journal of the marine biological association of the united ... · causeway; la jaonneuse; le...

Sacoglossan gastropods on native andintroduced hosts in Lough Hyne Irelandlarval retention and population asynchrony

cynthia d trowbridge1

colin little2

penny stirling2

and william f farnham3

1Department of Zoology Oregon State University PO Box 1995 Newport OR 97365 USA 2Beggars Knoll Long River RoadNewtown Westbury BA13 3ED UK 3Institute of Marine Sciences University of Portsmouth Portsmouth PO4 9LY UK

The north-eastern Atlantic sacoglossan gastropod Elysia viridis was studied on littoral and sublittoral shores of the BritishIsles from 2001 to 2007 to evaluate its potential role in the known decline of the invasive green alga Codium fragile sspfragile Across its European range this sacoglossan associates with eight genera of algal hosts within three algal ordersand two divisions these hosts include native and introduced macroalgae The lsquospecialistrsquo herbivore was investigated primar-ily within the rocky shore community at Lough Hyne Marine Reserve County Cork south-west Ireland with comparativeinformation from surveys of 95 other sites in the British Isles In Lough Hyne E viridis associated with the green algaeCodium Cladophora and Chaetomorpha as well as the red alga Griffithsia The sacoglossan associated with CodiumCladophora and Chaetomorpha as well as the red alga Halurus in Devon Isle of Wight and the Channel IslandsRecruitment of E viridis to Codium spp in the lough was substantially higher than in most areas of the British Isleswith 100 of the Codium hosts attacked during annual September surveys The strong autumn pulse of sacoglossan recruit-ment coupled with the asynchronous population dynamics compared to other shores indicates retention of planktotrophiclarvae within the lough The previously reported decline in sacoglossan density with tidal level was hypothesized to be causedby sublittoral predators Although our experiments failed to demonstrate an effect by large predators the foraging of theabundant fingerling fish and small crabs could not be tested conclusively Field experiments were consistent with theinterpretation of intense sacoglossan herbivory in certain areas of the lough However even if the sacoglossans contributeto the local decline of C fragile in the lough they cannot account for the regional decline of the alga on the north-easternAtlantic shores

Keywords sacoglossan gastropods native and intoduced hosts Lough Hyne Ireland larval retention population asynchrony

Submitted 30 October 2007 accepted 15 February 2008

I N T R O D U C T I O N

Seaweed incursions are proliferating on marine and estuarineshores around the world (reviewed by Trowbridge 2006)Siphonaceous green seaweeds including Caulerpa taxifolia(M Vahl) C Agardh C racemosa (Forsskal) J Agardh C bra-chypus (Harvey) and Codium fragile (Suringar) Hariot sspfragile (name change by Provan et al 2007) are among the eco-logically more important invaders One of the frequently citedexplanations for the ecological success of invasives is the escapefrom their native consumers parasites and pathogens intolsquoenemy-free spacersquo Ironically few studies have evaluated therole of consumers in the native range of invasive marinealgae (but see Trowbridge et al 2008) Furthermore whilemost studies on invasives focus on the establishment andexpansion phases in new incursions few have investigated thepatterns of and causal factors contributing to the subsequentdecline of the exotics In particular does consumer accumu-lation on exotics contribute to the decline of introduced algae

The ecology of the Asian Codium fragile ssp fragile wasrecently reviewed (under its previous name C fragile ssptomentosoides (Van Goor) Silva) but the proliferation andsubsequent current decline of the alga within Lough Hynehave never been investigated comprehensively Ecologicalinteractions between the generalist purple urchinParacentrotus lividus (Lamarck 1816) and C fragile havebeen examined experimentally (Kitching amp Thain 1983)The recent urchin decline within the lough (Barnes et al2002) may have contributed to the algarsquos decline becauseurchin herbivory facilitated the invasive species (Kitching ampThain 1983) which is a relatively poor competitor(Trowbridge 1998)

The population ecology of marine specialist herbivores onsympatric introduced and native algal hosts has been investi-gated in only a few cases compared with their terrestrialcounterparts The sacoglossan mollusc Elysia viridis(Montagu 1804) provides the opportunity to examine howdifferential host preferences could lead to significant effectson the introduced host without apparently harming a nativecongener Across its European range this species of slugassociates with eight genera of algal hosts within three algalorders and two divisions these hosts include native and intro-duced macroalgae

Corresponding authorCD TrowbridgeEmail trowbriconidorstedu

771

Journal of the Marine Biological Association of the United Kingdom 2008 88(4) 771ndash782 2008 Marine Biological Association of the United Kingdomdoi101017S0025315408001690 Printed in the United Kingdom

httpsdoiorg101017S0025315408001690Downloaded from httpswwwcambridgeorgcore Open University Library on 30 Jan 2017 at 101018 subject to the Cambridge Core terms of use available at httpswwwcambridgeorgcoreterms

In Lough Hyne a marine lough in south-west Ireland Elysiaviridis associates with several green algae including the nativeCodium vermilara (Olivi) Delle Chiaje and an introduced con-gener Codium fragile The populations of C vermilara havestayed relatively constant over many years but those ofC fragile have fluctuated wildly (Little et al 1992 Little per-sonal observation) It is likely that the introduced species firstbegan to spread within the lough in 1937 though at the timeit was referred to as C tomentosum Stackhouse (Renouf1939) (see Silva 1955 for clarifying the taxonomic confusion)(Codium tomentosum (sensu stricto) does not occur in thelough) The proliferation of C fragilewithin the lough occurredover the next 50 years but subsequent decline has reducedpopulations to a low level (Trowbridge et al unpublisheddata) Field observations in Scotland (Trowbridge 2002)suggested that herbivory by high densities of sacoglossanscould lead to the local decline of Codium populations

The aims in the work reported here were (1) to examinewhether Elysia viridis could potentially be a major consumerof the invasive alga in Lough Hyne (2) to quantify the patternsof grazing by Elysia on native and introduced Codium sppand (3) to evaluate the role of predators in limiting the popu-lations of Elysia and thus in influencing their effects on thealgal hosts

M A T E R I A L S A N D M E T H O D S

Study areaLough Hyne is a small (~1 km2) fully marine lough in CountyCork south-west Ireland Locations around the lough areidentified by numbers (see Figure 1) referred to as Renoufsectors (after Renouf 1931) The lough is connected to thenorth-east Atlantic Ocean via narrow tidal rapids (Figure 1)The narrow channel (~20 m wide) and the raised rocky sillrestrict the water flow exiting the lough during ebb tide Thelough has an asymmetrical semi-diurnal tidal cycle with~85 hours for tidal ebb and ~4 hours for tidal flood Themaximal tidal flow is ~3 msecond and the turbulent flowthrough the rapids kills many delicate invertebrate larvae(Bassindale et al 1948 Jessopp 2007) Because of therestricted seawater exchange the flushing time estimated forthe lough is 41 days (Johnson et al 1995) This period islonger than the larval period of Elysia viridis (Trowbridge2000 Trowbridge amp Todd 2001)

Sacoglossan populationsFrom 2001 to 2007 thalli of Codium spp were collected inSeptember at Lough Hyne and brought back to the laboratoryto examine Sacoglossans were manually removed (primarilyElysia viridis) with pipettes and forceps from several litres ofalgae each September The small size of the sacoglossans(mostly 5 mm and 1 mg) prevented us from weighingindividual animals therefore specimens were measured tothe nearest millimetre while crawling across a Petri dishwith clean seawater Based on a previous culturing study onE viridis specimens 5 mm would have settled and meta-morphosed in the previous 1ndash3 weeks (Trowbridge 2000)The algae were blotted and the wet weight measured to thenearest gram In 2003 Codium was dried in a microwave todetermine the slug abundance based on algal dry weight

The strong annual pulse of recruitment of Elysia viridis hadnever before been witnessed by the first author during earlierScottish and Irish work (1996ndash2002) Therefore during thecourse of surveying 95 shores of the British Isles for Codium(2001ndash2007) all E viridis encountered were measured andthe algal host species were recorded These sites included 41shores in County Cork 2 in County Clare 1 in NorthernIreland 13 in south-eastern Scotland 12 in southernEngland and 26 shores in the Channel Islands (Table 1)

The shores were surveyed from high-shore pools to low-shore emergent substratum for Codium spp All otherknown hosts of Elysia viridis were examined in the course ofthis survey all specimens of E viridis that were found werecollected except when local laws prohibited collection (egDunny Cove Ireland) Because of the slugsrsquo generally largesize specimens were gently blotted on paper towels andthen weighed to the nearest milligram

Radular teeth were dissected from a large number of speci-mens from each algal host and the tissues dissolved in 10Clorox bleach The cleaned radular ribbon was photographedwith a digital microscope and the blade-shaped teeth werecategorized as rounded-tips versus curved hook-like tipsMore detailed analyses of the radular teeth (eg length andwidth of the leading radular tooth and the number of ascend-ing and descending teeth) are being presented elsewhere Ourintention here was to evaluate whether tooth-shape per secould have precluded certain host changes within LoughHyne Tooth size and shape are known to be induced byalgal diet in this species (Jensen 1989 1993) Some

Fig 1 Map of the permanently established shore sectors as designated byRenouf (1931) at Lough Hyne County Cork Ireland Base map from Myerset al (1991) Individual grid units represent 100 m Symbols indicate theblocks of replicates of Codium transplant experiments

772 cynthia d trowbridge et al


sacoglossans die in the presence of algal food that conspecificsacoglossans can consume the basis of this trophic constraintmay be the inability to puncture cell walls of host algae withcertain tooth shapes Trophic induction of shape of newly pro-duced teeth occurs not only in sacoglossans (Jensen 19891993) but also in littorine snails (Padilla 1998)

Recruitment experimentsA series of field transplant experiments were conducted withCodium in Lough Hyne Because C fragile ssp fragile is aninvasive pest in some geographical regions (Trowbridge1998) and because Lough Hyne is a marine reserve Codium

Table 1 Spatial and temporal details of Codium surveys in the British Isles

Region County Island or District Sampling dates No of sites Site locations

Northern ChannelDevon April 2004 6 Grande Parade Walkway Plymouth Mount Batten

Jennycliff Bay Wembury Torquay TorbayThurlestone Bigbury Bay Brixham Bigbury Bay

Dorset May 2003 1 Chesil FleetIOW May 2003 May amp

September 20054 BembridgeLifeboat Station North Horse Ledge

Whitecliff Ledge Yellow LedgeSussex September 2002 1 Pagham Harbour

Southern ChannelGuernsey September 2002 May

2004 amp June 20059 Albecq Bay Bordeaux Grande Rocques Lihou

Causeway La Jaonneuse Le Catioroc Perelle BayMoulin Huet Port Soif Site 3 near Albecq

Jersey May 2003 amp May 2004 10 Corbiere Fliquet Bay Giffard Bay Greve de LecqLrsquoArchirondel La Rocque Queen Elizabeth CastlePortelet Harbour Rozel Harbour Sauchet Bay

Alderney June 2005 7 Braye Bay Catrsquos Bay Clonque Bay Crabby Bay SalineBay Saye Bay South St Esquere Bay

IrelandCounty Cork AugustndashSeptember

2002 April Augustamp September 2003September 2005

41 Knockadoon Youghal Bay Ballycotton south-westernside of Lifeboat Station West Channel of LoughMahon Curraghbinny Pier Myrtleville RobertrsquosCove Carrigadda Bay Nohaval Cove OysterhavenCoast Guard Slip Oysterhaven Ballinchashel CreekJarleyrsquos Cove Kinsale Harbour Sandy Cove BullenrsquosBay Quay Old Head of Kinsale Lispatrick Lowersouth of White Strand Coolmain Bay east side ofparking area Broadstrand Bay north sideGarrettstown Strand Old Kinsale Head InchydoneyBeach Clonakilty Bay Duneen Point Duneen BayDunnycove Point Red Strand Dirk Bay Dirk CoveDirk Bay east side of Galley Bay Little Island StrandRosscarbery Bay Owenahinchy Strand RosscarberyBay Pouladov East Pier Mill Cove Tralong Bay westside Tracarta Strand Flea Sound Toehead BayTragumna Bay near Drishane Island TralispeanTranabo Skull Harbour Roaringwater Bay CollaQuay Long Island Channel Roaringwater Bay AltarWedge Tomb Toormore Bay Galley Cove nearCrookhaven Gortnakilly Pier Collack Bay BantryBay Bantry Bay Seafoods Bantry Bay WP 283Zetland Pier near Adrigole Bantry Bay Serragh RockSullivan near Adrigole Bantry Bay Dunboy AnglingSite near Castletownbere Bantry Bay

County Clare September 2004 2 Finavarra Spanish Point North Miltown MalbayNorthern Ireland

County Down September 2001 1 The Narrows Strangford LoughScotland

Fife April 2003 2 Doo Craigs St Andrews Castle Rocks St AndrewsEast Lothian amp Borders April 2003 11 Milsey Bay North Berwick Wrecked Craigs Cockenzie

and Port Seton Hummell Rocks Gullane BentsGullane North Berwick Golf Club (2 sites) NorthBerwick Bay Seabird Centre Rocks North BerwickBay Belhaven Belhaven Bay Dunbar East BeachJohn Muir Way Dunbar Bayswell north of DunbarCastle Dunbar Longskelly Point Yellow Craigs Brigsof Fidra St Baldredrsquos Cradle Firth of Forth

elysia viridis in lough hyne 773


vermilara was used as a surrogate for C fragile to prevent theintroduced alga from spreading beyond its current distri-bution in the lough We presumed that this substitution wasjustified as juvenile sacoglossans were similar in size andabundance on the two species (Trowbridge 2004 this study)

In September 2004 specimens of defaunated weighedCodium vermilara were transplanted to three differentlocations (blocks) on the western shore W1415 W1617and W24 (Figure 1) At each location ~44 g (wet weight) ofC vermilara were attached to a plastic mesh bag (mesh sizeapproximately 17 38 mm) and each tied to a rock slab toensure samples remained on the benthos Half the algalthalli were attached inside the mesh bags to exclude large pre-dators the other half were outside the bags (N frac14 5 replicatesper treatment per block N frac14 30 transplants) Specimens wereretrieved in September 2005 and reweighed All sacoglossanswere removed counted and measured from each sample

In September 2005 Codium vermilara was transplanted toeight sites (blocks) around Lough Hyne in Renouf sectorsE45 E1516 I78 S34 S1011 W1415 W2425 and N1(Figure 1) ~24 g (wet weight) of Codium were attached toeach mesh bag which was tied with monofilament line to arock slab All algal transplants were attached outside themesh bags because there were no significant differences inalgal biomass or slug abundance inside versus outside bagsin the 2004ndash2005 experiment Transplants were collected inSeptember 2006 the remaining algae weighed and thenumber of Elysia viridis counted Given that E viridis iscommon on Codium spp on Scottish and Irish shores (includ-ing Lough Hyne and environs) (Trowbridge amp Todd 19992001 Trowbridge 2000 2001 2002 2004) slug recruitmentwas predicted to be moderate or high within Lough Hynedue to wave-sheltered conditions and potential retention oflarvae

To quantify the spatial pattern in sacoglossan recruitmenton a short time-scale (days) fronds of Codium vermilara weretransplanted in September 2006 to four blocks within LoughHyne in Renouf sectors N1 N11 W15 and W25 (Figure 1)Cable-ties were used to attach Codium (~10 g fronds) to thebrown algae Himanthalia elongata (Linnaeus) SF GrayFucus serratus Linnaeus and Cystoseira foeniculacea(Linnaeus) Greville replication was 20ndash30 Codium frondsfor each block At N1 transplanted Codium fronds includedthose with simple morphology as well as those with acomplex lsquowitch-broomrsquo morphology (due to the productionof extensive short adventitious branches) The recruitmentexperiment lasted 4ndash6 days Codium fronds were collectedthe remaining algae weighed and the slugs counted andmeasured

In September 2007 another short-term recruitment exper-iment was conducted in the North Basin four blocks (N12N89 N11 and E45) with 25ndash35 replicate transplants perblock Codium vermilara fronds (~3 g each) were transplantedfor 7 days as described above Fronds were collected theremaining algae weighed and the slugs counted and weighed

Codium vermilara was transplanted at three tidal levels alonga sloping bench in the South Basin (Renouf sector I17) Theupper transplants were attached to Ascophyllum nodosumthalli at the upper end of the fucoidrsquos tidal range the middletransplants were attached to the low-shore alga Himanthaliaelongata and the lower transplants were attached to theshallow sublittoral Cystoseira foeniculacea Based on relativetidal level the experimental groups were called mid-littoral

low-littoral and sublittoral respectively The experiment lasted4 days in September 2007 Transplants were collected and exam-ined as before

Feeding preference experimentsA series of feeding experiments was conducted to evaluate thepreferences of Elysia viridis for various seaweeds Two exper-iments conducted in 2002 (pairwise choice of Codium vermi-lara versus C fragile and C tomentosum versus C fragile)were already reported by Trowbridge (2004) In 2004 and2005 individual slugs were placed in plastic cups (~300 ml)with seawater and small amounts of algae There were threeseparate treatments (1) Codium vermilara only (2) alterna-tive alga only or (3) a pairwise choice of Codium and thealternative host In 2004 the alternative host wasCladophora rupestris (Linnaeus) Kutzing and in 2005 it wasthe red alga Griffithsia corallinoides (Linnaeus) TrevisanExperiments were conducted at ambient light and room temp-erature conditions Although the slugs derive energetic benefitfrom algal chloroplasts after several weeks of feeding(Trowbridge 2000) host preference per se is presumablyunrelated to local environmental conditions Experimentswere 1ndash2 days in length with the location of each slug mon-itored periodically through the experiment (Past experimen-tal work with E viridis indicated that short-term preferenceswere maintained over longer time scales of weeks to months(Trowbridge amp Todd 2001)) Categorical data were analysedfor the last census using Chi-square or Fisherrsquos exact testsdepending on the replication Algae were examined forgrazing damage at the end of each experiment as an additionalindicator of feeding

Predation experimentsElysia viridis was substantially more abundant on infralittoralalgal hosts than on sublittoral ones in Lough Hyne(Trowbridge 2004) One possible explanation was thatintense sublittoral predation by fish andor carnivorousinvertebrates could reduce densities of juvenile sacoglossansThis hypothesis was evaluated in two different ways Firsta predator-exclusion experiment was conducted in the field(2004ndash2005 transplant experiment previously described)The large mesh size (17 38 mm) excluded large fish andcrabs smaller predators could readily enter or reach into themesh bags

Second a series of short-term predation trials were con-ducted with invertebrate carnivores and omnivores A smallpredator was placed in a plastic container with seawaterversus no predator in the negative controls container sizevaried from one to several litres depending on predatorsize Ten juvenile E viridis were added per container(20 with prawns) slugs remaining after 1ndash2 days werecounted Replication was 3ndash6 containers per experimentaland no-predator control treatment In September 2005 and2006 we tested 3 genera of juvenile seastars (Asterias sppLinnaeus 1758 Marthasterias glacialis (Linnaeus 1758)Asterina gibbosa (Pennant 1777)) the nemertean Lineus long-issimus (Gunnerus 1770) juvenile prawns (Palaemon serratus(Pennant 1777)) and the green crab Carcinus maenasLinnaeus 1758



R E S U L T S

Sacoglossan populationsThe herbivore load of Elysia viridis on thalli of Codium spp inthe main body of Lough Hyne was quite variable rangingfrom 002 to 419 slugs per gram of algae (Table 2) In largepart this variation was caused by pulses of high sacoglossanrecruitment In September 2001ndash2007 there was a highdensity of juvenile slugs on the western shore of LoughHyne but slugs between 2 and 5 mm in length formed thebulk of the population in September in most years(Figure 2) The sizendashfrequency of sacoglossans based on col-lections made in September 2001 and 2003 (Figure 2) showedno significant difference in body length between the two sym-patric congeneric hosts C vermilara and C fragile (x2 frac14 3305 df P frac14 0654) The September recruitment pulse showedsubtle but highly significant interannual differences (x2 frac14587 15 df P 0001) In all cases slugs varied between 10and 85 mm long with modal sizes between 2ndash4 mm Noadult E viridis specimens were seen on Codium in the loughin September 2001ndash2007 despite extensive snorkel surveys

In the high current flow of the Rapids Elysia viridisoccurred on the filamentous green algae ChaetomorphaKutzing and Cladophora Kutzing The sacoglossans were notonly much more abundant (in number per gram algae) inthe Rapids than on the wave-sheltered west shore (W15 andW25) but also more variable in size For example slug abun-dance varied from 43 per gram algae in the Rapids to 01 onthe west shore (Figure 3) and individuals collected on thesame day were substantially smaller on Codium vermilara(maximum 4 mm) and Cladophora rupestris (maximum10 mm) in the lough than on hosts in the Rapids (Figure 3)In 2007 the population on Cladophora in the Rapids wasagain composed of a wide size-range of individuals (recentlysettled juveniles to those 20 mm) Large individuals weredisproportionately more frequent in 2007 than in 2006 (like-lihood ratio Chi-square G frac14 116 3 df P frac14 0009)

Radular tooth shape was highly variable All specimensinvestigated from Bryopsis many from Codium vermilaraand most from C fragile had blade-shaped teeth withrounded tips or with the currently used lsquoleading toothrsquopointed and hook-like (Table 3) Individuals from theseptate algae Cladophora and Griffithsia (Table 3) also hadrounded-tip teeth In contrast 89 of the conspecifics from

C tomentosum had hook-like teeth (Table 3) Some individ-uals had two types of teeth in the radulae particularly speci-mens collected from C vermilara or C fragile Someindividuals exhibited tooth-shape variation within thesection of the radula with newly formed unused teeth otherindividuals exhibited variation in the section with old dis-carded teeth still other slugs exhibited variation betweensections of the radulae

Marked asynchronyThe size of sacoglossans in Lough Hyne was substantially differ-ent from other locations in the British Isles This observedgeneral difference was based on five sets of observations (1)In September 2001 most Elysia viridis in Lough Hyne were5 mm long (Figure 2) In contrast on 11 September 2001large E viridis (several cm long) were found in StrangfordLough Northern Ireland on various taxa of red algae becauseof world events no quantitative data were collected (2) InSeptember 2002 the mean sacoglossan size on GuernseyChannel Islands was 12 mmmdashsubstantially larger than conspe-cifics in Lough Hyne For example when Lough Hyne Codiumslugs averaged ~3 mm (1 mg) Guernsey conspecifics onC tomentosum and Bryopsis plumosa were 1ndash2 orders ofmagnitude heavier (Figure 4A) (3) In April 2003 slug sizewas significantly larger in the lough than outside at KinsaleHead County Cork (Figure 4B) (4) In August 2003 smalladult slugs (10 mm) with egg masses were recorded from

Fig 2 Sizendashfrequency distribution of Elysia viridis in September 2001ndash2005on Codium spp on the western shore of Lough Hyne County Cork Black andgrey bars indicate thalli of Codium fragile and C vermilara collected from theshore hatched bars indicate thalli of C vermilara collected from a 1-yeartransplant experiment

Table 2 Sacoglossan abundance in Lough Hyne The abbreviation lsquoexprsquoand term lsquowildrsquo indicate collections from experimentally deployed versus

established wild thalli respectively

Year Algal host No slugsgramwet weight

Location(Renouf sectors)

2003 Codium vermilara 020 W382004 Codium vermilara 019 (005ndash041) W17182005 Codium vermilara 018 W1415 (exp)2005 Codium vermilara 002 W24 (exp)2005 Codium vermilara 013 W24 (wild)2006 Codium vermilara 019ndash419 W1415 W2425

N1 E45 N11and S34 (exp)

2007 Codium vermilara 016ndash099 N12 N89 N11 E45 and I17 (exp)



Dunny Cove County Cork (specimens were not measured asthe site is a no-take reserve) the following week sacoglossansin the lough were 5 mm (Figure 2) (5) In September 2004a wide range of slug sizes (2ndash22 mm) was found atFinavarra County Clare with a mean value of 84 mm thesewest-coast sacoglossans were significantly larger than thelough conspecifics (Figure 4D Fisherrsquos exact test P 0001)The results collectively indicate considerable temporal variationas well as host-specific variation in slug phenology Given thatthorough surveys of Codium spp were made on 41 other shoresin County Cork (Table 1) and that juvenile sacoglossans of

other species were found (eg Limapontia Johnston 1836)the population dynamics of E viridis inside Lough Hynewere considered asynchronous with other conspecific popu-lations in the British Isles

High but variable recruitmentSacoglossan recruitment varied significantly along the westernshore of Lough Hyne In September 2003 comparisons of thesizendashfrequency distribution of Elysia viridis differed signifi-cantly between Renouf sectors W18 and W38 (x2 frac14 2064 df P 0001) Slugs in W18 were slightly larger (~1 mm)indicative of earlier recruitment andor faster juvenilegrowth Furthermore there was a gradient in abundance ofrecent recruits along the western shore

In the Codium-transplant experiment deployed in threeblocks on the western shore peak sacoglossan recruitment inSeptember 2005 occurred in the northern-most block(Figure 5) In a stepwise general linear model block treatmentand final algal mass collectively accounted for 51 of the vari-ation in sacoglossan abundance yet only the block effect wasstatistically significant (F frac14 797 2 df P frac14 0003) To simplifythe experimental results the non-significant treatment effect(inside versus outside mesh bag) was dropped the ensuinganalysis indicated a significant difference in the percentageof Codium mass persisting in the different blocks (Figure 5AKruskalndashWallis H frac14 80 P frac14 0018) Codium grew the bestat the southern end of the western shore (W24) and sacoglos-san density (numberg) was low there slug density was onaverage an order of magnitude higher on the more northerntransplants (W1415) Although these latter results were notstatistically significant (Figure 5B KruskalndashWallis H frac14 47P frac14 0095) slug and algal abundance were inversely relatedFinally there was no significant difference in sacoglossanlength on experimental (transplanted) versus unmanipulatedcontrol (lsquowildrsquo) thalli (x2 frac14 636 5 df P frac14 0273) collected atthe same site although there may have been a reduction inslug density (not quantified)

In the second experiment 40 transplants were deployedand all but 2 were recovered (Figure 5C) The proportion ofCodium mass remaining after 1 year varied significantlyamong the 8 blocks (KruskalndashWallis H frac14 303 P 0001)Of the 19 remaining transplants (ie those thalli not 100eliminated) slug abundance tended to vary among the6 blocks (Figure 5D KruskalndashWallis H frac14 99 P frac14 0079)When blocks were combined (due to low sample size) slug

Fig 3 Sizendashfrequency distribution of Elysia viridis collected from fourdifferent algal hosts (Cladophora sp Cladophora rupestris Chaetomorphasp and Codium vermilara) on 17 September 2006 The collection locations(Rapids or Renouf sectors) and the abundance of slugs (number per gramwet weight) are indicated

Table 3 Percentages of differentially shaped radular teeth in specimens of Elysia viridis from different algal hosts lsquoMixedrsquo refers to radulae containingblade-shaped teeth with rounded tips versus pointed hook-like tips N refers to number of radulae examined

Hosts Algal construction Roundedtips only

Leading toothrounded tip mixed

Leading tooth pointedand hooked mixed

Pointed andhooked teeth only

N

Green Algal HostsBryopsis Siphonous 100 5Codium vermilara Siphonous 33 7 33 27 15Codium tomentosum Siphonous 11 89 18Codium fragile Siphonous 69 22 6 63 32Cladophora Septate 92 8 11

Red Algal HostsGriffithsia Septate 67 33 3Unidentified Ceramiales Septate 100 4

Dutch specimens courtesy of P Van Bragt



Fig 4 Body size of Elysia viridis in different areas of the British Isles on different algal hosts Codium vermilara C tomentosum Bryopsis plumosa andChaetomorpha sp The rectangles in the box-plots (AndashC E) depict data between 25th and 75th percentiles error bars indicate 10th and 90th percentilesblack dots represent outliers solid horizontal line in each bar indicates median and dashed line indicates mean In September 2004 a balance was notavailable so body length histograms are presented (D) for Lough Hyne versus open-coast (Finavarra County Clare) sites For comparative purposes LoughHyne data for September 2002 are shown as 1 mg (not actually weighed due to small size see text)

Fig 5 Recruitment of Elysia viridis to experimental transplants of Codium vermilara in two long-term experiments (AndashB) September 2004ndash2005 in three blockson the western shore of Lough Hyne and (CndashD) September 2005ndash2006 in eight blocks around the lough The percentage mass change (growth and loss) of Codiumtransplants is illustrated in upper panels sacoglossan abundance is shown in the lower panels The numbers above each bar indicate the number of replicates errorbars indicate 1+ SE



abundance varied strongly among shores (KruskalndashWallisH frac14 91 P frac14 0028) with highest density on the north-westshore (sector N1)

In 2006 recruitment rates varied from 02ndash06 slugsgduring the 4 to 6-day experiment (Figure 6A) Slug abundancevalues were comparable to those on lsquowildrsquo established Codiumvermilara near transplants showing that there was no bias dueto transplant methodology andor algal stress However theremay have been some bias due to day-to-day variation becausetransplants in different blocks could not be collected on thesame day due to severe weather and waves The small sizeof recruits (Figure 6B) indicates they were probably larval set-tlers not benthic slugs relocating

The two recruitment experiments in 2007 also yielded highrates of sacoglossan recruitment In the 1-week North Basinexperiment recruitment averaged about 01 to 10 Elysia pergram of Codium The rock promontory in N11 had thehighest mean densities of the four blocks tested (Figure 6C)Interference from members of the public caused damage orloss of most of the replicates in block N12 and many ofthose in N11 we recovered only 20 and 63 of the replicatesrespectively after 1 week The use of large numbers of repli-cates partially offset the severity of the damage except in N12

Sacoglossan recruitment to transplants varied with tidallevel but not as predicted (Figure 6D) Most of the Codiumfronds transplanted into the mid-zone and some of those inthe low-zone became flaccid and desiccation-stressed duringthe 4-day experiment that coincided with unusually warmweather Slug recruitment varied significantly with tidal level(KruskalndashWallis H frac14 169 P 0001) Either no Elysia

recruited to the upper transplants or none survived the con-ditions during exposure (Figure 6D) At the other two tidallevels there was no significant difference between densitieson the low littoral and shallow sublittoral transplants (~02slugs per gram Codium Figure 6D)

Variable host use feeding and effectson the hostsIn pairwise-choice and no-choice experiments juvenile Elysiaviridis from Codium strongly preferred Codium vermilara toCladophora rupestris In no-choice situations only ~20 ofslugs even approached the unfamiliar hosts no slugs did soin choice-situations (Figure 7A) Sacoglossan response toalgal hosts did not vary significantly between choice andno-choice treatments (Fisherrsquos exact test P frac14 0367) At theend of the 2-day experiment sacoglossans in the two treat-ments with Codium present were significantly larger thanthose with just Cladophora present (G-test G frac14 110 4 dfP frac14 0026) Because the individuals had not been measuredat the outset it was not possible to rule out the possibility ofinitial difference We re-measured individuals on Cladophoraafter an additional 8 days and found no significant sluggrowth on that alga (G frac14 15 2 df P frac14 0479)

In 2004 we found a few thalli of Griffithsia corallinoides onthe south shore of Lough Hyne with numerous juvenileE viridis on the alga In 2005 juveniles collected fromCodium vermilara strongly preferred Codium to Griffithsiahowever a high percentage crawled on the red alga and

Fig 6 Short-term recruitment rates (A) and body size (B) of Elysia viridis on Codium vermilara fronds outplanted for 4ndash6 days in shallow subtidal areas of LoughHyne in 2006 Frond morphology is indicated simple (either unbranched or branched dichotomously) versus complex (extensive adventitious branches oflsquowitch-broomrsquo shape) Transplant locations refer to Renouf sectors Recruitment experiments conducted for 1 week (C) and 4 days (D) in September 2007Numbers above each bar indicate the replication whereas letters indicate which groups differed significantly (similar letters ns different letters significantdifference)



attempted to feed when alternative hosts were not available(Figure 7B) There was a statistically significant differentresponse in E viridis in choice versus no-choice situations(Fisherrsquos exact test P 0050) The reciprocal experimentwas not possible due to low abundance of E viridis onGriffithsia in the lough and the paucity of the alga

Role of predationNone of the invertebrate predators tested significantly reducedsacoglossan densities in laboratory trials In fact the only pre-dators that consumed any Elysia at all were a few individualsof green crab (Carcinus maenas) even this predator speciesdid not significantly reduce slug numbers (MannndashWhitneyU-test U frac14 45 1 df P frac14 0165) On the shore the 2004ndash2005 Codium transplant experiment inside versus outsidemesh cages did not have a significant treatment effect(ANOVA F128 frac14 293 P frac14 0098) demonstrating that largepredators (ones excluded by 38 mm mesh) were probablynot ecologically significant in controlling slug densities

D I S C U S S I O N

Sacoglossan populationsThe abundance of Elysia viridis in Lough Hyne was high com-pared to most other north-east Atlantic shores that have been

surveyed (Trowbridge amp Todd 1999 2001 Trowbridge 20012002 Trowbridge amp Farnham 2004 Trowbridge et al 2004)A few other lsquohot-spotsrsquo in the British Isles include (1) ObanArgyll in western Scotland where sacoglossans occurred onup to 100 of the intertidal thalli of Codium fragile anddecimated the local algal population (Trowbridge 2002)(2) Thurlestone in Devon (this study) (3) Clachan Seil tidalrapids Argyll (Trowbridge amp Todd 2001) and (4) SpanishRapids Loch Maddy North Uist in the Outer Hebrides(Trowbridge unpublished data) In most other locations sur-veyed by the first author in the British Isles E viridis was fre-quently observed but not at the high densities documented atLough Hyne

On Australasian shores high densities of sacoglossansperiodically caused grazing damage and thallus loss toCodium fragile (reviewed by Trowbridge 1998 2004) Highdensities of Placida dendritica (Alder amp Hancock 1843)have been noted for Clare Island County Mayo Ireland(Colgan 1911) and Lane County and Lincoln CountyOregon USA (Trowbridge 1992a b) When abundantP dendritica causes extensive damage to differentiated andundifferentiated Codium hosts including branch loss andthallus fragmentation (Trowbridge 1992a 1993 1998)

One potential reason for the high abundance of Elysia viridisin Lough Hyne may be that the semi-enclosed nature of thelough may retain the veliger larvae Recent studies haveshown increased local abundance of plankton and larval reten-tion in areas with restricted or limited water exchange (Gaines ampBertness 1992 Ballard amp Myers 2000 Greenwood et al 2001Johnson amp Costello 2002 Rawlinson et al 2005 and referencestherein) The flushing rate of seawater in Lough Hyne wasestimated to range between 125 and 41 days (Johnson et al1995 Jessopp amp McAllen 2007) Based on previous culturework on E viridis the minimum larval period (from hatchingto metamorphosis) was 28 days at 158C (Trowbridge 2000)The slightly warmer seawater in Lough Hyne (16ndash188C inAugust and September) could lead to an even shorter larvalperiod Such slow flushing rates in Lough Hyne could limitlarval dispersal of sacoglossans (and other taxa) Thus thehigh settlement rates and juvenile densities of Elysia viridismay reflect larval retention in the lough

A second complementary explanation is that the high pro-ductivity due to enhanced phytoplankton densities andwarmer seawater in the lough (Johnson amp Costello 2002)could enhance the survival and the growth rate of larval saco-glossans Although veliger larvae of E viridis are known tofeed on the cryptophyte Rhodomonas Karsten 1898 inculture studies (Trowbridge 2000 Trowbridge amp Todd2001) it is not known the extent to which sacoglossanlarvae would feed on the dinoflagellates that dominate thesummer phytoplankton assemblage These two mechanisms(larval retention and phytoplankton productivity) individuallyor synergistically could produce high sacoglossan densities

Marked asynchronyThe asynchrony between Lough Hyne populations of Elysiaviridis and outer coast ones in Ireland southern Englandand the British Isles could be due largely to the hydro-dynamics of Lough Hyne Recent work on the survival ratesof larval invertebrates passing out through the tidal rapids atLough Hyne (Jessopp 2007) documented the high mortalityof thin-shelled gastropod veligers and many other larval

Fig 7 Results of algal selections of juvenile Elysia viridis in September 2004(A) and September 2005 (B) Slugs were collected from Codium vermilara inLough Hyne and assigned to one of three treatments no-choice Codiumno-choice alternative hosts or pairwise choice of Codium and alternativehosts Replication for each of the three treatments is indicated Thestatistical tests compare categorical algal selection of slugs in choice andno-choice situations A significant difference indicates a differential responsebetween choice and no-choice situations



types Although Jessopp did not investigate sacoglossan veli-gers the shells are extremely delicate (Trowbridge personalobservation) Thus the tidal rapids may well be a partialbarrier separating slug populations within the lough fromthose on the outer coast Molecular work by Bell ampOkamura (2005) on Nucella lapillus (Linnaeus 1758) alsosupports the hypothesis that the rapids represent a dispersalbarrier however the whelk has direct development so thespecies dynamics would inherently differ from those of saco-glossans with planktotrophic larvae

Another second contributory factor is that the semi-enclosed lough becomes stratified with warm surface-watertemperatures in the summer For example during thisstudy the water temperature inside the lough was around168C in September 2001ndash2007 water outside was ~3ndash48Ccolder The higher water temperatures would lead to morerapid larval and juvenile growth in the lough relative tocoastal waters The temperature differential could produce aphenological asynchrony between sacoglossan populationsinside versus outside the lough

Another potential causal mechanism of the asynchronouspopulations could be the complex water movements withinthe region of this study Thermal fronts and local upwellinghave been identified in the Irish Sea Celtic Sea and westernend of the English Channel (eg Holligan 1981 Simpson1981 Raine et al 1990) These fronts and upwelling regionssupport different phytoplankton populations (eg diatomversus dinoflagellate dominance) than closely adjacent watermasses (Holligan 1981 Raine et al 1990) Differencesin the thermal regime and the phytoplankton communitycould affect veliger larval survival and growth potentially con-tributing to differential recruitment and post-metamorphicgrowth Detailed phenological comparisons have not beenmade to our knowledge for any invertebrate or macroalgalspecies between lough populations and other areas of theBritish Isles

Variable host use feeding and effects on thehostsFrom transplant experiments it appears that Elysia may havesevere local effects on its host plants particularly in the 2005ndash2006 experiment (Figure 5C) Thus the first objective of ourstudy was supported namely sacoglossans could potentiallybe a major consumer of Codium fragile in the loughBecause the densities of slugs in Lough Hyne are very highElysia herbivory could have contributed to the post-1970sdemise of the invasive Codium fragile in the lough analogousto that observed in Oban Scotland (Trowbridge 2002)Nevertheless the sacoglossans do not seem to have affectedpopulations of the native C vermilara This alga grew slightlydeeper and the similar densities of recruits on sympatric thalli(Trowbridge 2002 amp unpublished observations) indicate thatthe differential attack may be a habitat-specific effect ratherthan a species-specific one

Within Lough Hyne Elysia viridis feeds on a variety of algalhosts in at least four genera of algae The degree to which slugsform ecologically distinct sub-populations versus ecologicallyintegrated populations on these hosts is unclear The basisof feeding preferences and food choice in our study is influ-enced in part by parental diet at least by genus of host(Trowbridge amp Todd 2001) However some individuals can

learn to feed on alternative hosts after a lag time other con-specifics cannot switch hosts (Jensen 1989 1993) Toothshape per se may constrain switches between septate andsiphonous algal hosts (eg Chaetomorpha versus Codium inJensen 1989 1993) However tooth shape does not fullyaccount for the capacity (or lack thereof) to host-switch inElysia viridis most individuals in Lough Hyne have blade-shaped teeth with rounded tips (C tomentosum does notoccur inside the lough and Chaetomorpha is not commonthere except in the Rapids) Yet there was limited short-termcapacity to switch between Codium and Cladophora in thepresent study (Figure 7) as well as limited long-term capacityin a previous study (Trowbridge amp Todd 2001)

Patrick Krug (personal communication) has found nogenetic differentiation among conspecific specimens ofElysia viridis on different algal hosts implying that the slugdynamics are ecologically produced Increased abundance ofan algal hostmdashsuch as the invasive Codium fragilemdashcouldcreate a biological control-feedback limiting that alga theabundant algal host would be attacked disproportionatelyand the offspring would be predisposed to attack the host rela-tive to alternative hosts Even if there were no maternal effectsfor congeneric algal species host lsquoapparencyrsquo could lead to thedisproportionate attack of C fragile Clearly the role of larvalchoice of hosts on the shore merits further investigation

Role of predationThe sacoglossan recruits are concentrated at the upper tidallimit of Codium spp (Trowbridge 2004) One causal expla-nation could be that there is intense though undocumentedpredation on subtidal slugs The field transplant experimentwith Codium vermilara inside versus outside coarse meshbags (38 mm opening) demonstrated no significant treatmenteffect indicating that large predators were not responsible forthe abundance versus depth pattern previously documentedThe high densities of small fish (juveniles as well as smallspecies such as gobies) observed in the lough indicate thatpredation by small predators could be ecologically importanteven though large predators seemingly were not Microscopicpredators could also be important for example mites onCladophora rupestris readily consumed larvae and post-larvaeof Elysia viridis (Trowbridge unpublished observation)

An alternative causal mechanism that may account for thesacoglossan abundance versus depth distribution is the possi-bility that veliger larvae are concentrated into shallow watersGreenwood et al (2001) documented the shallow distributionof gastropod larvae in the lough although they did not identifythe veligers The results of our tidal-level experiment do notsupport this hypothesis though we did not include a 1-m sub-littoral treatment

A related not mutually exclusive hypothesis is that thelarvae andor post-larvae may move toward light (eg positivephototaxis) Experimental work by Grosberg (1982) indicatedthat pre-settlement behaviour (such as zonation within theplankton) can strongly influence the distribution of post-metamorphic individuals Elysia viridis does ingest retainedfunctional chloroplasts from its benthic hosts Althougha benefit in growth is not seen within the first three weeksof post-metamorphic life (Trowbridge 2000) sacoglossansmay still migrate to the algal hosts in the shallowest water toenhance the amount of light captured and used in subsequentfunctional kleptoplasty Although quantitative data are



meagre sacoglossans tend to exhibit high host fidelity aftersettlement juveniles are presumed to remain on a host indi-vidual unless dislodged by waves Demonstrating thishowever would be a logistical challenge given the small sizeof the juvenile slugs

In conclusion sacoglossan recruitment in Lough Hyne isimpressively high The causal mechanisms producing suchhigh densities on algal hosts have not been fully elucidatedbut may include (i) larval retention in the lough (ii) popu-lation asynchrony inside versus outside the lough (whichcould prolong the temporal peak of the slugs) and (iii) wave-sheltered conditions The role of slug genotype tooth mor-phology and abundance of alternative hosts all merit futureconsideration in understanding the complex slugndashalgal inter-actions on north-east Atlantic shores

A C K N O W L E D G E M E N T S

We thank A Miles G Pilling HA Ross T FarnhamA Laferriere S Schroeder and R Lord for their field assist-ance with this project We thank D OrsquoDonnell P Grahamand R McAllen for permission to work at Lough Hyne andmany valuable logistical conversations We thank P VanBragt for specimens of Dutch Elysia viridis from red algaeand P Krug for preliminary genetic analyses We appreciatethe time and care with which the referees evaluated ourMS their comments substantially improved our paperThis research was conducted with the approval of the IrishNational Parks and Wildlife (scientific permits R2007R2506 R3205 R1904 R1103 and R402) This materialis based upon work supported by the National ScienceFoundation under Grant No 0211186 Any opinions findingsand conclusions or recommendations expressed in thismaterial are those of the authors and do not necessarilyreflect the views of the National Science Foundation

R E F E R E N C E S

Ballard L and Myers A (2000) Observations on the seasonal occurrenceand abundance of gelatinous zooplankton in Lough Hyne Co Corksouth-west Ireland Biology and Environment 100B 75ndash83

Barnes DKA Verling E Crook A Davidson I and OrsquoMahoney M(2002) Local population disappearance follows (20 yr after) cycle col-lapse in a pivotal ecological speciesMarine Ecology Progress Series 226311ndash313

Bassindale R Ebling FJ Kitching JA and Purchon RD (1948) Theecology of the Lough Ine Rapids with special reference to water cur-rents Journal of Ecology 36 305ndash322

Bell JJ and Okamura B (2005) Genetic diversity in a marine naturereserve re-evaluating diversity criteria in reserve design Proceedingsof the Royal Society of London Series B Biological Sciences 2721067ndash1074

Colgan N (1911) Marine mollusca Clare Island Survey part 22Proceedings of the Royal Irish Academy 31B 1ndash136

Gaines SD and Bertness MD (1992) Dispersal of juveniles and variablerecruitment in sessile marine species Nature 360 579ndash580

Greenwood A OrsquoRiordan R and Barnes DKA (2001) Seasonality andvertical zonation of zooplankton in a semi-enclosed sea lough Journalof the Marine Biological Association of the United Kingdom 81213ndash220

Grosberg RK (1982) Intertidal zonation of barnacles the influence ofplanktonic zonation of larvae on vertical distribution of adultsEcology 63 894ndash899

Holligan PM (1981) Biological implications of fronts on the northwestEuropean continental shelf Philosophical Transactions of the RoyalSociety of London Series A 302 547ndash562

Jensen KR (1989) Learning as a factor in diet selection by Elysia viridis(Montagu) (Opisthobranchia) Journal of Molluscan Studies 5579ndash88

Jensen KR (1993) Morphological adaptations and plasticity of radularteeth of the Sacoglossa (frac14Ascoglossa) (Mollusca Opisthobranchia)in relation to their food plants Biological Journal of the LinneanSociety 48 135ndash155

Jessopp MJ (2007) The quick and the dead larval mortality due to tur-bulent tidal transport Journal of the Marine Biological Association ofthe United Kingdom 87 675ndash680

Jessop MJ and McAllen RJ (2007) Water retention and limited larvaldispersal implications for short and long distance dispersers inmarine reserves Marine Ecology Progress Series 333 27ndash36

JohnsonMP and Costello MJ (2002) Local and external components ofthe summertime plankton community in Lough Hyne Ireland a stra-tified marine inlet Journal of Plankton Research 24 1305ndash1315

Johnson MP Costello MJ and OrsquoDonnell D (1995) The nutrienteconomy of a marine inlet Lough Hyne South West IrelandOphelia 41 137ndash151

Kitching JA and Thain VM (1983) The ecological impact of the seaurchin Paracentrotus lividus (Lamarck) in Lough Ine IrelandPhilosophical Transactions of the Royal Society of London Series B300 513ndash552

Little C Morritt D and Stirling P (1992) Changes in the shore faunaand flora of Lough Hyne Irish Naturalistsrsquo Journal 24 87ndash95

Myers AA Little C Costello MJ and Partridge JC (1991) Theecology of Lough Hyne Proceedings of a conference 4ndash5 September1990 Dublin Royal Irish Academy

Padilla DK (1998) Inducible phenotypic plasticity of the radula inLacuna (Gastropoda Littorinidae) Veliger 41 201ndash204

Provan J Booth D Todd NP Beatty GE and Maggs CA (2007)Tracking biological invasions in space and time elucidating the trueinvasive history of the green alga Codium fragile using old DNADiversity and Distributions DOI 101111j1472-4642200700420

Raine R OrsquoMahony J McMahon T and Roden C (1990) Hydrographyand phytoplankton of waters off south-west Ireland Estuarine Coastaland Shelf Science 30 579ndash592

Rawlinson KA Davenport J and Barnes DKA (2005) Tidal exchangeof zooplankton between LoughHyne and the adjacent coast EstuarineCoastal and Shelf Science 62 205ndash215

Renouf LPW (1931) Preliminary work of a new biological station(Lough Ine Co Cork IFS) Journal of Ecology 19 410ndash438

Renouf LPW (1939) Faunistic notes from the south coast of CountyCork Eire Annals and Magazine of Natural History 4 520ndash525

Silva PC (1955) The dichotomous species of Codium in Britain Journalof the Marine Biological Association of the United Kingdom 34565ndash577

Simpson JH (1981) The shelf-sea fronts implications of their existenceand behaviour Philosophical Transactions of the Royal Society ofLondon Series A 302 531ndash546

Trowbridge CD (1992a) Mesoherbivory the ascoglossan sea slugPlacida dendritica may contribute to the restricted distribution of itsalgal host Marine Ecology Progress Series 83 207ndash220



Trowbridge CD (1992b) Phenology and demography of a marine special-ist herbivore Placida dendritica (Gastropoda Opisthobranchia) alongthe central coast of Oregon Marine Biology 114 443ndash452

Trowbridge CD (1993) Interactions between an ascoglossan sea slug andits green algal host branch loss and role of epiphytes Marine EcologyProgress Series 101 263ndash272

Trowbridge CD (1998) Ecology of the green macroalga Codium fragile(Suringar) Hariot 1889 invasive and non-invasive subspeciesOceanography and Marine Biology an Annual Review 36 1ndash64

Trowbridge CD (2000) The missing links larval and post-larvaldevelopment of the ascoglossan opisthobranch Elysia viridis Journalof the Marine Biological Association of the United Kingdom 801087ndash1094

Trowbridge CD (2001) Coexistence of introduced and native congenericalgae Codium fragile and C tomentosum on Irish rocky intertidalshores Journal of the Marine Biological Association of the UnitedKingdom 81 931ndash937

Trowbridge CD (2002) Local elimination of Codium fragile on Argyllshores indirect evidence of sacoglossan herbivory Journal of theMarine Biological Association of the United Kingdom 82 1029ndash1030

Trowbridge CD (2004) Emerging associations on marine rocky shoresspecialist herbivores on introduced macroalgae Journal of AnimalEcology 73 294ndash308

Trowbridge CD (2006) A global proliferation of non-native marine andbrackish macroalgae Invited chapter in Critchley AT Ohno M andLargo DB (eds) World seaweed resourcesmdashan authoritative reference

and system DVD-ROM version 10 Amsterdam ETI Bioinformatics(publisher)

Trowbridge CD and Todd CD (1999) The familiar is exotic II Codiumfragile ssp tomentosoides on Scottish rocky intertidal shores BotanicalJournal of Scotland 50 161ndash179

Trowbridge CD and Todd CD (2001) Host-plant changes of marinespecialist herbivores ascoglossan sea slugs on the introducedCodium fragile Ecological Monographs 71 219ndash243

Trowbridge CD and Farnham WF (2004) Spatial variation in littoralCodium assemblages on Jersey Channel Islands (southern EnglishChannel) Botanica Marina 47 501ndash503

Trowbridge CD Farnham WF and White L (2004) Thriving popu-lations of the native macroalga Codium tomentosum on Guernseyrocky shores Journal of the Marine Biological Association of theUnited Kingdom 84 873ndash877

and

Trowbridge CD Hirano YJ and Hirano YM (2008) Sacoglossanopisthobranchs associated with the green macroalgae Codium sppon Pacific rocky shores of Japan Venus 66 175ndash190

Correspondence should be addressed toCynthia D TrowbridgeDepartment of Zoology Oregon State UniversityPO Box 1995 Newport OR 97365 USAemail trowbriconidorstedu



In Lough Hyne a marine lough in south-west Ireland Elysiaviridis associates with several green algae including the nativeCodium vermilara (Olivi) Delle Chiaje and an introduced con-gener Codium fragile The populations of C vermilara havestayed relatively constant over many years but those ofC fragile have fluctuated wildly (Little et al 1992 Little per-sonal observation) It is likely that the introduced species firstbegan to spread within the lough in 1937 though at the timeit was referred to as C tomentosum Stackhouse (Renouf1939) (see Silva 1955 for clarifying the taxonomic confusion)(Codium tomentosum (sensu stricto) does not occur in thelough) The proliferation of C fragilewithin the lough occurredover the next 50 years but subsequent decline has reducedpopulations to a low level (Trowbridge et al unpublisheddata) Field observations in Scotland (Trowbridge 2002)suggested that herbivory by high densities of sacoglossanscould lead to the local decline of Codium populations

The aims in the work reported here were (1) to examinewhether Elysia viridis could potentially be a major consumerof the invasive alga in Lough Hyne (2) to quantify the patternsof grazing by Elysia on native and introduced Codium sppand (3) to evaluate the role of predators in limiting the popu-lations of Elysia and thus in influencing their effects on thealgal hosts

M A T E R I A L S A N D M E T H O D S

Study areaLough Hyne is a small (~1 km2) fully marine lough in CountyCork south-west Ireland Locations around the lough areidentified by numbers (see Figure 1) referred to as Renoufsectors (after Renouf 1931) The lough is connected to thenorth-east Atlantic Ocean via narrow tidal rapids (Figure 1)The narrow channel (~20 m wide) and the raised rocky sillrestrict the water flow exiting the lough during ebb tide Thelough has an asymmetrical semi-diurnal tidal cycle with~85 hours for tidal ebb and ~4 hours for tidal flood Themaximal tidal flow is ~3 msecond and the turbulent flowthrough the rapids kills many delicate invertebrate larvae(Bassindale et al 1948 Jessopp 2007) Because of therestricted seawater exchange the flushing time estimated forthe lough is 41 days (Johnson et al 1995) This period islonger than the larval period of Elysia viridis (Trowbridge2000 Trowbridge amp Todd 2001)

Sacoglossan populationsFrom 2001 to 2007 thalli of Codium spp were collected inSeptember at Lough Hyne and brought back to the laboratoryto examine Sacoglossans were manually removed (primarilyElysia viridis) with pipettes and forceps from several litres ofalgae each September The small size of the sacoglossans(mostly 5 mm and 1 mg) prevented us from weighingindividual animals therefore specimens were measured tothe nearest millimetre while crawling across a Petri dishwith clean seawater Based on a previous culturing study onE viridis specimens 5 mm would have settled and meta-morphosed in the previous 1ndash3 weeks (Trowbridge 2000)The algae were blotted and the wet weight measured to thenearest gram In 2003 Codium was dried in a microwave todetermine the slug abundance based on algal dry weight

The strong annual pulse of recruitment of Elysia viridis hadnever before been witnessed by the first author during earlierScottish and Irish work (1996ndash2002) Therefore during thecourse of surveying 95 shores of the British Isles for Codium(2001ndash2007) all E viridis encountered were measured andthe algal host species were recorded These sites included 41shores in County Cork 2 in County Clare 1 in NorthernIreland 13 in south-eastern Scotland 12 in southernEngland and 26 shores in the Channel Islands (Table 1)

The shores were surveyed from high-shore pools to low-shore emergent substratum for Codium spp All otherknown hosts of Elysia viridis were examined in the course ofthis survey all specimens of E viridis that were found werecollected except when local laws prohibited collection (egDunny Cove Ireland) Because of the slugsrsquo generally largesize specimens were gently blotted on paper towels andthen weighed to the nearest milligram

Radular teeth were dissected from a large number of speci-mens from each algal host and the tissues dissolved in 10Clorox bleach The cleaned radular ribbon was photographedwith a digital microscope and the blade-shaped teeth werecategorized as rounded-tips versus curved hook-like tipsMore detailed analyses of the radular teeth (eg length andwidth of the leading radular tooth and the number of ascend-ing and descending teeth) are being presented elsewhere Ourintention here was to evaluate whether tooth-shape per secould have precluded certain host changes within LoughHyne Tooth size and shape are known to be induced byalgal diet in this species (Jensen 1989 1993) Some

Fig 1 Map of the permanently established shore sectors as designated byRenouf (1931) at Lough Hyne County Cork Ireland Base map from Myerset al (1991) Individual grid units represent 100 m Symbols indicate theblocks of replicates of Codium transplant experiments






































R E S U L T S



























N






















D I S C U S S I O N


























R E F E R E N C E S











































and








































R E S U L T S



























N






















D I S C U S S I O N


























R E F E R E N C E S











































and
















N






















D I S C U S S I O N


























R E F E R E N C E S











































and





















D I S C U S S I O N


























R E F E R E N C E S











































and





















R E F E R E N C E S











































and


















and





journal of the marine biological association of the united ... · causeway; la jaonneuse; le...

Documents