlement of Chthamalus.

Evidence from the previous work on Semibalanus would suggest that this was unlikely as treatments with barnacle settlement factor did not increase Semibalanus settlement and other work has shown that barnacle settlement factor is unlikely to last long in the intertidal (Wethey 1984a). It was therefore unlikely that barnacle settlement cues were present in the treatments that had been 89 Chapter Five

cleared of adult barnacles. This theory however, cannot be totally discounted; particularly as the present work was carried out on natural substrate which is thought to absorb chemical cues more readily than artificial substrates such as those used in the Semibalanus work (Crisp and Barnes

1954).

Various species of barnacle are known to show territoriality in their behaviour,with individuals settling gregariously over a distance of centimetres, but demonstrating a tendency to Hui space themselves over a scale of millimetres (Crisp 1961, Barnett and Crisp 1979, Moyse and

1981, Kent et al. 2003). Therefore although up to 20% of treatments where adult con-specifics were present consisted of bare rock, space may still have appeared limited at the scale that

Chthamalus settlers were investigating. The result of such behaviour would be significantly higher been numbers of Chthamalus individuals recruiting to the treatments where all adult barnacles had removed. Alternatively some species of barnacle are able to distinguish their own species from others and preferably settle near con-specifics avoiding settling near other species (Larman and

Gabbott 1975). Therefore the increased settlement in treatments where adult barnacles had been removed may not indicate that this habitat was preferable, but that treatments where other species were present were not as preferable.

There was no significant effect of Semibalanus0+ abundance on the settlement success of

Chthamalus spat at any of the three locations. Therefore further work would be required to provide dominant support for the hypothesis of Southward and co-workers that Chthamalus becomes the barnacle species at most shore heights in southwest Britain during warm climatic conditions when they are able to take advantage of the free space made available by the reduction in the number of settlers and survivors of Semibalanus (Southward and Crisp 1956, Southward 1967,1991,

Southward et at. 1995).

5.4.5 The consequences of climate change on the inter- and Intraspecific interactions of a

northern and southern species of barnacle

My results indicate that the number of adult con-specifics present in the population influences Semibalanus recruitment success, with an increased density of adults resulting in increased recruitment of Semibalanus juveniles. Similar evidence for this comes from an

Australian species of barnacle that also relied on cues from adult conspecifics for good recruitment

(Jeffery 2000). Here a reduced adult population had far reaching consequences for barnacle 90 Chapter Five

population dynamics for many generations (Jeffery 2000). Thus as climatic warming Increases resulting in a reduction of adult Semibalanusin the overall barnacle population, there Is likely to be a direct negative effect on the chance of Semibalanuscyprids settling on the shore, irrespectiveof the numbers of cyprids present in the plankton. To compound this there will also be increased

Allee effects with reduced numbers of adults resulting in a reduction In mating opportunities, resulting in fewer larvae entering the plankton.

The results presented here also indicate that both inter and intra specific interactions are likely to have both positive and negative impacts on the relative abundance of these two barnacle taxa. With increased climatic warming it is likely that Semibalanus will become locally extinct along the coastline of southwest Britain in a similar way experienced in the 1950s (Southward 1967).

Semibalanus are likely to become progressively limited to more northerly latitudes, lower on the shore and refuges in sheltered bays. Although both Semibalanus and Chthamalus appear to occupy the same ecological niche they do differ in size with Semibalanus generally larger than

Chthamalus. Semibalanus has an important role in regulating rocky shore community dynamics because limpets are not as effective grazers across Semibalanus covered rock, which allows

Fucus propagules, which have settled amongst the barnacles, to escape from the grazing activities of limpets. With an increase in the abundance of the smaller Chthamalus at the expense of the larger Semibalanus the interaction between limpets, barnacles and canopy forming algae could alter with impacts for rocky shore community dynamics. The potential effects of a change in the strength and direction of this interaction for rocky shore community structure and ecosystem functioning is discussed in chapter six.

91 Chapter Six

Six

General Discussion

6.1 Thesis overview and summary

There is now strong evidence from both terrestrial and marine realms of shifts In species distributions and abundances (Barry et al. 1995, Parmesan 1996, Walther et al. 2002, Hiscock et al. 2004, Kendall et al. 2004, Parmesan et al. 2005). Although none of these studies have been able to demonstrate a causal link between these changes and anthropogenic climate change

(McCarty 2001), many of the changes in species distributions and abundance have in part been attributed to anthropogenic forcing of the climate (Hughes 2000, Parmesan and Yohe 2003). For the management and conservation of natural habitats it is essential to understand how entire assemblages as well as individual species will respond to environmental change. Most studies attempting to predict species range shifts in response to climatic warming have used a'climate envelope' approach (Jeffree and Jeffree 1996, Davies et al. 1998, Pearson and Dawson 2003,

Huntley et at. 2004, Thomas et at. 2004). However, this approach has recently been criticised because it does not take into account dispersal capability, biological interactions, and habitat quality, availability and connectivity, including refuges all of which are likely to have considerable influence on the responses of individual species to climate change (Brown et al. 1997, Davies et al.

1998, Helmuth and Hofmann 2001, Helmuth et al. 2002).

My work has used the rocky intertidal as a convenient model system to investigatethe role of direct and indirect biological interactions in modulating species responses to climate change. In particular I have investigated how predictions of species range shifts based on 'climate envelope' approaches may be modified by interactions between species. In this concluding chapter I briefly summarise the main findings of my work, before discussing the likely impacts of population changes and/or species loss in response to increased warming on rocky shore community structure, dynamics and ecosystem functioning. I also discuss the effects of changes in the intensity and temporal variance of natural processes and how this will complicate predictions of species shifts. Finally, I discuss the role of biotic interactions in modifying species responses to climate change with reference to the use of the 'climate envelope' approach in making predictions of species range shifts.

92 Chapter Six

Patella vulgata, a northern/boreal species, and P. depressa, a southern/lusitanian species, are the dominant midshore limpets in southwest Britain (Southward and Southward 1978,

Southward et at. 1995, Thompson et at. 2000, Jenkins et al. 2001, Boaventura et at. 2002a, Kendall et at. 2004). As a result of similarities In their ecology and behaviour It has been assumed that they have a similar functional role in controlling macroalgal abundance and therefore play a key role in rocky shore community structure and ecosystem functioning (Southward and Southward 1978,

Southward at at. 1995, Thompson et al. 2000, Jenkins et at. 2001, Boaventura et at. 2002a, Kendall et at. 2004, Jenkins et at. 2005). However, little has been done to quantify the spatial distribution, particularly in relation to Fucus patches, of P. depressa or on the relative role of each limpet species in controlling macroalgal abundance. In agreement with previous studies (Burrows and

Lodge 1950, Southward and Southward 1978, Hawkins 1983, Hawkins and Hartvoll 1983, Hartnoll and Hawkins 1985, Hawkins et at. 1992, Johnson et al. 1997) 1have shown that P. vu/gata preferentially aggregated beneath Fucus patches resulting in a more aggregated distribution across the whole shore (Chapter two). In contrast, I have shown that P. depressa is more randomly distributed across open rock and it appears likely that P. depressa avoids creating home scars beneath Fucus patches over small and large spatial scales and was much (Chapter two). In

Chapter three I demonstrated that P. depressa did not avoid creating home scars beneath Fucus patches to avoid competition with P. vulgata. Therefore other factors must be responsible for the avoidance of Fucus patches by P. depressa. Taken together, these results have shown that P. vulgata and P. depressa differ in their spatial distribution across the midshore on shores of southern Britain. Since the relative abundances of these two species are predicted to alter in response to climatic warming (Southward et al. 1995), this will result in a change in the overall spatial distribution of limpets on the shore, which is likely to have broad-scale implications for rocky shore community dynamics.

Canopy forming macroalgae are predicted to become restricted to more northern latitudes in response to climatic warming (Barry et al. 1995, Franklin and Forster 1997, Coelho et at. 2000).

Therefore Fucus patches were experimentally removed to investigate the response of P. vulgata and P. depressa to the loss of canopy forming macroalgae (Chapter three). P. vulgata experienced increased mortality and relocated home scars, often beneath new Fucus patches. In contrast, P. depressa showed little response to the loss of Fucus in terms of behaviour or mortality. These results indicated that' as climatic warming increases P. vulgata may become more reliant on 93 Chapter Six

habitats such as Fucus patches that ameliorate environmentalconditions. As a result of the loss of

Fucus the range of P. depressa may expand more rapidly than predicted as more of its preferred habitat, open rock, becomes available.

In chapter four I investigatedthe relative grazing activity of P. vulgata and P. depressato see if they both had the same role in controlling macroalgal abundance. During the course of the experiment P. vulgata grazed more than P. depressa. This was particularly evident during peaks in the spawning activity of P. depressa, when grazing intensity was significantly reduced. As a result of the reduction in grazing pressure by P. depressa, Fucus germlings were able to settle and grow to a point where they escaped the grazing activity of limpets. In treatments with over 40% Fucus' cover, grazing was greatly reduced for the remainder of the experiment resulting in small P. depressa having a lower growth increment than small limpets in other treatments. This result indicates P. depressa is not as effective as P. vulgata in controlling macroalgal abundance on shores of southwest Britain.

It is thought that a dynamic equilibrium exists between the barnacles Semibalanus balanoides and Chthamalus spp. with temperature and other environmental factors influencing their relative abundances. Changes in these climatic factors alter the strength of competitive interactions between these taxa (Southward 1967,1991, Southward et al. 1995). Therefore in chapter five I experimentally simulated climatically driven fluctuations in the abundance of these two taxa to inform predictions on the likely outcomes of climatic warming on barnacle population structure. Initially the effect of adult Semibalanus abundance on the recruitment success of its gregarious con-specific settling cyprids was Investigated. Treatments with a higher density of adults consistently experienced increased recruitment success compared to treatments with low densities of adult con-specifics and treatments without con-specifics, although this effect was not always statistically significant. These results indicated that the presence of adult con-specifics may have an important role in maintaining population size, and that a reduction in the adult population may affect the settlement success of future generations (see also Jeffery 2000, Kent et al. 2003). In the second part of this chapter I investigated the effects of simulated good and poor recruitment of

Semibalanus on the recruitment success of Chthamalus. There was no significant difference In the abundance of Chthamalus to recruit to treatments with high or low densities of Semibalanus 0+ at any of the three locations. Therefore further work needs to be undertaken to support the hypotheses put forward by Southward and co-workers that following a reduction in the abundance 94 Chapter Six

of the superior competitor, Semibalanus(Connell 1961b), Chthamalusis able to take advantageof the increased space available and increase in abundance (Southward 1991, Southwardet al. 1995,

Burrows, Hawkins and Leaper, unpub).

My results have shown that species which are thought to be very similar In terms of their biology ecology and biology may exhibit subtle differences in their behaviour or aspects of their

which can fundamentally alter how they interact with the abiotic and biotic environment. As a result

switches in the relative abundance of these species are likely to have broad-scale implications for

community structure and ecosystem functioning.

6.2 Consequences of Increased climatic warming for rocky shore community structure

and dynamics

Most of the work investigating the importance of limpets in controlling macroalgal

abundance has been carried out on shores where P. vulgata is the sole limpet present (Burrows

and Lodge 1950, Southward and Southward 1978, Hawkins 1983, Hawkins and Hartnoll 1983,

Hartnoll and Hawkins 1985, Hawkins et at. 1992). On shores where more than one species of

limpet is present, the relative importance of the grazing activity of the different species has not

been investigated (Thompson et at. 2000, Jenkins et at. 2001, Boaventura et al. 2002a, Jenkins et

al. 2005). In the past it has been assumed that due to similarities in the radula structure (Hawkins

et al. 1989), ecology and behaviour (Thompson et at. 2000, Jenkins et at. 2001, Boaventura et al.

2002a, Boaventura et at. 2002b) of P. vulgata and P. depressa that they had the same effect in

controlling macroalagal abundance.

My work has shown that these two limpet species have different spatial distributionson the

shore, interact differently with Fucus patches and have different roles in controlling the abundance

of macroalage on the shore. These differences in behaviour and ecology result in these two

species having different effects on the probability of Fucus escapes occurring. The effect Fucus

has on the two limpet species is also different. P. vulgata has been shown to control the

abundance of macroalgae on shores in the NE Atlantic by grazing juvenile algae in the microbial

film (Jones 1948, Burrows and Lodge 1950, Southward and Southward 1978, Hawkins 1981,

Hawkins and Hartnoll 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992, Burrows and

Hawkins 1998) and to a lesser extent by grazing on mature Fucus holdfasts (Hawkins et al. 1989,

Hawkins, pers. comm. ). Conversely adult and juvenile P. vulgata preferentially aggregate beneath 95 Chapter Six

Fucus patches (Chapter 2&3; Burrows and Lodge 1950, Southward and Southward 1978,

Hawkins 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992). Thus when Fucus was experimentally removed they experienced increased mortality and moved home scars, often large beneath new Fucus patches (Chapter 3). It would therefore appear that P. vulgata has a negative effect on juvenile Fucus, but conversely, when Fucus is lost there Is also a large negative impact on both adult and juvenile P. vulgata (Fig 6.1). In contrast, P. depressa has less of a negative impact on Fucus since mature P. depressa graze less during the settlement time of

Fucus. A higher proportion of adult P. depressa would therefore aid Fucus patch establishment

(Chapter 4; Fig 6.1). However, once mature, Fucus appears to have a negative effect on juvenile P. depressa (Chapter 4; Fig 6.1). As a result of the differences in the interaction of P. depressa and

P. vulgata with Fucus, it appears that both limpet species have indirect positive effects on each other by aiding in the creation of the other species preferred habitat (Fig 6.1).

It appears that P. vulgata can strongly affect communitydynamics by controlling macroalgal abundance on the shore. In contrast, P. depressa has less of an effect in controlling macroalgal abundance. Recent work has suggested that only a few species In a given assemblage have strong effects on community dynamics, whereas most will have weak effects either as a result of low abundance or small per capita effects (Berlow 1999). The effects of weak interactors have been shown to significantly alter natural communities because they play an importänt stabilizing role and may also magnify spatiotemporal variation in community structure (Berlow 1999). The loss, or addition, of any species, a weak or strong interactor, can lead to pronounced changes in community composition and structure (McCann 2000). Therefore if as predicted the relative abundance of P. depressa (weak effects) increases at the expense of P. vulgata (strong effects) there are likely to be broad-scale affects for rocky shore community structure.

Further work is needed to test the generality of this model (Fig 6.1) on shores that are exposed to different levels of wave action and also across a broader geographical area. In particular there is a need to investigate if there Is a latitudinal cline in the grazing activity of P. depressa. In previous chapters I suggested that in southwest Britain, near the northern range limit of P. depressa, this species may need increased temperatures experienced on open rock to aid in digestion or for other physiological reasons. If this is the case then P. depressa may have a more significant role in controlling macroalgal abundance in regions with increased temperature. If the grazing activity of P. depressa does increase across a latitudinal gradient then as climatic warming 96 Chapter Six

increasesthe small negative effect P. depressa has on juvenile Fucus at present may become a stronger. This may have implicationsfor the abundance of both P. vulgata and Fucus on shores In southwest Britain.

Adults Juveniles Fucus P. vulgata S veslculosus S ' Juveniles/ SI Adults adults

11,

1

1ý 1 ý"

-- Adults Juveniles

P. depressa Fucus vesiculosus Juveniles Adults

Fig 6.1. Interactionsbetween P. vulgata,P. depressaand Fucus vesiculosuson rocky shores In southwest Britain. - negativeinteraction; + positive interaction. The width of arrows representsthe strengthof the interaction. Dotted lines are indirect interactions,while full lines are direct Interactionsbetween species.

Barnacles, particularly Semibalanusbalanoides, also aid in the creation of Fucus patches on moderately exposed shores in the NE Atlantic (Burrows and Lodge 1950, Hawkins 1981,

Hawkins 1983, Hawkins and Hartnoll 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992).

Limpets are not as effective at grazing across barnacle covered rock, therefore barnacles are able to provide a refuge from limpet grazing activity allowing new Fucus patches to become established

(Hawkins 1981, Hawkins 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992). Most work investigating the interactions between limpets, Fucus and barnacles has been carried out on shores where Semibalanus and P. vulgata are the only limpet and barnacle species present

(Burrows and Lodge 1950, Hawkins 1983, Hartnoll and Hawkins 1985, Hawkins et al. 1992). At lower latitudes Semibalanus is joined by the barnacle species Chthamalus spp where there is also a greater diversity and abundance of grazers (Ballantine 1961). As Chthamalus is smaller than

Semibalanus, limpets may be able to graze more effectively across Chthamalus, and this species may not be as effective a refuge for Fucus from limpet grazing. Recent work has shown that there are regional differences in the effects of limpets on the abundance of macroalgae on the shore

(Jenkins et al. 2005). At more northern locations there was a strong deterministic response to

97 Chapter Six

reduced grazing with increased Fucus growth which was consistent between replicates, Further south there was less Fucus growth and it was more variable within replicates (Jenkins et al. 2005).

It was suggested that the regional differences were a result of the increased probability of algal escapes with increasing latitude as a result of differences in the supply of algal propagules as well as differences in the abundance and perhaps diversity of grazers (Jenkins et al. 2005). However, the switch from shores dominated by Semibalanus to those dominated by Chthamalus may also influence the chance of Fucus escapes occurring.

If limpets are more effective grazers across the smaller Chthamalus,the strength and direction of interactions between Fucus and barnacles may alter depending on the numerically dominant species of barnacle. On Semibalanus and P. vulgata dominated shores, Semibalanus has a direct negative effect on P. vulgata by inhibiting its ability to graze (Fig 6.2; Hawkins and

Hartnoll 1982). In contrast, Semibalanus has a direct positive effect on Fucus abundance by providing a refuge from limpet grazing encouraging Fucus escapes, therefore creating Fucus patches, the preferred habitat of P. vulgata (Fig 6: 2). Therefore although Semibalanus has a direct negative effect on P. vulgata, as a consequence of the interaction between Semibalanus and

Fucus, Semibalanus also has an indirect positive effect on P. vulgata (Fig 6.2). P. vulgata has a direct negative effect on Semibalanus because it is known to be able to plough barnacle cyprids from the substrate (Connell 1961a, Dayton 1971, Menge 1976, Hawkins 1983, but see Holmes et at. 2005 where limpets had a positive effect on cyprid settlement success) and barnacle plates have been found in the gut of P. vulgata (Hawkins et al. 1989). P. vulgata also has an indirect positive effect on Semibalanus by consuming the'Fucus which limits settlement of barnacle cyprids through the sweeping of their fronds (Menge 1976, Hawkins 1983, Jenkins et at. 1999c; Fig 6.2).

On shores dominated by Chthamalusit Is likely that the interaction between limpets, Fucus and barnacles could alter. Chthamalusmay have a negative effect on Fucus abundance as limpets may be more effective grazers across this smaller species of barnacle resulting in a reduction in Fucus escapes (Fig 6.2). As a result of Chthamalusnot providing a refuge for Fucus,

Chthamalus may have an indirect negative effect on P. vulgata because its preferred habitat,

Fucus patches, is likely to become patchier (Fig 6.2). Further manipulative experiments need to be undertaken to investigate how the interaction between limpets, Fucus and barnacles alters in response to changes in the species involved and an increase in grazer diversity.

98 Chapter Six

Grazers, macroaigae and barnacles interact in complicatedways and are expected to respond in individualisticways to environmentalchange. These species interactionshinder our ability to easily predict the impacts of climatic warming on rocky shore communities. However,with a better understandingof how biotic interactionsalter species responses to climacticwarming and how this affects community scale processes we will have a better chance of providing biologically realistic predictions on the responsesof species and communities to climatic change.

a) Cooler climatic conditions/northernlatitudes +

Fucus Semibalanus vesiculosus

"; . ..

.. + ".

P. vulgata /f

b) Warmer climatic conditions/southern latitudes

ý` ------'ý Chthamalus Fucus vesiculosus

1 j1 11 11 11 1 11 1 1 11 1 1 1 11 1 IS 1 1 i1 T 1 1 1 1 11 11 11 11 11 11 1/ 11 11

T

P. vulgata P. depressa

+ il

Fig 6.2. Interactionsbetween limpets, barnaclesand Fucus on a) shores dominatedby Semibalanusand P. vulgata and b) shores with P. vulgata and P. depressa and dominatedby Chthamalus.- - - negative interaction; + positive interaction. The width of arrows representsthe strength of the Interaction. Dotted lines are Indirect interactions,while full lines are direct interactionsbetween species.

99 Chapter Six

6.3 Consequences of changes in species abundances and ranges for rocky shore

ecosystem functioning

Species interactions may affect ecosystem processes by directly modifying pathways of energy and material flow (Chapin III et al. 2000) or indirectly by modifying the abundances of species with strong ecosystem effects (Power et al. 1996, Chapin III et al. 2000). Trophic interactions, in particular, can alter ecosystem processes by directly modifying fluxes of energy and materials (Chapin III et al. 2000). In the case of rocky shores of the NE Atlantic, limpets, in particular P. vulgata, control the abundance of macroalgae on the shore and therefore influence whether these shores are net importers or exporters of primary production (Hawkins et al. 1992).

The role of canopy forming macroalgae in near-shore nutrient cycling may have important ecological and economic consequences not only for intertidal environments, but also for coastal waters (Benedetti-Cecchi et al. 2001). Declines in some near shore fisheries have been blamed on a reduction in overall productivity of marine coastal assemblages, possibly as a result of a decline in canopy forming macroalgae (Worm et al. 2000, Benedetti-Cecchi et al. 2001). With changes in the relative abundance of P. vulgata and P. depressa it is likely that there will be changes in the strength of trophic interactions, which will have cascading effects on rocky shore and coastal water primary production and ecosystem functioning.

In the NE Atlantic the diversity and abundance of grazers increaseswith decreasing latitude (Ballantine 1961, Hawkins et al. 1992, Jenkins et al. 2001). Therefore an understanding of the importance of increased diversity or the level of 'redundancy' in systems with increased diversity will be important for understanding the balance between grazers and macroalgae. This is particularly important as grazer assemblages on shores in Britain are likely to alter in terms of abundance and diversity with climatic warming. Recent work carried out in rockpools has shown that a reduction in grazer diversity affected algal cover depending on the Identity of the species lost

(O'Connor and Crowe in press). My work has shown that even congeneric grazer species with similar radula structure can have different effects on macroalgal abundance. Further experiments which manipulate grazer diversity and abundance are needed in order to fully understand the likely effects of increased grazer diversity and abundance on the chance of macroalgal escapes, which in turn will influence rocky shore and near shore ecosystem functioning.

100 Chapter Six

6.4 Changes In the variance of climatic events Increases the difficulty In predicting

species responses to climate change

Most studies investigating ecological responses to climate change only Incorporate climate variables at a constant rate (Hobble et al. 1999, Callaway et at. 2002, Klanderud 2005). Recent work has highlighted that under predicted climate scenarios there will be an increase in extreme

events, changing the intensity and temporal variance of natural processes (Benedetti-Cecchi

2003). For example, in southwest Britain, the time of greatest warming will alter depending on the

season. In summer warming is likely to be greater at day than at night. In contrast, during the winter greater warming is expected at night rather than during the day (South West Climate

Impacts Partnership [SWCCIP] 2003). Precipitation rates are also predicted to alter and become

more variable depending on the season with dryer summers and wetter winters (SWCCIP 2003).

In addition, changes in environmental conditions will occur simultaneously and are likely to interact, with potentially differing effects than those predicted from a single climate variable on species

responses to climate change. Recent work has shown that the interaction of temperature and

ultra-violet radiation had a different effect than the effect of temperature alone on the early life

stages of two algal species (Hoffman et at. 2003). As a result of the Interaction of environmental factors and changes in the intensity and temporal variance of natural processes our ability to

predict species range shifts will be hampered further (Benedetti-Cecchi 2003). For example, warmer wetter conditions are likely to reduce desiccation stress resulting in species being more able to tolerate increased air temperature. If conditions are warm and dry intertidal species will be affected not only by the increase in temperature, but also an increase In desiccation stress. This

may result in an increase in heat stress mortality, particularly for high shore species, during warm dry conditions.

Studies which incorporate biotic interactions in scenarios with differing variance and mean values of anticipated climate will help in our understanding of likely changes in assemblages at scales relevant to the management and conservation of those systems. However the complexity of abiotic factors and biological interactions makes precise and accurate predictions of species range shifts extremely difficult.

101 Chapter Six

6.5 'Climate envelope' approach vs biotic Interactions

The climate envelope approach has been widely used to predict species range shifts under various climate change scenarios (Scott and Poynter 1991, Jeffree and Jeffree 1994, Jeffree and

Jeffree 1996, Berry et at. 2002, Pearson and Dawson 2003,2005). More recently this approach

has been criticised because the interaction of abiotic and biotic factors are likely to modify how

species will respond to climatic warming (Brown et at. 1997, Davies et al. 1998, Martin 2001,

Callaway et at. 2002, Schmitz et at. 2003, Jiang and Morin 2004). Through descriptive studies and

experimental manipulation of natural assemblages my thesis has shown that habitat ameliorating

species are able to influence the balance of species with northern and southern biogeographic

distributions and that species, which are thought to be very similar in terms of their physiology and

ecology, can exhibit subtle differences in behaviour, which may fundamentally alter community

dynamics in benthic habitats.

My results have demonstratedthat the interactions between three groups of organisms, in

this case Fucus and two northern and southern species of limpet and barnacle, can result in

complicated biological interactions. As a result of increased climatic warming these interactions

are also likely to alter. It is therefore unlikely that predictions of species range shifts based on

species responses to a single climatic variable will be reliable. The results of my thesis support

previous work which has also criticised the predictive power of the 'climate envelope' approach in

calculating the future biogeographic distributions of species as they respond to climatic warming

(Brown et al. 1997, Davies et al. 1998). It has been suggested that using the 'climate envelope'

approach tends to broaden the realised niche of species (Berry et al. 2002) and underestimate the shifts in species boundaries. As a consequence Davies et al. (1998) suggest that the 'climate envelope' approach should only be used as a null model, while Pearson and Dawson (2003) suggest the approach can only predict species range shifts, at regional (200-2000 km) to global (>

10 000 km) scales. At smaller spatial scales, which are relevant to the management and conservation of natural systems, the combination of abiotic factors and biotic interactions reduce the reliability of'climate envelope' approaches (Pearson and Dawson 2003). I therefore suggest that the 'climate envelope' approach is reasonable for the forecasting of potential niches at scales greater than a few hundred kilometres, but at smaller scales it cannot be used to predict changes in abundance or rates of change due to abiotic and biotic interactions modifying how species respond to climate. At smaller spatial scales manipulative experiments could provide information on how 102 Chapter Six

the interactionsbetween populations and assemblageswill alter In response to climate change and if used in conjunction with the 'climate envelope' approach, allow biologically realistic predictionsof species range shifts.

103 References

Albritton, D. L., L. G. Meira Filho, U. Cubasch, D. Dal, Y. Ding, D. J. Griggs, B. Hewitson, J. T.

Houghton, I. Isaksen, T. Karl, M. McFarland, V. P. Meleshko, J. F. B. Mitchell, M. Noguer,

B. S. Nyenzi, M. Oppenheimer, J. E. Penner, S. Pollonais, T. Stocker, and K. E. Trenberth.

2001. Technical summary. Pages 881 in J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer,

P. J. van der Linden, X. Dal, K. Maskell, and C. E. Johnson, editors. Climate change 2001:

The scientific basis. Contribution of working group 1 to the third assessment report of the

intergovernmental panel on climate change. Cambridge University Press, Cambridge,

United Kingdom and New York, NY, USA.

Alcock, R. 2003. The effects of climate change on rocky shore communities in the Bay of Biscay,

1895-2050. University of Southampton, Southampton.

Anderson, M. J., and A. J. Underwood. 1994. Effects of substratum on the recruitment and

developmentof an intertidal estuarine fouling assemblage.Journal of ExperimentalMarine

Biology and Ecology 184:217-236.

Andrewartha, H. G., and L. C. Birch. 1954. The distribution and abundance of animals. University

of Chicago Press, Chicago.

Arrontes, J. 2002. Mechanismsof range expansion in the intertidal brown alga Fucus serratus in

northern Spain. Marine Biology 141:1059-1067.

Ballantine, W. J. 1961 a. A biologically-defined exposure scale for the comparative description of

rocky shores. Field Studies 1:73-84.

Ballantine,W. J. 1961b. The population dynamics of Patella vulgata and other limpets. University of London, London.

Barnes, H. 1958. Regarding southern limits of Ba/anusbalnoldes. Oikos 9:139-157.

Barnes, R. S. K., P. Calow, and P. J. W. Olive. 1993. The Invertebrates:a new synthesis, 2nd

edition. Blackwell Science, Oxford.

Barnett, B. E., and D. J. Crisp. 1979. Laboratory studies of gregarious settlement in Balanus

balanoides and Elminius modestus In relation to competition between these species.

Journal of the Marine Biological Association of the UK 59:581-590.

Barry, J. P., C. M. Baxter, R. D. Sagarin, and S. E. Gilman. 1995. Climate-related, long-term faunal

changes in a Californian rocky intertidal community. Science 267: 672-675. 104 Baxter, J. M. 1983. Allometric relationshipsof Patella vulgata shell characters at three adjacent

sites Sandwick Bay in Orkney. Journal of Natural History 17:743-755.

Beardall, J., S. Beer, and J. A. Raven. 1998. Biodiversityof marine plants in an era of climate

change:Some predictions based on physiologicalperformance. Botanica Marina 41: 113- 123.

Beaugrand,G., P. C. Reid, F. Ibanez,J. A. Lindley, and M. Edwards. 2002. Reorganizationof

North Atlantic marine copepod biodiversity and climate. Science 296: 1692-1694.

Beebee, T. 1995. Amphibian breeding and climate. Nature 374:219-374.

Begon, M., J. L. Harper, and C. R. Townsend, editors. 1996. Ecology: individuals, populationsand

communities,3rd edition. Blackwell Science.

Benedetti-Cecchi, L. 2003. The importance of the variance around the mean effect size of

ecological processes. Ecology 84: 2335-2346.

Berlow, E. L. 1999. Strong effects of weak interactions in ecological communities. Nature 398:330-

334.

Berry, P. M., T. P. Dawson, P. A. Harrison, and R. G. Pearson. 2002. Modelling potential impacts

of climate change on the bioclimatic envelope of species in Britain and Ireland. Global

Ecology and Biogeography11: 453-462.

Bertness, M. D., and R. Callaway. 1994. Positive interactions in communities.Trends in Evolution

and Ecology 9: 191-193.

Bertness, M. D., and G. H. Leonard. 1997. The role of positive interactions in communities: lessons

from intertidal habitats. Ecology 78: 1976-1989.

Bertness, M. D., G. H. Leonard, J. M. Levine, and J. F. Bruno. 1999. Climate-driveninteractions

among rocky intertidal organisms caught between a rock and a hot place. Oecologia 120:446-450.

Bertness, M. I., and P. J. Ewanchuk.2002. Latitudinal and climate-drivenvariation in the strength

and nature of biological interactions in New England salt marshes. Oecologia 132:392-401.

Birch, L. C. 1957. The meanings of competition.American Naturalist 91:5-18.

Boaventura, D., M. Alexander, P. D. Santina, N. D. Smith, P. Re, L. C. da Fonseca, and S. J.

Hawkins. 2002a. The effects of grazing on the distribution and composition of low-shore

algal communities on the central coast of Portugal and on the southern coast of Britain.

Journal of ExperimentalMarine Biology and Ecology 267:185-206. 105 Boaventura,D., L. C. D. Fonseca, and S. J. Hawkins. 2002b. Analysis of competitive interactions

between the limpets Patella depressa Pennant and Patella vulgata L. on the northern coast

of Portugal. Journal of ExperimentalMarine Biology and Ecology 271:171-188.

Boaventura,D., R. Pedro, L. C. da Fonseca, and S. J. Hawkins. 2002c. Intertidal rocky shore

communities of the continentalPortuguese coast: analysis of distribution patterns. Marine

Ecology - Pubblicazioni Della Stazione Zoologica di Napolii 23: 69-90.

Boaventura, D., L. Cancela Da Fonseca, and S. J. Hawkins. 2003. Size matters: Competition within

populations of the limpet Patella depressa. Journal of Animal Ecology 72:435-446.

Bolton, J. J., and R. J. Anderson. 1990. Correlation between intertidal seaweed community

composition and sea water temperaturepatterns on a geographical scale. Botanica Marina 33: 447-457.

Bowman, R. S. 1981. The morphology of Patella spp. juveniles In Britain and some phylogenetic

inferences. Journal of the Marine Biological Association, UK 61:647-666.

Bowman, R. S., and J. R. Lewis. 1977. Annual fluctuations in the recruitmentof Patella vulgata L.

Journal of the Marine BiologicalAssociation, UK 57:793-815.

Bowman, R. S., and J. R. Lewis. 1986. Geographicalvariation in the breeding cycles and

recruitment of Patella spp. Hydrobiologia 142: 41-56.

Boyden, C. R., and J. R. Zeldis. 1979. Preliminaryobservations using an attached microphonic

sensor to study feeding behaviour of an intertidal limpet. Estuarine and Coastal Marine

Science 9: 759-769.

Branch, G. M. 1981. The biology of limpets: Physical factors, energy flow, and ecological

interactions.Oceanography and Marine Biology Annual Review, London 19:235-380.

Brawley, S. H., and L. E. Johnson. 1991. Survival of fucoid embryos in the intertidal zone depends

upon developmentalstage and microhabitat. Journal of Phycology27: 179-186.

Breeman,A. M. 1988. Relative importance of temperature and other factors in determining

geographic boundariesof seaweeds: experimental and phenologicalevidence.

Helgolander Meeresuntersuchungen42: 199-241.

Breeman,A. M. 1990. Expectedeffects of changing sea-water temperatureson the geographic

distribution of seaweed species. Pages 69-76 in J. J. Beukema,W. J. Wolff, and J. J. W.

M. Brouns, editors. Expected effects of climate change on marine coastal ecosystems. Kluwer Academic Publishers.

106 Brown, J. H., T. J. Valone, and C. G. Curtin. 1997. Reorganizationof an arid ecosystemIn

response to recent climate change. Proceedingsof the National Academy of Sciences

USA 94: 9729-9733.

Burrows, E. M., and S. M. Lodge. 1950. A note on the inter-relationshipsof Patella, Balanus and

Fucus on a semi-exposedcoast. Annual Report of the Marine Biological Station, Port Erin

62:30-34.

Burrows, M. T. 1988. The comparative biology of Chthamalus stellatus (Poli) and Chthama/us

montagul Southward. Manchester University.

Burrows, M. T., S. J. Hawkins, and A. J. Southward. 1992. A comparison of reproduction in co-

occurring Chthamalid barnacles, Chthamalusstellatus (Poli) and Chthamalusmontagul

(Southward). Journal of Experimental Marine Biology and Ecology 160: 229-249.

Burrows, M. T., and S. J. Hawkins. 1998. Modelling patch dynamics on rocky shores using

deterministiccellular automata. Marine Ecology Progress Series 167:1-13.

Callaway, R. M., R. W. Brooker, P. Choler, Z. Kikvidze, C. J. Lortie, R. Michalet, L. Paolini, F. I.

Pugnaire, B. Newingham,E. T. Aschehoug, C. Armas, D. Kikodze, and B. J. Cook. 2002.

Positive interactionsamong alpine plants increase with stress. Nature 417:844-848.

Calvo, S., I. Barbara, and J. Cremades. 1999. Benthic algae of salt marshes (CorrubedoNational

Park, NW Spain): the flora. Botanica Marina 42:343-353.

Chapin III, F. S., E. S. Zavaleta, V. T. Eviner, R. L. Naylor, P. M. Vitousek, H. L. Reynolds, D. U.

Hooper, S. Lavorel, 0. E. Sala, S. E. Hobble, M. C. Mack, and S. Diaz. 2000.

Consequences of changing biodiversity. Nature 405: 234-242.

Chelazzi, G. 1990. Eco-ethologicalaspects of homing behaviour in molluscs. Ethology, Ecology

and Evolution 2: 11-26.

Chelazzi, G., G. Terranova, and P. Della Santina. 1990. A field technique for recording the activity

of limpets. Journal of the Molluscan Society 56:595-600.

Clarke, A. 1996. The influence of climate change on the distribution and evolution of organisms.

Pages 377-407 in I. A. Johnston and A. F. Bennett, editors. Animals and temperature:

Phenotypicand evolutionaryadaptation. Society for experimentalbiology seminar series, Cambridge University Press.

107 Coelho, S. M., J. W. Rijstenbil,and M. T. Brown. 2000. Impacts of anthropogenicstresses on the

early development stages of seaweeds.Journal of Aquatic Ecosystem Stress and

Recovery 7:317-333.

Colebrook,J. M. 1979. Continuousplankton records: seasonal cycles of phytoplanktonand

Copepods in the North Atlantic Ocean and the North Sea. Marine Biology 51:23-32.

Connell, J. H. 1961a. Effects of competition,predation by Thais lapillus, and other factors on

natural populations of the barnacle Balanus balanoides. Ecological Monographs 31: 61-

104.

Connell, J. H. 1961 b. The influence of interspecific competition and other factors on the distribution

of the barnacle Chthamalus stellatus. Ecology 42: 710-723.

Connell, J. H. 1972. Community interactionson marine rocky intertidal shores. Annual Review of

Ecology and Systematics 3: 169-192.

Crick, H. Q. P., C. Dudley, D. E. Glue, and D. L. Thomson. 1997. UK birds are laying eggs earlier.

Nature 388:526.

Crick, H. Q. P., and T. H. Sparks. 1999. Climate change related to egg-laying trends. Nature

399: 423-424.

Crisp, D. J. 1958. The spread of Elminius modestus Darwin in north-west Europe. Journal of the

Marine Biological Association of the UK 37:483-520.

Crisp, D. J. 1961. Territorial behaviour in barnacle settlement. Journal of ExperimentalBiology

38: 429-446.

Crisp, D. J., and H. Barnes. 1954. The orientation and distribution of barnacles at settlementwith

particular reference to surface culture. Journal of Animal Ecology 23: 142-162.

Crisp, D. J., and A. J. Southward. 1958. The distribution of intertidal organisms along the coasts of

the English Channel. Journal of the Marine Biological Association, UK 37:157-208.

Crisp, D. J., and P. S. Meadows. 1963. Adsorbed layers: the stimulus to settlement in barnacles.

Proceedingsof the Royal Society of London Series B- Biological Sciences 158:364-387.

Crisp, D. J., A. J. Southward, and E. C. Southward. 1981. On the distribution of the intertidal

barnacles Chthamalusstellatus, Chthamalusmontagul and Euraphla depressa. Journal of

the Marine Biological Association of the UK 61:359-380.

Cushing, D. H. 1961. On the failure of the Plymouth Herring fishery. Journal of the Marine

BiologicalAssociation of the U.K. 41:799-816.

108 Cushing, D. H., and R. R. Dickson. 1976. Biological response in the sea to climate changes.

Advances in Marine Biology 14:1-22.

Davis, A. J., L. S. Jenkinson, J. H. Lawton, B. Shorrocks, and S. Wood. 1998a. Making mistakes

when predicting shifts in species range in response to global warming. Nature 391:783-

786.

Davis, A. J., J. H. Lawton, B. Shorrocks, and L. S. Jenkinson. 1998b. Individualisticspecies

responses invalidate simple physiological models of community dynamics under global

environmentalchange. Journal of Animal Ecology 67:600-612.

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

subsequent utilization of space in a rocky intertidal community. Ecological Monographs

41: 351-389.

Deal, M. S., M. E. Hay, D. Wilson, and W. Fenical. 2003. Galactolipids rather than phlorotannins as

herbivore deterrents in the brown seaweed Fucus vesiculosus.Oecologia 136:107-114.

Delany, J., A. A. Myers, and D. McGrath. 1998. Recruitment,immigration and population structure

of two coexisting limpet species in mid-shore tidepools, on the West Coast of Ireland.

Journal of ExperimentalMarine Biology and Ecology 221:221-230.

Della Santina, P., E. Naylor, and G. Chelazzi. 1994. Long term field actography to assess the

timing of foraging excursions in the limpet Patella vulgata L. Journal of Experimental

Marine Biology and Ecology 178:193-203.

Easterling, D. R., B. Horton, P.D. Jones, T.C. Peterson, T. R. Karl, D.E. Parker M.J. Salinger, V.

Razuvayev, N. Plummer, P. Jamason, and C. K. Folland 1997. Maximum and minimum

temperature trends for the globe. Science 277:364-367.

Easterling, D. R., G. A. Mechl, C. Parmesan, S. A. Changnon, T. R. Karl, and L. 0. Mearns. 2000.

Climate extremes: Observations,modelling and impacts. Science 289:2068-2074.

Evans, R. G. 1948. The lethal temperaturesof some common British littoral molluscs. Journal of Animal Ecology 17:165-. 173.

Feminella,J. W., M. E. Power, and V. H. Resh. 1989. Periphyton responses to invertebrategrazing

and riparian canopy in three northern California coastal streams. FreshwaterBiology 22:445-457.

Fields, P. A., J. B. Graham, R. H. Rosenblatt, and G. N. Somero. 1993. Effects of expected global

climate change on marine faunas. Trends in Evolution and Ecology 8:361-367. 109 Fischer-Piette,E. 1948. Sur les elements do prosperite des Patelles et sur lour specificite.Journal de Conchyliologie88: 45-96.

Fischer-Piette,E. 1955. Repartition,le long des cotes septentrionalesdo I'Espagne,des principales

especes peuplant les rochers intercotidaux.Annales de ('Institut Oceanographique31: 37-

124.

Fischer-Piette,E., and M. Prenant. 1956. Distributiondes Cirripedes intercotidauxd'Espagne

septentrionale. Bulletin du Centre d'etudes et de recherches scientifiques Biarritz 1: 7-19.

Fischer-Piette,E., and M. Prenant. 1957. Quelques donnees ecologiques sur les cirripedes

intercotidauxdu Portugal, de I' Espagne du Sud et du Nord du Maroc. Bulletin du Centre

d'6tudes et de recherches scientifiques Biarritz 1: 361-368.

Fischer-Piette,E., and J.-M. Gaillard. 1959. Les patelles, au long des cotes atlantiques iberiques et Nord-Marocaines. Journal de Conchyliologie 99: 135-200.

Forbes. 1853. The natural history of Europes seas. John van Vorst, London.

Foster, B. A. 1971. On the determinantsof the upper limit of intertidal distribution of barnacles

(Crustacea: Cirripedia).Journal of Animal Ecology 40:33-48.

Fox, J. W., and P. J. Morin. 2001. Effects of intra- and interspecific interactions on species

responses to environmentalchange. Journal of Animal Ecology 70:80-90.

Franklin, L. A., and R. M. Forster. 1997. The changing irradiance environment:consequences for

marine macrophyte physiology, productivity and ecology. EuropeanJournal of Phycology

32: 207-232.

Fretter, V., and A. Graham. 1976. The Prosobranch molluscs of Britain and Denmark. I.

Pleurotomariacea,Fissurellacea and Patellacea.Journal of Molluscan Studies

Supplement 1:37.

Gaines, S., S. Brown, and J. Roughgarden. 1985. Spatial variation in larval concentrationsas a

cause of spatial variation in settlement for the barnacle, Balanus glandula. Oecologia 67:267-272.

Genner, M. J., D. W. Sims, V. J. Wearmouth, E. J. Southall, A. J. Southward, P. A. Henderson,and

S. J. Hawkins. 2004. Regional climatic warming drives long-term community changes of

British marine fish. Proceedingsof the Royal Society of London Series B- Biological Sciences 271:655-661.

110 Graham, R. W., and E. C. Grimm. 1990. Effects of global climate change on the patterns of

terrestrial biological communities.Trends in Evolution and Ecology 5:289-292.

Guerra, M. T., and M. J. Gaudencio. 1986. Aspects of the ecology of Patella spp on the Portuguese

coast. Hydrobiologia142: 57-69.

Hansson, L. J., I. R. Hudson, R. J. Seddon, 0. Shaw, and J. C. Thomason. 2003, Massive

recruitment of the barnacle Semibalanusbalanoides in the Clyde Sea (Scotland,UK) in the

spring of 2000. Journal of the Marine Biological Association of the UK 83: 923-924.

Harley, C. D. G. 2002. Light availability indirectly limits herbivore growth and abundance in a high

rocky intertidal community during the winter. Limnology and Oceanography47: 1217-1222.

Harrington, R., I. Woiwod, and T. Sparks. 1999. Climate change and trophic Interactions.Trends In

Evolution and Ecology 14:146-150.

Hartnoll, R. G., and J. R. Wright. 1977. Foraging movements and homing behaviour In the limpet

Patella vulgata L. Animal Behaviour25: 806-810.

Hartnoll, R. G., and S. J. Hawkins. 1985. Patchiness and fluctuations on moderately exposed rocky

shores. Ophelia 24:53-63.

Hawkins, S. J. 1981. The influenceof Patella grazing on the fucoid-barnaclemosaic on moderately

exposed rocky shores. Kieler Meeresforsch5: 537-543.

Hawkins, S. J. 1983. Interactionof Patella and macroalgae with settling Semibalanusbalanoides

(L.). Journal of ExperimentalMarine Biology and Ecology 71:55-72.

Hawkins, S. J., and R. G. Hartnoll. 1982. Settlement patterns of Semibalanusbalanoides (L.) in the

Isle of Man (1977-1981).Journal of ExperimentalMarine Biology and Ecology 62:271-283.

Hawkins, S. J., and R. G. Hartnoll. 1983a. Changes In a rocky shore community: an evaluation of

monitoring. Marine EnvironmentalResearch 9: 131-181.

Hawkins, S. J., and R. G. Hartnoll. 1983b. Grazing of intertidal algae by marine Invertebrates.

Oceanographyand Marine Biology Annual Review, London 21:195-282.

Hawkins, S. J., and R. G. Hartnoll. 1985. Factors determining the upper limits of intertidal canopy-

forming algae. Marine Ecology Progress Series 20:265-271.

Hawkins, S. J., R. G. Hartnoll, J. M. Kain, and T. A. Norton. 1992. Plant animal interactionson hard

substrata in the north-east Atlantic. in D. M. John, S. J. Hawkins, and J. H. Price, editors.

Plant animal interactions in the marine benthos. Clarendon Press, Oxford.

111 Hawkins, S. J., D. C. Watson, A. S. Hill, S. P. Harding, M. A. Kyriakides, S. Hutchinson,and T. A.

Norton. 1989. A comparisonof feeding mechanisms in microphagous,herbivorous,

intertidal, Prosobranchsin relation to resource partitioning. Journal of MolluscanStudies

55:151-165.

Hawkins, S. J., and A. J. Southward. 1992. The Torrey Canyon oil spill: recovery of rocky shore

communities. Pages 716 in G. W. Thayer, editor. Restoring the nation's marine

environment. Maryland Sea Grant, College Park, Maryland.

Hawkins, S. J., A. J. Southward, and M. J. Genner. 2003. Detection of environmentalchange In a Total marine ecosystem - evidence from the western English Channel. Science of the Environment 310: 245-256.

Heck, K. L., and G. S. Wetstone. 1977. Habitat complexity and invertebrate species richness and

abundance in tropical seagrass meadows. Journal of biogeography4: 135-142.

Helmuth, B., C. D. Harley, P. M. Halpin, M. O'Donnell, G. E. Hofmann, and C. A. Blanchette.2002.

Climate change and latitudinal patterns of intertidal thermal stress. Science 298:1015-

1017.

Helmuth, B., and G. E. Hofmann.2001: Microhabitats,thermal heterogeneity,and patterns of

physiologicalstress in the rocky intertidal zone. Biological Bulletin 201:374-384.

Helmuth, B. S. 1998. Intertidal mussel microclimates:predicting the body temperature of a sessile

invertebrate. Ecological Monographs68: 51-74.

Herbert, R. J. H., S. J. Hawkins, M. Sheader, and A. J. Southward. 2003. Range extension and

reproductionof the barnacle Balanus perforatus in the eastern English Channel. Journal of

the Marine BiologicalAssociation of the U.K. 83:73-82.

Hicks, G. R. 1980. Structure of phytal harpacticoid copepod assemblagesand the Influence of

habitat complexity and turbidity. Journal of ExperimentalMarine Biology and Ecology 44:157-192.

Hill, A. S., and S. J. Hawkins. 1991. Seasonal and spatial variation of epilithic microalgae

distribution and abundance and its ingestion by Patella vulgata on a moderately exposed

rocky shore. Journal of the Marine Biological Association of the UK 71:403-423.

Hills, J. M., J. C. Thomason, J. L. Milligan, and M. Richardson. 1998. Do barnacle larvae respond

to multiple settlement cues over a range of spatial scales? Hydrobiologia3751376: 101-111.

112 Hiscock, K., A. J. Southward, I. Tittley, and S. J. Hawkins. 2004. Effects of changing temperature

on benthic marine life in Britain and Ireland. Aquatic Conservation: Marine and Freshwater

Ecosystems 14:333-362.

Hiscock, K., A. J. Southward, I. Tittley, A. Jory, and S. J. Hawkins. 2001. The Impact of climate

change on subtidal and intertidal benthic species in Scotland. Report to Scottish National

Heritage from the Marine BiologicalAssociation of the United Kingdom - Final report. Marine Biological Association, Plymouth.

Hobbie, S. E., A. Shevtsova, and F. S. Chapin. 1999. Plant responses to species removal and

experimentalwarming in Alaskan tussock tundra. Oikos 84:417-434.

Hoffman, J. R., L. J. Hansen, and T. Klinger. 2003. Interactionsbetween UV radiation and

temperature limit inferencesfrom single-factorexperiments. Journal of Phycology39: 268-

272.

Holbrook, S. J., R. J. Schmitt, and J. J. S. Stephens. 1997. Changes in an assemblageof

temperate reef fishes associated with a climate shift. Ecological Applications 7: 1299-1310.

Holmes, S. P., G. Walker, and J. van der Meer. 2005. Barnacles, limpets and periwinkles:the

effects of direct and indirect interactions on cyrprid settlement and success. Journal of Sea

Research 53: 181-204.

Houghton,J. T., L. G. Meira Filho, B. A. Callander, N. Harris, A. Kattenberg,and K. Maskell. 1996.

Climate Change1995:The science of climate change. Cambridge University Press,

Cambridge, UK.

Hughes, L. 2000. Biological consequencesof global warming: Is the signal already apparent?

Trends in Evolution and Ecology 15:56-61.

Hulme, P. E. 1994. Seedling herbivory in grassland: relative impact of vertebrate and invertebrate

herbivores. Journal of Ecology 82:873-880.

Hulme, M., and G. Jenkins. 1998. Climate change scenarios for the United Kingdom UKCIP

Technical report no. 1.

Hulme, M., G. J. Jenkins, X. Lu, J. R. Turnpenny, T. D. Mitchell, R. G. Jones, J. Lowe, J. M.

Murphy, D. Hassell, P. Boorman, R. McDonald, and S. Hill. 2002. Climate Change

Scenarios for the United Kingdom: The UKCIP02 Scientific Report. Tyndall Centre for

Climate Change Research, School of Environmental Sciences, University of East Anglia,

Norwich, UK.

113 Huntley, B. 1991. How plants respond to climate change - Migration rates, individualismand the

consequencesfor plant-communities.Annals of Botany 67: 15-22 suppl.

Huntley, B., R. E. Green, Y. C. Collingham, J. K. Hill, S. G. Willis, P. J. Bartlein, W. Cramer, W. J.

M. Hagemeijer,and C. J. Thomas.2004. The performanceof modelsrelating species

geographicaldistributions to climate is independent of trophic level. Ecology Letters 7:417-

426.

Hurrell, J. W., and H. VanLoon. 1997. Decadal variations In climate associated with the north

Atlantic oscillation. Climate Change 36:301-326.

Hutchins, L. W. 1947. The bases for temperaturezonation in geographical distribution. Ecological

Monographs 17: 325-335.

Jeffery, C. J. 2000. Settlement in different-sizedpatches by the gregarious intertidal barnacle

Chamaesipho tasmanica Foster and Anderson in New South Wales. Journal of

ExperimentalMarine Biology and Ecology 252:15-26.

Jeffree, C. E., and E. P. Jeffree. 1996. Redistributionof the potential geographical ranges of

Mistletoe and Colorado Beetle in Europe in response to the temperature componentof

climate change. Functional Ecology 10:562-577.

Jeffree, E. P., and C. E. Jeffree. 1994. Temperature and the biogeographicaldistributions of

species. Functional Ecology 8:640-650.

Jenkins, S. R. 1995. The ecology of sheltered, canopy dominated shores. University of Liverpool.

Jenkins, S. R., S. J. Hawkins, and T. A. Norton. 1999a. Direct and indirect effects of a macroalgal

canopy and limpet grazing in structuring a sheltered inter-tidal community. Marine Ecology Progress Series 188:81-92.

Jenkins, S. R., S. J. Hawkins, and T. A. Norton. 1999b. Interaction between a fucoid canopy and

limpet grazing in structuring a low shore intertidal community. Journal of Experimental

Marine Biology and Ecology 233:41-63.

Jenkins, S. J., T. A. Norton, and S. J. Hawkins. 1999c. Settlement and post-settlementinteractions

between Semibalanusbalanoides (L.) (Crustacea: Cirripedia) and three species of fucoid

canopy algae. Journal of Experimental Marine Biology and Ecology 236: 49-67.

Jenkins, S. R., P. Aaburg, G. Cervin, R. A. Coleman, J. Delany, P. Della Santina, S. J. Hawkins, E.

LaCroix, A. A. Meyers, M. Lindegarth, A.-M. Power, M. F. Roberts, and R. G. Hartnoll.

2000. Spatial and temporal variation in settlement and recruitmentof the intertidal barnacle 114 Semibalanusbalanoides (L) (Crustacea:Cirripedia) over a European scale. Journal of

ExperimentalMarine Biology and Ecology 243:209-225.

Jenkins, S. R., F. Arenas, J. Arrontes, J. Busselt,J. Castro, R. A. Coleman, S. J. Hawkins, S. Kay,

B. Martinez, J. Oliveros, M. F. Roberts, S. Sousa, R. C. Thompson, and R. G. Hartnoll.

2001. European-scaleanalysis of seasonal variability in limpet grazing activity and

microalgal abundance. Marine Ecology Progress Series 211: 193-203.

Jenkins, S. J., R. A. Coleman, P. Della Santina, S. J. Hawkins, M. T. Burrows, and R. G. Hartnoll.

2005. Regional scale differences in the determinism of grazing effects in the rocky

intertidal. Marine Ecology Progress Series 287:77-86.

Jiang, L., and P., J. Morin. 2004. Temperature-dependent interactions explain unexpected

responses to environmentalwarming in communitiesof competitors. Journal of Animal

Ecology 73: 569-576.

Johnson, M. P., M. T. Burrows, R. G. Hartnoll, and S. J. Hawkins. 1997. Spatial structure on

moderately exposed rocky shores: patch scales and the interactions between limpets and

algae. Marine Ecology Progress Series 160:209-215.

Johnson, M. P., M. T. Burrows, R. G. Hartnoll, and S. J. Hawkins. 1998. Individual based

simulations of the direct and indirect effects of limpets on a rocky shore Fucus mosaic.

Marine Ecology Progress Series 169:179-188.

Johnson, M. P., and S. J. Hawkins. 1998. Insights from a general model of herbivores and

macroalgae on rocky shores. Oikos 82: 603-608.

Jones, N. S. 1948. Observationsand experiments on the biology of Patella vulgata at Port St Mary,

Isle of Man. Proceedingsof the Liverpool Biological Society 56:60-77.

Jormalainen,V., T. Honkanen, R. Koivikko, and J. Eranen. 2003. Induction of phlorotannin

production in a brown alga: defence or resource dynamics? Oikos 103:640-650.

Kendall, M. A., P. Williamson, and P. R. Garwood. 1985. Annual variations In the recruitmentof

Sembalanusbalanoides on the North Yorkshire coast 1969-1981.Journal of the Marine

Biological Association, UK 65:1009-1030.

Kendall, M. A., and M. L. Bedford. 1987. Reproduction and recruitment of the barnacle Chthamalus

montagui at Aberystwyth (mid-Wales). Marine Ecology Progress Series 38:305-308.

115 Kendall, M. A., M. T. Burrows, A. J. Southward, and S. J. Hawkins. 2004. Predictingthe effects of

marine climate change on the invertebrate prey of the birds of rocky shores. Ibis 146:40-

47.

Kent, A., S. J. Hawkins, and C. P. Doncaster.2003. Population consequences of mutual attraction

between settling and adult barnacles.Journal of Animal Ecology 72:941-952.

Keough, M. J., and P. T. Raimondi. 1995. Responses of settling invertebrate larvae to bioorganic

films: Effects of different types of films. Journal of Experimental Marine Biology and Ecology 185:235-253.

Klanderud, K. 2005. Climate change effects on species interactions In an alpine plant community.

Journal of Ecology 93: 127-137.

Knight, M., and M. Parke. 1950. A biological study of Fucus vesiculosus L. and Fucus serratus L.

Journal of the Marine Biological Association of the UK 29: 439-514.

Knight-Jones,E. W. 1953. Laboratory experiments on gregariousnessduring settlement in Balanus

balanoides and other barnacles.Journal of Experimental Biology 30:584-598.

Larman, V. N., and P. A. Gabbott. 1975. Settlement of cyprid larvae of Balanus balanoldes and

Elminius modestus induced by extracts of adult barnacles and other marine animals.

Journal of the Marine Biological Association of the UK 55: 183-190.

Leonard, G. H. 2000. Latitudinal variation in species interactions:A test in the New England rocky

intertidal zone. Ecology 81:1015-1030.

Le Tourneux, F., and E. Bourget. 1988. Importanceof physical and biological settlement cues used

at different spatial scales by the larvae of Semibalanus balanoides. Marine Biology 97:57- 66.

Lewis, J. R. 1964. The ecology of rocky shores. English Universities Press, London.

Lewis, J. R. 1986. Latitudinal trends in reproduction, recruitment and population characteristicsof

some rocky littoral molluscs and cirripedes. Hydrobiologia 142:1-13.

Lewis, J. 1996. Coastal benthos and global warming: strategies and problems. Marine Pollution Bulletin 32: 698-700.

Lewis, J. R., and R. S. Bowman. 1975. Local habitat-inducedvariations in the population dynamics

of Patella vulgata L. Journal of Experimental Marine Biology and Ecology 17:165-203.

116 Lewis, J., R. S. Bowman, M. A. Kendall, and P. Williamson. 1982. Some geographicalcomponents

in population dynamics: possibilitiesand realities in some littoral species. Netherlands

Journal of Sea Research 16:18-28.

Little, C., and J. A. Kitching. 1996. The Biology of Rocky Shores. Oxford University Press, Oxford.

Lobban,C. S., and P. J. Harrison.1997. Seaweed ecology and physiology.Cambridge University

Press.

Lubchenco, J. 1983. Littorina and Fucus: Effects of herbivores, substratum heterogeneity, and

plant escapes during succession. Ecology 64:522-533.

Lubchenco,J., and S. D. Gaines. 1981. A unified approach to marine plant-herbivoreinteractions.

1. Populationsand communities.Annual Review of Ecology and Systematics. 12:405-437.

Luning, K. 1990. Seaweeds their environment,biogeography and ecophysiology.John Wiley and

Sons, London.

McCarty, J. P. 2001. Ecological consequencesof recent climate change. Conservation Biology

15:320-331.

McCann, K. S. 2000. The diversity-stabilitydebate. Nature 405:228-233.

Martin, T. E. 2001. Abiotic vs. biotic influenceson habitat selection of coexisting species: climate

change impacts? Ecology 82:175-188.

Meehl, G. A., W. M. Washington,W. D. Collins, J. M. Arblaster, A. Hu, L. E. Buja, W. G. Strand,

and H. Teng. 2005. How much more global warming and sea level rise? Science

307: 1769-1772.

Menge, B. A. 1976. Organization of the New England rocky intertidal community: role of predation,

competition and environmentalheterogeneity. Ecological Monographs46: 355-393.

Menge, B. A. 1978. Predation intensity in a rocky intertidal community: effect of an algal canopy,

wave action and desiccation on predator feeding rates. Oecologia 34: 17.35.

Menge, B. A. 1995. Indirect effects in marine rocky intertidal Interactionwebs: Patterns and

importance. Ecology 65:21-74.

Menge, B. A., and J. P. Sutherland. 1987. Community regulation: variation in disturbance,

competition, and predation in relation to environmental stress and recruitment.The

American Naturalist 130: 730-757.

Miron, G., E. Bourget, and P. Archambault. 1996. Scale of observation and distribution of adult

conspecifics: their influence in assessing passive and active settlement mechanisms in the 117 barnacle Balanus crenatus (Brugiere).Journal of ExperimentalMarine Biology and Ecology 201:137-158.

Moore, H. B. 1936. The biology of Balanus balanoides V. distribution In the Plymouth area. Journal

of the Marine BiologicalAssociation of the UK 20:701-716.

Moore, H. B., and J. A. Kitching. 1939. The biology of Chthamalusstellatus (Poll). Journal of the

Marine BiologicalAssociation of the UK 23:521-541.

Morais, S., D. Boaventura, L. Narciso, P. Re, and S. J. Hawkins. 2003. Gonad development and

fatty acid composition of Patella depressa Pennant (Gastropoda:Prosobranchia)

populations with different patterns of spatial distribution, in exposed and sheltered sites.

Journal of Experimental Marine Biology and Ecology. 294: 61-80.

Moyse, J., and E. Hui. 1981. Avoidance by Balanus balanoides cyprids of settlement on conspecific

adults. Journal of the Marine Biological Association of the U. K. 61:449-460.

Murias Dos Santos, A. E. F. 2000. Intertidalecology of northern Portuguese rocky shores.

University of Southampton,Southampton.

Newell, R. C., V. I. Pye, and M. Ahsanullah. 1971. Factors affecting the feeding rate of the winkle

Littorina littorea. Marine Biology 9: 138-144.

Newell, R. C. 1979. Biology of intertidal animals, Third edition. Marine Ecological Surveys LTD.

O'Connor, N. E., and T. P. Crowe. in press. Biodiversityloss and ecosystem functioning:

distinguishingbetween number and identity of species. Ecology.

Oreskes, N. 2004. The scientific consensus on climate change. Science 306:1686.

O'Riordan, R. M., J. Delany, D. McGrath, A. A. Myers, A. -M. Power, N. F. Ramsey, D. Alvarez, T.

Cruz, F. G. Pannacciulli, P. Range, and G. Relini. 2001. Variation in the sizes of

Chthamalid barnacle post-settlementcyprids on European shores. Marine Ecology 22:307-

322.

Orton, J. H., and A. J. Southward. 1961. Studies on the biology of limpets. IV. The breeding of

Patella depressa Pennant on the north Cornish coast. Journal of the Marine Biological

Association of the UK 41: 653-662.

Orton, J. H., A. J. Southward, and J. M. Dodd. 1956. Studies on the biology of limpets. II. The

breeding of Patella vulgata L. in Britain. Journal of the Marine Biological Associationof the UK 35: 149-176.

Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100: 65-75. 118 Paine, R. T. 1979. Disaster, catastrophe and local persistence of the sea palm Postelsia

palmaeformis. Science 205:685-687.

Paine, R. T. 2002. Trophic control of production in a rocky intertidal community. Science 296: 736-

739.

Pannacciulli, F. G. 1995. Populationecology and genetics of Europeanspecies of intertidal

barnacles. University of Liverpool.

Parmesan, C. 1996. Climate and species range. Nature 382: 765-766.

Parmesan, C., and G. Yohe. 2003. A globally coherent fingerprint of climate impacts across natural

systems. Nature 421:37-42.

Parmesan, C., S. Gaines, L. Gonzalez, D. M. Kaufman, J. Kingsolver, A. T. Peterson, and R. D.

Sagarin. 2005. Empirical perspectives on species borders: from traditional biogeography to

global change. Oikos 108: 58-75.

Pearson, R. G., and T. P. Dawson. 2003. Predictingthe impacts of climate change on the

distribution of species: are bioclimate envelope models useful? Global Ecology and

Biogeography12: 361-371.

Pearson, R. G., and T. P. Dawson. 2005. Long-distanceplant dispersal and habitat fragmentation:

identifying conservationtargets for spatial landscape planning under climate change.

Biological Conservation 123:389-401.

Petchey, 0. L., P. T. McPhearson,T. M. Casey, and P. J. Morin. 1999. Environmentalwarming

alters food-web structure and ecosystem function. Nature 402:69-72.

Petry, P., P. B. Bayley, and D. F. Markle. 2003. Relationships between fish assemblages,

macrophytesand environmentalgradients in the Amazon River floodplain. Journal of Fish

Biology 63:547-579.

Power, M. E., D. Tilman, J. A. Estes, B. A. Menge, W. J. Bond, L. S. Mills, G. Daily, J. C. Castilla, J.

Lubchenco, and R. T. Paine. 1996. Challenges in the quest for keystones. BioScience

46:609-620,

Quinn, G. P., and M. J. Keough. 2002. Experimentaldesign and data analysis for biologists, First Edition. Cambridge University Press.

Raffaelli, D., and S. J. Hawkins. 1999. Intertidal Ecology. Kluwer Academic Publishers,Dordrecht.

Rainbow, P. S. 1984. An introduction to the biology of British littoral barnacles. Field Studies 6: 1- 51.

119 Rittschof, D., E. S. Branscomb, and J. D. Costlow. 1984. Settlement and behaviour in relation to

flow and surface in larval barnacles, Balanus amphitrite Darwin. Journal of Experimental

Marine Biology and Ecology 82: 131-146.

Roberts, M. F. 2002. The ecology and behaviourof Patella species. Universityof Southampton,

Southampton.

Rodriguez-Trelles,F., and M. A. Rodriguez. 1998. Rapid micro-evolutionand loss of chromosomal

diversity in Drosophila in response to climate warming. Evolutionary Ecology 12: 829-838.

Roemmich, D., and J. McGowan. 1995. Climatic warming and the decline of zooplankton in the

California Current. Science 267:1324-1326:

Russell, F. S., A. J. Southward, G. T. Boalch, and E. I. Butler. 1971. Changes in biological

conditions in the English Channel off Plymouth during the last half century. Nature

234: 468-470.

Sagarin, R. D., J. P. Barry, S. E. Gilman, and C. H. Baxter. 1999. Climate-related change in an

intertidal community over short and long time scales. Ecological Monographs 69:465-490.

Schmitz, 0. J., E. Post, C. E. Burns, and K. M. Johnston. 2003. Ecosystem responses to global

climate change: moving beyond colour mapping. BioScience 53: 1199-1205.

Scott, D., and M. Poynter. 1991. Upper temperature limits for trout in New Zealand and climate

change. Hydrobiologia222: 147-151.

Sims, D. W., V. J. Wearmouth, M. J. Genner, A. J. Southward, and S. J. Hawkins. 2004. Low-

temperature-drivenearly spawning migration of a temperate marine fish. Journal of Animal

Ecology 73: 333-341.

Southward,A. J. 1953. Proceedingsof the Liverpool Biological Society 59:1.

Southward,A. J. 1960. On changes of sea temperature in the English Channel. Journal of the

Marine BiologicalAssociation of the U.K. 39:449-458.

Southward, A. J. 1963. The distribution of some plankton animals in the English Channel and

approaches 3 theories about long-term biological changes, including fish. Journal of the

Marine BiologicalAssociation, UK 43: 1-29.

Southward,A. J. 1967. Recent changes in abundance of intertidal barnacles in south-west

England: a possible effect of climatic deterioration.Journal of the Marine Biological Association, UK 47:81-95.

120 Southward,A. J. 1976. On the taxonomic status and distribution of Chthamalusstellatus

(Cirripedia) in the North-EastAtlantic region: with a key to the common intertidal barnacles

of Britain. Journal of the Marine Biological Association of the UK 56:1007.1028.

Southward,A. J. 1980. The western English Channel - an inconsistent ecosystem? Nature 285:361-366.

Southward,A. J. 1991. Forty years of changes in species composition and population density of

barnacles on a rocky shore near Plymouth. Journal of the Marine Biological Association,

UK 71:495-513.

Southward,A. J., and D. J. Crisp. 1954. Recent changes in the distribution of the intertidal

barnacles Chthamalus stellatus Poli and Balanus balanoides in the British Isles. Journal of

Animal Ecology 23: 163-177.

Southward, A. J., and J. H. Orton. 1954. The effects of wave action on the distribution and numbers

of the commoner plants and animals living on the Plymouth Breakwater.Journal of the

Marine Biological Association, UK 33: 1-19.

Southward,A. J., and D. J. Crisp. 1956. Fluctuations in the distribution and abundance of intertidal

barnacles. Journal of the Marine BiologicalAssociation, UK 35:211-229.

Southward,A. J., S. J. Hawkins, and M. T. Burrows. 1995. Seventy years' observationsof changes

in distribution and abundance of zooplankton and intertidal organisms in the western

English Channel in relation to rising sea temperature.Journal of Thermal Biology 20:127-

155.

Southward, A. J., 0. Langmead, N. J. Hardman-Mountford,J. Aiken, G. T. Boalch, P. R. Dando, M.

J. Genner, I. Joint, M. A. Kendall, N. C. Halliday, R. P. Harris, R. Leaper, N. Mieszkowska,

R. D. Pingree, A. J. Richardson,D. W. Sims, T. Smith, A. W. Walne, and S. J. Hawkins.

2005. Long-term oceanographicand ecological research in the Western English Channel.

Advances in Marine Biology 47: 1-105.

Southward, A. J., and E. C. Southward. 1978. Recolonizationof rocky shores In Cornwall after use

of toxic dispersants to clean up the Torrey Canyon spill. Journal of the Fisheries Research

Board of Canada 35:682-706.

Stephenson,T. A., and A. Stephenson. 1972. Life between tidemarks on rocky shores. W. H. Freeman, San Francisco.

121 Stevens, M. J., and G. R. North..1996. Detection of the climate response to the solar cycles.

Journal of the Atmospheric Sciences 53:2594-2608.

Stouffer, R. J., S. Manabe, and K. Ya. Vinnikov. 1994. Model assessment of the role of natural

variability in recent global warming. Nature 367:634-636.

Svensson, C. J., S. R. Jenkins, S. J. Hawkins, and P. Aberg. 2005. Population resistanceto climate

change: modelling the effects of low recruitment in open populations. Oecologia 142:117-

126.

SWCCIP. 2003. Warming to the idea: Meeting the challenge of climate change in the South West.

Taniguchi, H., S. Nakano, and M. Tokeshi. 2003. Influencesof habitat complexity on the diversity

and abundance of epiphytic invertebrateson plants. Freshwater Biology 48:718-728.

Thomas, C. D., A. Cameron, R. E. Green, M. Bakkenes,L. J. Beaumont,Y. C. Collingham, B. F. N.

Erasmus, M. F. de Siqueira, A. Grainger, L. Hannah, L. Hughes, B. Huntley, A. S. van

Jaarsveld, G. F. Midgley, L. Miles, M. A. Ortega-Huerta,A. T. Peterson, O. L. Phillips, and

S. E. Williams. 2004. Extinctionrisk from climate change. Nature 427: 145-148.

Thompson, G. B. 1980. Distributionand population dynamics of the limpet Patella vulgata in Bantry

Bay southwest Ireland. Journal of ExperimentalMarine Biology and Ecology 45:173-218.

Thompson, R. C., B. J. Wilson, M. L. Tobin, A. S. Hill, and S. J. Hawkins. 1996. Biologically

generated habitat provision and diversity of rocky shore organisms at a hierarchy of spatial

scales. Journal of ExperimentalMarine Biology and Ecology 202:73-84.

Thompson, R. C., L. E. Johnson, and S. J. Hawkins. 1997. A method for spatial and temporal

assessment of gastropod grazing intensity in the field: the use of raduia scrapes on wax

surfaces. Journal of ExperimentalMarine Biology and Ecology 218:63-76.

Thompson, R. C., T. A. Norton, and S. J. Hawkins. 1998. The influence of epilithic microbial films between laboratory on the settlement of Semibalanusbalnoides cyprids -a comparison

and field experiments. Hydrobiologia375/376: 203-216.

Thompson, R. C., M. F. Roberts, T. A. Norton, and S. J. Hawkins. 2000. Feast or famine for

intertidal grazing molluscs: a mis-match between seasonal variations in grazing intensity

and the abundance of microbial resources. Hydrobiologia440: 357-367.

Thompson, R. C., T. A. Norton, and S. J. Hawkins. 2004. Physical stress and biological control

regulate the balance between producers and consumers in marine intertidal biofilms.

Ecology 85: 1372-1382.

122 Trenberth, K. E., and J. W. Hurrell. 1994. Decadal atmosphere-oceanvariations In the Pacific.

Climatic Dynamics 9:303-319.

Underwood,A. J. 1979. The ecology of intertidal gastropods.Advances in Marine Biology 16:111-

210.

Underwood,A. J. 1997. Experimentsin Ecology: There logical design and interpretationusing

analysis of variance. Cambridge University Press.

Underwood, A. J. 2000. Experimental ecology of rocky intertidal habitats: what are we learning?

Journal of ExperimentalMarine Biology and Ecology 250:51-76.

Underwood,A. J., and P. Jernakoff. 1981. Effects of Interactionsbetween Algae and Grazing

Gastropods on the Structure of a Low-Shore Intertidal Algal Community. Oecologia 48: 221-

233.

Vadas, R. L., W. A. Wright, and S. L. Miller. 1990. Recruitment of Ascophyllum nodosum: wave

action as a source of mortality. Marine Ecology Progress Series 61: 263-272.

Veiga, A. J., J. Cremades, and I. Barbara. 1998. A catalogue of the marine benthic algae of the

Sisargas Islands (NW Iberian Peninsula, Spain). Boletim do Museo Municipal do Funchal

Supplement No. 5:481-493.

Visbeck, M. H., J. W. Hurrell, L. Polvani, and H. M. Cullen. 2001. The North Atlantic Oscillation:

Past, present and future. Proceedingsof the National Academy of Sciences USA

98:12876-12877.

Wahl, M. 1989. Marine epibiosis. I. Fouling and antifouling; some basic aspects. Marine Ecology

Progress Series 58: 175-189.

A. Menzel, Parmesan, T. J. C. Beebee, J. Fromentin, 0. Walther, G.-R., E. Post, P. Convey, C. -M. Hoegh-Guldberg,and F. Bairlein. 2002. Ecological responses to recent climate change.

Nature 416: 389-395.

Wethey, D. 1983. Geographic limits and local zonation: The barnacles Semibalanus (Balanus) and

Chthamalus in New England. Biological Bulletin 165:330-341.

Wethey, D. S. 1984a. Spatial pattern in barnacle settlement: day to day changes during the

settlement season. Journal of the Marine Biological Association of the UK 64:687-698.

Wethey, D. S. 1984b. Sun and shade mediate competition in the barnacles Chthamalusand

Semibalanus:a field experiment. Biological Bulletin 167:176.185.

123 Wethey, D. S. 1986. Ranking.of settlement cues by barnacle larvae: Influenceof surface contour.

Bulletin of Marine Science 39:393-400.

Winer, B. J. 1971. Statistical principles in experimentaldesign, 2nd edition. McGraw-HillBook

Company.

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

nitrogen retention regulated by consumer pressure and nutrient loading. Limnologyand

Oceanography 45: 339-349.

Yule, A. B., and G. Walker. 1984. The temporary adhesion of barnacle cyprids: effects of some

differing surface characteristics.Journal of the Marine BiologicalAssociation of the UK

64:429-439.

Zacherl, D., S. D. Gaines, and S. I. Lonhart. 2003. The limits to biogeographicaldistributions:

insights from the northward range extension of the marine snail, Kelletia kelletil (Forbes,

1852). Journal of biogeography30: 913-924.

124 Appendix I

Table 1. Analysis of variance for the abundance of limpets found in three treatments: P. vulgata only, P. depressa only and both species together for Kingsand and Andum Point. Variances were homogenous following square root (x + 1) transformation (Cochran's test C= 0.4662, ns).

Source SS DF MS F P F versus Location (Lo) 6.25 1 6.25 84.32 <0.01 Residual Treatment(Tr) 0.73 2 0.37 7.81 0.11 Lo x Tr Location x Treatment 0.09 2 0.05 0.63 0.54 Residual Residual 1.78 24 0.07 Total 8.85 29 SNK tests Location Andurn < Kingsand

20 **

vom,18 0 c 16 14 12 u ö 10 c8 6 4 2 0

® Andum Point o Kingsand

Fig 1. Limpet abundance in treatment enclosures at Andurn Point and Kingsand. " P<0.01.

125 Appendix2

Settlement Factor preparation

Settlement factor was prepared by: 1. grinding 100gwet weight of whole Semibalanusbalanoides with a mortar and pestle. 2.20ml of distilled water was then added to the pestle and ground before being decanted to a separate container. This process was repeated five times to give a final volume of 100ml 3. the liquid was then centrifuged at 12 000 xg for five minutes 4. the supernate was then decanted and boiled for ten minutes 5. the supernatewas then centrifugedfor five minutesat 12 000 xg 6. settlement factor was made up the morning of deployment and painted onto appropriate settlement plates.

126