meiofauna in the gollum channels and the whittard canyon ... · gollum channel system were sampled...

OPEN 3 ACCESS Freely availab le on line PLos one

Meiofauna in the Gollum Channels and the Whittard Canyon, Celtic Margin— How Local Environmental Conditions Shape Nematode Structure and FunctionJeroen Ingels1*, Alexei V. Tchesunov2, Ann Vanreusel11 M arine Biology D epartm en t, G hen t University, G hent, Belgium, 2 D ep artm en t o f Inverteb rate Zoology, Faculty o f Biology, M oscow S ta te University, Vorobyovy Gory,

M oscow, Russia

AbstractThe Gollum Channels and Whittard Canyon (NE Atlantic) are two areas that receive high Input of organic matter and phytodetrl tus from euphotlc layers, but they are typified by different trophic and hydrodynamic conditions. Sediment blogeochemlstry was analysed In conjunction with structure and diversity of the nem atode community and differences were tested between study areas, water depths (700 m vs 1000 m), stations, and sediment layers. The Gollum Channels and Whittard Canyon harboured high meiofauna abundances (1054-1426 Ind. 10 cm -2 ) and high nematode diversity (total of 181 genera). Next to enhanced meiofauna abundance and nematode biomass, there were signs of high levels of organic matter deposition leading to reduced sedimentary conditions, which In turn structured the nematode community. Striking In this respect was the presence of large numbers of 'chemosynthetlc ' Astom onem a nematodes (A stom onem a southwardorum, Order Monhysterida, Family Siphonolaimidae). This genus lacks a mouth, buccal cavity and pharynx and possesses a rudimentary gu t containing Internal, symbiotic prokaryotes which have been recognised as sulphur-oxidising bacteria. Dominance of Astom onem a may Indicate the presence of reduced environments In the study areas, which Is partially confirmed by the local blogeochemlcal environment. The nem atode communities were mostly affected by sediment layer differences and concomitant trophic conditions rather than other spatial gradients related to study area, water depth or station differences, pointing to small-scale heterogeneity as the main source of variation In nematode structure and function. Furthermore, the positive relation between nematode standing stocks, and quantity and quality of the organic matter was stronger when hydrodynamic disturbance was greater. Analogically, this study also suggests that structural diversity can be positively correlated with trophic conditions and that this relation Is tighter when hydrodynamic disturbance Is greater.

C ita tio n : Ingels J, T chesunov AV, V anreusel A (2011) M eiofauna in th e Gollum C hannels and th e W hittard Canyon, Celtic M argin— How Local Environm ental C onditions S hape N em atode S tructure and Function. PLoS ONE 6(5): e20094. doi:10.1371/journal.pone.0020094

E d ito r: Martin Solan, University o f A berdeen, U nited Kingdom

R eceived March 14, 2011; A c cep ted April 11, 2011; P u b lish ed May 18, 2011

C o p y rig h t: © 2011 Ingels e t al. This is an open-access article d istribu ted u n d er th e te rm s o f th e Creative C om m ons A ttribution License, w hich perm its unrestricted use, distribution , and reproduction in any m edium , provided th e original au th o r and source are credited .

F u n d in g : This research w as funded by th e European Com m ission's Sixth Fram ew ork P rogram m e HERMES pro jec t (h ttp ://w w w .eu-herm es.ne t, EC con trac t GOCE511234) and th e European C om m ission's S eventh Fram ew ork P rogram m e HERMIONE pro ject (h ttp ://w w w .eu-herm ione.net, g ran t ag ree m en t n um ber 2263541). Participation o f AVT in th is research w as su p p o rted by th e Russian Fund for Basic Research (g ran t n u m b er 09-04-01212). The funders had no role in s tudy design , d a ta collection and analysis, decision to publish, o r p repara tion o f th e m anuscrip t.

C o m p e tin g In te rests : The au tho rs have declared th a t no com peting in terests exist.

* E-mail: je roen .ingels@ ugent.be

IntroductionT h e Celtic M argin situated in the N E Atlantic is a highly

productive system w ith significant p rim ary production in the surface waters, w hich consequently supplies deep-sea sediments with high levels o f organic m atter (OM) and carbon [1,2] com pared to o ther deep-sea areas. T his is especially the case for mid-slope depths, w hich are influenced by additional export from the shelf o r upper slope, w here sediment organic loads are higher[3]. At the Porcupine Seabight an d further south along the M eriadzek T errace , the m argin is incised by num erous canyons and channels. T hese geohydrological structures provide conduits for the transport o f sedim ent from the shelf to the abyssal p lain and over-bank turbidity currents, w hich deposit on the intervening terraces and spurs [4], bu t they also accum ulate high am ounts o f sediments and O M , m aking them im portan t in budgeting carbon sinks and sources [5,6], In addition, at the Celtic M argin cascading o f dense w ater masses down the slope is likely to occur [7] and m ay

entra in fresh chlorophyll m aterial rapidly dow n slope, as reported by Hill et al. [8], T hese observations all imply the presence of enhanced levels o f O M and phytodetrital m aterial in canyons and channel systems in this area, w hich will have implications for the trophic conditions, the biogeochem ical processes an d ultim ately the benthic fauna. In particu lar m eiobenthic densities and standing stocks in canyons are under control o f the flux o f O M from surface productivity to the seafloor an d resulting trophic conditions in the sedim ent [9,10,11], In addition to enhanced (although variable) O M levels, canyons and channel systems generally display high hydrodynam ic activity in the form o f tidal currents, episodic slumps, turbidity flows, an d density induced internal waves and cascading events, w hich are likely to have a significant effect on m eiobenthos by physically d isturbing the sediments they reside in [10,11,12,13,14,15,16,17,18],

T h e effect o f natural disturbances on com m unities have continuously instigated ecologists to develop and test conceptual models that explain patterns a t various tem poral and spatial scales

PLoS ONE I w ww.plosone.org 1 May 2011 | Volume 6 | Issue 5 | e20094

http://www.eu-hermes.net

http://www.eu-hermione.net

mailto:[email protected]

http://www.plosone.org

Structure and Function of Marine Canyon N em atodes

based on biological interactions a n d /o r abiotic processes; e.g. In term ediate D isturbance Hypothesis [19], D ynam ic Equilibrium M odel [20]. In applying these models to the benthic environm ent, changes in sedim entary trophic param eters and physical disturbances m ay regulate deterministic biotic processes, leading to the large and small-scale patterns in benthic fauna. Specifically applied to the deep sea, Levin e t al. [21] presented a schematic m odel based on five ecological forces (productivity, flow, bottom -w ater oxygen, sedim ent heterogeneity, biotic disturbance) to predict regional-scale variation of local diversity (1-10 m 2). For the smaller benthic fauna, the m eiofauna (32-1000 pm), one could even look a t smaller spatial scales and include the vertical sedim ent profile in explaining com m unity an d diversity patterns. Such patterns along the vertical sedim ent profile m ay be explained by trophic conditions and oxygen concentration in the different sedim ent layers; e.g. T R O X m odel for Foram inifera [22]. Following this concept, Soetaert e t al. [23] suggested that similar processes m ay control benthic nem atode com m unity patterns. Based on various interpretations o f environm ental variation in the different schematic m odels it is likely that there is a hierarchical interaction betw een environm ental drivers w hich regulates abundance and diversity o f the m eiofauna on various spatial and tem poral scales.

C anyons and channel systems provide suitable habitats to test the effect o f natu ral disturbances on benthic faunas. H igh organic loads and variable sedim entary trophic conditions, variable frequency an d intensity o f d isturbance events, and sediment heterogeneity are all features ascribed to canyon and channel systems and m ay be used to explain benthic faunal patterns. Furtherm ore, it has becom e increasingly clear th a t variability betw een canyons and even w ithin one canyon system is very high and th a t productivity an d d isturbance processes can shape m eiobenthic com m unities differently [10,11,24], D ue to this high variability, studies investigating how processes th a t affect the sedim entary environm ent influence the benthic fauna in canyons and channel systems should be conducted on a scale th a t no t only allows for sedim entary contrasts on a cm-scale, b u t also includes com parisons betw een different canyons or channels and w ater depths. T h e extent to w hich sedim entary conditions an d processes control standing stocks, and structural and functional diversity o f m eiobenthic com m unities in canyons and channels w ould provide im portan t inform ation on how benthic com m unities are shaped and w hat drives diversity in these environm ents.

T h e Gollum C hannel system and W hittard C anyon along the Irish M argin have only recently been subject to m ore intensive investigations, prim arily w ithin the fram ew ork o f the E U FP6 H E R M E S project, an d therefore little previous inform ation is available [25,26], especially on the biology o f the benthos [27,28]. In this first study on nem atode com m unities - as representative for the m etazoan m eiofauna - from these two areas we tested nem atode standing stocks, diversity, an d com m unity structure for differences betw een study sites, w ater depths, an d sedim ent depths. Differences a re in terpreted in relation to biogeochem ical p a ram eters, trophic characteristics and sedim ent disturbance processes. Given the high hab ita t heterogeneity and environm ental variability in these canyons an d channels, we hypothesise th a t the nem atode assemblages will be largely controlled by small-scale, local environm ental conditions ra th e r th an large-scale differences betw een canyons or w ater depths.

Materials and Methods

S am plin g a n d s tu d y a reaD uring the H E R M E S R V Belgica 2 0 0 6 /1 3 cruise, 23-29 Ju n e

2006 [29] to the Irish M argin, m eiofauna samples were taken

using a M idicorer (M D C, O cean Scientific In ternational Ltd), yielding samples with a virtually undisturbed sedim ent-water interface. T h e M D C was equipped with four core tubes with an inner d iam eter o f 100 m m , b u t 57 m m diam eter subcores were taken for m eiofauna and nem atode analyses. T w o sites in the Gollum C hannel system were sam pled a t ca. 700 and 1000 m dep th (G700 an d G 1000, respectively) as well as two stations in betw een two north-eastern branches o f the W hittard C anyon at similar depths (W700 an d W 1000, respectively). A n overview and details o f the sam pling areas are provided in Fig. 1.

T h e Gollum C hannel system has been described as a tributary channel system incising the upper slope of the south-eastern Porcupine Seabight [30,31]. I t converges into one m ain channel, the Gollum C hannel, w hich opens into the Porcupine Abyssal Plain. Samples were taken in the m ost northerly channel o f the Baggins b ran ch of the channel system, the Bilbo C hannel [32]. T h e Gollum C hannel system lies along the N E Atlantic M argin w here the w ater mass characteristics in the upper 1000 m are dom inated by E astern N orth Atlantic W ater and M editerranean Outflow W ater. This results in relatively w arm (8—10°C) and saline w ater (ca. 35.5) betw een 700 and 1000 m depth, with weak vertical tem peratu re and density gradients [33]. H ydrographic activity in the Bilbo C hannel was investigated by W hite [33] and indicated m ean currents o f 4.7 cm s 1 with a m axim um recorded speed o f 53 cm s *. T h e flow in the canyon channel is dom inated by the sem i-diurnal tide and is fairly dynam ic, resulting in currents strong enough to produce significant turbidity. Like m any other canyon and channel systems, there appears to be a significant down-slope flow of suspended m aterial w ithin the Bilbo C hannel[34].

T h e W hittard C anyon consists o f a num ber o f deeply incised branches th a t extend from the shelf b reak south o f the G oban Spur. Further dow n canyon, the walls are lower b u t the b ranch ing na ture o f this canyon continues towards the canyon fan a t abyssal depths [26]. T h e sam pling stations in the W hittard C anyon were situated on an interfluvial a rea covered in fine sediments in betw een two upper N E branches (Fig. 1). In the upper reaches o f the canyon down-slope sedim ent transport is dom inated by turbidity currents. T hese m ud-rich flows are o f sufficient size to overspill the canyon walls, leading to deposition o f m ainly fine sediments, e.g. m uds an d clays, on the adjacent continental slope[4]. A long the Celtic Shelf edge a m ean northw estern along-slope flow of ca. 6 cm s 1 has been reported [35].

B iog eoch em is t ryE ach 1-cm-thick slice from three replicate cores a t each

sam pling site (two for G700) were analysed for granulom etric and geochem ical param eters dow n to the 5-cm level. G rain size distribution was m easured with a M alvern M astersizer 2000 (0.02— 2000 pm size range). For total sedim entary organic carbon (TOC) and n itrogen (TN), samples were lyophilised, hom ogenised and acidified w ith 1 % HC1 until com plete decarbonisation and dried before being analysed w ith a PD Z E uropa A N C A -G SL elem ental analyser (UC Davis, USA). Chloroplastic P igm ent Equivalents (CPE: sum of chlorophyll a (chl-a) and its degradation products (phaeopigments)) [36] served to estim ate the am o u n t o f sedim entary O M derived from prim ary production. Sedim ent samples were lyophilised and hom ogenised, after w hich pigm ents were extracted in 90% acetone, separated using reverse-phase H P L C , and m easured w ith a Gilson fluorescence detector according to W righ t & Jeffrey [37]. T h e ratio chl-a:phaeopigm ents was used as a m easure o f ‘freshness’ o f the photosynthetically derived O M in the sediments, the ratio ch l-a :T O C gives an indication o f the bioavailability o f the bulk O M .



GollumChannels

PorcupineS e a b ig h t

GobanSpur Whittard Canyon —W e ste rn "

Approaches


i r

14‘W

52'N

48 ’N

48“ 5 0 'N

Figure 1. Overview (left) and detailed maps (right) of the study area with the Gollum stations (top right) and W hittard stations (bottom right). G1000: Gollum Channel at 1090 m depth, G700: Gollum Channel at 755 m depth, W1000: Whittard Canyon at 1160 m depth, W700: Whittard Canyon at 762 m depth. doi:10.1371/journal.pone.0020094.g001

M eiofauna a n d n e m a t o d e analysisA t each station, three replicate samples were sliced p e r cm down

to 5 cm sedim ent dep th an d fixed in buffered 4% form alin. Samples were w ashed over a 1000-pm m esh and then sieved on a 32-pm m esh to re ta in the m eiofauna fraction. T h e m eiofauna was then elutriated by the centrifugation-flotation technique using L U D O X [38,39] and stained with Rose Bengal. All m etazoan m eiobenthic organism s were classified a t h igher taxon level following Higgins & Thiel [40] an d counted under a stereo- m icroscope (50 xm agnification). For nem atode analysis, 100-150 individuals w ere picked out random ly from each sedim ent slice and transferred using the form alin-ethanol-glycerine technique [38] and m ounted on glass slides. N em atode identifications followed the pictorial key to nem atode genera o f W arwick et al. [41], taxonom ic literature from the nem atode library of G hent University, and the NeM ys database [42]. All individuals were grouped into four feeding types sensu W ieser [43] : selective deposit feeders (1A), non-selective deposit feeders (IB), epistratum feeders (2A), and predators/scavengers (2B). O ne additional feeding type was defined for gutless nem atodes o f the genus Astomonema. Because these nem atodes have a degenerated alim entary canal with absence o f m outh and buccal cavity, they do not belong to any o f W ieser’s four feeding types [43]. Length (excluding filiform

tail tips) and m axim al w idth were m easured using a Leica D M R com pound m icroscope an d Leica LAS 3.3 im aging software. N em atode wet weight was calculated according to A ndrassy’s [44] form ula and translated into nem atode dry weight w ith a dry-to- wet ratio o f 0.25 [39]. T o calculate total nem atode biomass associated w ith sedim ent layers, sam pling stations and feeding types, average biomass was calculated for each nem atode genus and m ultiplied by their density per sedim ent layer, for each layer, station and feeding type.

D ata analysisT h e m ultivariate com m unity da ta on genus level (standardised,

square-root transform ed, Bray-Curtis similarity was used to calculate resem blance) was analysed by m eans o f non-param etric perm utational A N O V A (PE R M A N O V A [45,46]) to assess differences betw een the Gollum and W hittard areas, betw een w ater depths, an d betw een sedim ent layers. T he da ta set was analysed using a 4-factor m ixed m odel design (factors: C anyon (Ca; fixed), W ater D epth (WD; fixed), Sedim ent D epth (SD; fixed), Cores (Co; to account for replicate variability; random and nested in C ax W D ) in PE R M A N O V A + for P R IM E R [46]). Since the vertical layers in the sedim ent were no t replicated within each core, a split-plot design was used leading to a repeated m easures




analysis w hereby the m ain-factor test was followed by pair-wise com parisons w ithin each C a x W D com bination to investigate significant in teraction effects in the full-model test. A non-m etrical M ulti-dim ensional scaling plo t (MDS) was used to visualise the PE R M A N O V A results and illustrate the three-w ay interaction term s. T h e M D S was obtained by calculating the centroids o f the three-w ay interaction cell groupings in the full m ultivariate Bray- Curtis space, followed by calculating the distances am ong them (PER M A N O V A + for P R IM E R [46]). A two-way crossed (Ca,W D) univariate PE R M A N O V A test was perform ed on average biom ass da ta (data was fourth-root transform ed, resem blance calculated w ith Bray-Curtis). Averaging the biom ass data over replicates gave a m ore realistic estim ate o f the true biomass pe r sedim ent layer, station an d canyon.

A two-way crossed (factors: C anyon and W ater Depth) S IM PE R (Similarity Percentages - species contributions) analysis was perform ed to reveal w hich genera are responsible for the m ultivariate com m unity patterns w ithin an d betw een canyons and w ater depths.

S tructural diversity o f the com m unity was calculated using Hill's [47] diversity num bers (H 0, H ¡, H 2, H lnj), estim ated num ber o f genera (EG(51), [48,49]). Functional o r trophic diversity (TD) was calculated as the reciprocal value o f the trophic index ( 0 _1) by H eip e t al. [50], so th a t h igher values correspond with h igher T D , based on relative abundance data.

T h e full set o f 18 available environm ental variables was tested for collinearity (Draftsm an plot an d Spearm an correlation matrix) and (redundant) variables with correlations (r2) > 0 .9 were om itted from the m odel; chl-a, chl-a:phaeo, ch l-a :T O C ,a nd C PE :T O C needed log(X+0.1) transform ation to com pensate for skewness. For analysis o f selected geochem ical variables (chl-a, CPE, chl- a:phaeo, ch l-a :T O C , T O C , T N , C:N, m ean grain size), separate univariate PE R M A N O V A tests were perform ed (data was norm alised, Euclidean distance was used to calculate resem blance). T h e PE R M A N O V A design was identical to the one used for com m unity analysis. R E L A T E and D IS T M L (distance-based linear model) routines were perform ed [51,52] to analyse and m odel the relationship betw een the nem atode genera assemblages and the environm ental variables (chl-a, CPE, chl-a:phaeo, chl- a :T O C , C P E :T O C , T O C , T N , C:N, m ean grain size). T he assem blage D IS T M L was constructed using the step-wise selection procedure and the adjusted R 2 as selection criterion to enable the fitting o f the best explanatory environm ental variables in the m odel [51]. In addition, D IS T L M models w ere calculated for selected biotic param aters (Nem atode density, relative abundance ‘chem osynthetic’ Astomonema nem atodes, total nem atode biomass, biomass ‘chem osynthetic’ Astomonema, trophic diversity (TD), Hill's diversity num bers H 0, H ¡, and expected num ber o f genera (EG(51)). Euclidean distance was used as resem blance m easure in all D IS T L M procedures. C om plem entary to these analyses, non- param etric K endall-T au correlations were com puted betw een the same selected biotic param eters and abiotic variables for each of the can y o n /ch an n e l systems.

Results

S e d im e n t charac ter is t icsBiogeochemical properties o f the sedim ent are shown in Figs. 2

and 3. Grain-size fractions differed significandy betw een study areas and individual stations (p< 0 .01 , Table SI). H ighest m uddy clay and silt content were observed a t W 700 and highest sand content at G700 (Fig. 3). A t station W 1000, the sedim ent becam e finer with increasing sedim ent depth, whilst a t the o ther stations average grain size values declined slightly with sedim ent depth (Fig. 3).

Sedim entary T O C (p<0.05) and T N (pCO.Ol) contents differed significantly betw een stations (C ax W D , T able SI). T N also differed significantly betw een canyons and w ater depths (p< 0 .05 , T ab le SI), w hich was supported by significant pair-wise com parisons w ithin each canyon or w ater dep th group (data no t shown). C :N only differed significandy betw een stations (p< 0 .05 , Table SI). Station W 700 was characterised by highest average total-core T O C an d T N content, followed by G1000, G 700, and W 1000, respectively (Fig. 2). A t stations G 1000 and W 700, average total- core T O C , T N and C :N values were h igher com pared to G700 and W 1000 (Fig. 2), indicating h igher am ounts o f O M with greater fractions o f terrestrial a n d /o r very degraded O M a t these stations. No clear gradients were observed along the vertical sedim ent profile for T O C , T N , an d C :N (Fig. 3).

C hl-a differed significandy betw een canyon areas an d w ater depths (p< 0 .05 , T able SI), b u t the test showed a significant interaction term (C axW D ), w hich indicates station differences (pCO.Ol, T ab le SI). Subsequent pair-wise com parisons within each canyon an d w ater dep th level showed th a t these were due to significant differences betw een W 700 an d the o ther stations (data no t shown, b u t see Fig. 3). C PE levels contrasted significandy betw een w ater depths (p< 0 .05 , T able SI), b u t these were also caused by contrasting values betw een W 700 and the o ther stations. C PE an d chl-a concentrations w ere clearly higher a t W 700 th an at o ther stations (Fig. 2, 3) and decreased significandy w ith sediment dep th a t all stations (p< 0 .05 , T able S I; Fig. 3). T h e ‘freshness’ (chl-a:phaeo) and bio-availability (chl-a: T O C ) of O M decreased significandy w ith sedim ent dep th a t each station (p< 0 .01 , Table S I; Fig. 3), b u t did no t contrast betw een canyon areas, w ater depths o r stations (Table SI). Sedim ents a t W 700 showed higher chl-a:phaeo an d ch l-a :T O C com pared to the o ther stations, indicating the presence o f m ore ‘fresher’ an d bio-available O M .

M e ta z o a n m e io fau n aA total o f 21 h igher m eiofauna groups were identified o f which

20 occurred in the Gollum area and 15 in the W hittard a rea (Table S2). T he m ost diverse station was G 1000 with 18 different groups. N em atodes were consistendy the dom inan t taxon (90 .2- 95.2%), followed by nauplius larvae (1.51-4.30% ) an d copepods (1.06-2.94% ). O th e r taxa such as polychaetes an d tardigrades occurred typically in lower abundances (< 1 %). H ighest m eiofauna abundance was recorded a t G 1000 (Table S2, Fig. 2), bu t was still similar to abundances a t o ther stations because o f high variability betw een replicates. T o tal m eiofauna abundances consistendy decreased with increasing sedim ent depth.

N e m a to d e sNematode community and biomass. T h e PE R M A N O V A

com m unity results indicated significant differences betw een the Gollum and W hittard areas, w ater depths, sedim ent layers, and stations (p< 0 .01 , T ab le 1). T h e significant three-w ay interaction term (C ax W D x S D ) called for pair-wise com parisons w ithin each two-factor com bination to investigate the na ture o f the effect (Table S3) and the M D S plot in Fig. 4 clearly shows the reasons for the three-w ay in teraction suggested by the pair-wise com parisons. T h ere is the p a tte rn o f a g radient across the different sediment layers from the upper left (0-1) to the lower right (4-5), b u t the size o f the sedim ent-layer effect is very different in the different canyons an d at different w ater depths (Fig. 4). M ore specifically, W 1000 has a very m odest sedim ent-layer effect, while W 700 has a m uch larger effect, spreading practically all the way across the first M D S axis (Fig. 4). Overall, the surface layers (0-1 cm) are m uch m ore similar than the deeper layers, since all 0 -1 cm points are m ore densely grouped. T he difference betw een w ater depths is




2000

?*

CU

c 1000 ro■o

\

- 0.16 - 0.035

- 300 - 0.14 - 0.030

- 0.12- 0.025

- 0.10

- 200

- 0.08

- 0.06

- 0.020

- 0.015

- 100- 0.04

- 0.010

- 0.02 - 0.005

- 0 - 0.00 - 0.000

G700 G1000 W700

I I Meiofauna abundance (¡nd.lOcnrr2)I I N ematode abundance ( in d . lO c m 2)A Biomass (jig dwt)

W1000

O TOC (%)

□ Chl-a:phaeo © Chl-a :TOC

5T300

3LOLOroEo

CD

OCUra

. cQ.rä■

. cU

UO

- 1.2

- 1.0

- 0.8

- 0.6

- 0.4

- 0.2

0.0

\SoxUo

Figure 2. Average meiofauna and nem atode abundance, nem atode biomass, and selected sedim entary biogeochemical variables for each station (average total core values 0 -5 cm, 10 cm-2 ). TOC: total sedim entary organic carbon, Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating 'freshness' o f the phytodetrital OM, chl-a:TOC: chlorophyll a divided by total organic carbon con ten t indicating bioavailability of the bulk OM, dwt: dry w eight. G1000: Gollum Channel a t 1090 m d ep th , G700: Gollum Channel a t 755 m dep th , W1000: W hittard Canyon a t 1160 m d ep th , W700: W hittard Canyon at 762 m d ep th . Values averaged over replicates; error bars d en o te standard deviations. doi:10.1371/journal.pone.0020094.g002

m uch greater for the W hittard samples th an it is for the Gollum samples (Table S3, pair-wise com parisons w ithin C a x S D ; Fig. 4) and a t 1000 m w ater dep th the differences betw een Gollum and W hittard are m ore p rom inent th an a t 700 m w ater dep th (Table S3, pair-wise com parisons w ithin W D xSD , Fig. 4). In general, the M D S plot illustrates clearly the extent to w hich differences betw een sedim ent layers exceed differences betw een canyons and w ater depths (Fig. 4), w hich is supported by the estim ated com ponents o f variation (Table 1), and points to small-scale heterogeneity as the m ain source of variation in nem atode structure. M oreover, the im portan t role o f sedim ent dep th in shaping com m unity com position was clearly illustrated by the clustering o f 0—1 cm layers and the p ronounced dissimilarity betw een the 4 -5 cm layers from different stations (Fig. 4).

T h e dom inant genera for each station are given in T able 2. Sabatieria was clearly the dom inant genus in the Gollum area, its relative proportions being highest a t G 1000 (27 an d 16% for G 1000 and G 700, respectively). Sabatieria occurred in the W hittard C anyon, bu t in m uch lower densities (3.5—4.5%). In contrast, station W 1000 was dom inated by Leptolaimus and Molgolaimus (14.4 an d 13.5%, respectively), whilst at station W 700 mouthless nem atodes o f the genus Astomonema (-4. southwardorum, containing endo-symbiotic bacteria) constituted m ore th an 10% of the total com m unity.

T h e S IM PE R test revealed similarity values for the Gollum and W hittard a rea o f 51.5 an d 53.8% , respectively (Table 3). As

expected from its dom inant occurrence at the Gollum stations, Sabatieria clearly contributed m ost to similarity w ithin this a rea (14.5%). For the W hittard area, a b ro ad er range o f genera were responsible for similarity betw een stations. Sim ilarity betw een the 700 m stations was lower th an betw een the 1000 m stations (49.9 and 55.5% , respectively), and the genus Sabatieria contributed most to similarity w ithin these groups. Dissimilarity betw een the canyon and w ater-depth groups (57.5-59.4% ) was h igher th an similarity w ithin groups (49.9-55.5% ), with Sabatieria and Astomonema as the m ain discrim inating genera.

T otal nem atode biomass was significantly greater in the Gollum C hannel than in the W hittard area (Fig. 2, Table 4) w ith highest values recorded at station G 1000 and lowest a t W 1000 (Fig. 2). No significant differences were observed betw een w ater depths. Biomass typically decreased tow ards the deeper sedim ent layers.

Structural and functional nematode diversity. O ne hundred and eighty one genera were found am ong a total o f 6051 identified nem atodes (H 0, T able 5). T h e Gollum area was m ost diverse with 152 genera, com pared to 132 genera a t the W hittard stations, w ith 103 genera occurring in bo th areas. W ithin the Gollum C hannel, only 83 genera were shared betw een stations, whilst a t the W hittard stations, 64 genera were shared. T h e shallower stations (G700 and W 700) harboured m ore genera (127 an d 112, respectively) th an the deeper stations G 1000 and W 1000 (108 an d 84 genera, respectively). Generally, H ill’s indices




Chl-a (n g /g )

o-i

1-2

2-3

3-4

4-5

-0.02 0.04 0.06 0.08 0.10 0.12

TOC (%)

o-i

1-2

2-3

3-4

4-5

0.0 0.5 1.0 1.5 2.5

C lay (%)

o-i

1-2

2-3 ■iW-

3-4

4-5

0 5 15 20 25

Q.<yTJ+-»c0JE

T3<UKft

CPE (Me/e)

o-i

1-2

2-3

3-4

4-5

-0,1 0.0 0.1 0,2 0.3 0.4 0.5

0-1

1-2

2-3

3-4

4-5

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

S ilt (%)

0-1

1-2

2-3

3-4

4-5

0 10 20 30 40 50 60 70 80

Chl-a :p h a e o

0-1

1-2

2-3

3-4

4-5

-0,1 0.0 0.1 0,2 0.3 0.4

0-1

1-2

2-3

3-4

4-5

0 5 10 15 20 25 30 35

S a n d (%)

o-i

1-2

2-3

3-4

4-5

0 20 40 60 80 100

Chl-a :TOC

o-i

1-2

2-3

3-4

4-5

-0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12

■ G700■ G1000■ W 700■ W 1 0 0 0

Figure 3. Average values for selected sedim entary biogeochemical parameters along the vertical sediment profile. Chl-a: chlorophyll a; CPE: chloroplastic p igm ent equivalents; Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating 'freshness' of the phytodetrital OM; Chl-a:TOC: chlorophyll a divided by total organic carbon con ten t indicating bioavailability of th e bulk OM; TOC: total organic carbon concentration; TN: total nitrogen concentration; C:N: m olar carbon/n itrogen ratio; Clay, Silt, Sand: volum e percen t clay, silt, sand con ten t, respectively. G1000: Gollum Channel a t 1090 m d ep th , G700: Gollum Channel a t 755 m dep th , W1000: W hittard Canyon a t 1160 m d ep th , W700: W hittard Canyon at 762 m d ep th . Values averaged over replicates; error bars d en o te standard deviations. doi:10.1371/journal.pone.0020094.g003

indicated that W 700 was m ost diverse and equitable, followed by G700, whilst the deeper stations harb o u red less diverse and less even com m unities. At all stations, structural diversity did not show a consistent decrease o r increase w ith sedim ent depth.

T rophic diversity (TD) is shown in T able 5. T rophic diversity was h igher at the 700 m stations than at the 1000 m stations, and highest at W 700 (4.37). This was due to relatively high num bers o f the gutless nem atodes belonging to the genus Astomonema at this station. Lowest T D was observed a t station W 1000 (3.10). At the G ollum stations, the non-selective deposit feeders (IB) dom inated in term s of relative abundance, whilst at the W hittard stations the selective deposit feeders (1A) dom inated the assem blage. At all stations biom ass was dom inated by the non-selective deposit feeders (IB). Striking was the increasing num bers o f Astomonema

nem atodes w ith increasing sedim ent dep th a t station W 700. This was also the case, bu t to a m uch lesser extent, a t G 1000, where non-selective deposit feeders (IB) w ere dom inant.

Relation between structure and function of nematodes and environmental variables. T h e R E L A T E analysis indicated that the spatial p a tte rn based on the environm ental variables is significantly related to the patterns inherent to the com m unity structure (p = 0.27, p < 0 .01). O f the original set o f 18 environm ental param eters, 9 were re ta ined for further analysis based on collinearity analysis (Draftsm an plot, Fig. SI). N ine variables with correlation r~ values > 0 .9 (considered redundant) were om itted for the D IS T L M procedures; the rem aining variables and their pair-wise spearm an correlations are shown in T able S4. T h e best fitted m odel, based on the retained




0-1♦ " 0-1

......

1-2 í2

2-3

2D Stress: 0.14

0-1■ .......

0-1A"

1-2

1-2 \■2-3y

4-5 i

♦ .\ 2-3 ▲

' \ 3-4 ▲

3-4

\ 4-5 '♦

2-3 '', 3-4

y 3-4

A4-5

CaxWD A G1000 ▼ G700

4-5 ■ W1000 ♦ W700

Figure 4. Non-metric MDS plot (Bray Curtis similarity) based on standardised and square-root transformed nem atode genera relative abundance data. W700: W hittard 762 m w ater dep th , W1000: W hittard 1 1 6 0 m w ater d ep th , G700: Gollum 755 m w ater d ep th , G1000: Gollum 1090 m w ater d ep th . Labels d en o te sedim ent layer (cm). D otted lines connect the succeeding sed im ent layers per station.doi:10.1371/journal.pone.0020094.g004

environm ental variables, that explains the genera assemblage is shown in T able S5 and illustrated in Fig. 5. T hese results show that availability o f O M (chl-a:TOC) explains nearly 17% of the variation observed. O th er m ain contributors in explaining nem atode genera assemblage variation in genera assemblages betw een samples include T N (14.2%), m ean grain size (11.3%) and chl-a content (6.4%) o f the sediments. O th er variables contributed < 5 % in explaining assem blage variation. O f the 9 variables, 8 w ere used by the D IS T L M procedure to construct the best-fitting m odel, together explaining 64.3% of total variation. T h e D IS T L M results for the univariate biotic param eters are sum m arised in Table S6. T h ey reveal th a t the relative abundance and biom ass o f Astomonema are m ainly explained by T O C and T N in the sediments, while T D variability is m ainly explained by T N . N em atode density an d diversity indices are controlled by chl-a- derived param eters, indicating the im portan t role o f the quality and quantity o f phytodetrital O M for these biotic param eters.

Fig. 5 shows the D IS T L M results by m eans o f a dbR D A plot, with the relative abundance o f Astomonema superim posed. T h e vectors o f the environm ental variables retained by the D IS T L M procedure as fitting the best explanatory m odel indicate the im portan t role o f T O C , T N and C :N in explaining their abundance.

K endall-T au correlations betw een structural com m unity p a ram eters and environm ental variables are shown in T able 6. T h ere were differences betw een the Gollum an d W hittard stations, w ith 34 and 21 significant correlations, respectively, indicating different levels o f interactions betw een the two areas.

In the Gollum area, nem atode density, nem atode biomass, H (h and H j w ere positively correlated w ith the quality and bioavailability o f O M (Chl-a, CPE, ‘freshness’ (Chl-a:phaeo), bioavailability (C hl-a:TO C ). N em atode density was positively correlated with T N and silt content and showed a negative relation with sand content. T h e relative abundance and biomass o f Astomonema showed no correlation w ith chl-a and its derivatives. H ow ever, the biomass o f this genus was positively correlated w ith T O C , T N , and C :N while its relative abundance was positively correlated with T O C and C :N . A a tz i Astomonema biom ass was greater in finer sediments th an in coarser sediments as inferred from the correlations w ith grain-size fractions. T rophic diversity was positively correlated w ith ch l-a :T O C . G enus diversity (H 0, H¡) was positively correlated with chl-a and chl-a-derived variables.

In the W hittard area, nem atode density was positively correlated w ith ‘freshness’ and availability o f O M (chl-a:phaeo, chl-a:T O C ), whilst a negative relation was found w ith clay content. Astomonema relative abundance an d biomass was positively correlated w ith T O C an d C :N an d negatively correlated with sand content. N em atode biomass was positively correlated with chl-a and all chl-a-derived variables, indicating its dependence on photosynthetically derived m aterial. H (h H j , an d EG(51) were positively related to ‘freshness’ values (Chl-a:phaeo).

Discussion

Q u a n t i ty a n d qua li ty o f o rg an ic m a t te r a n d its e ffec t on m e io fa u n a a b u n d a n c e a n d b io m ass

A t the Celtic M argin area, the high productivity in surface waters leads to significant sedim entation o f O M to the deep sea [53,54], w hich m ay positively control m eiofauna standing stocks [55,56,57,58,59,60], M eiofauna abundances and biomass were

T a b le 1. Results from multivariate PERMANOVA analyses for differences in nematode com m unity structure.

S ource d f SS MS Pseudo-F P (p erm ) U n iq u e perm sEstim . o f c o m p , o f v a ria tio n

Ca 1 10490 10490 48.268 0.0021 9826 285.89

WD 1 8657.3 8657.3 39.836 0.0024 9839 222.89

SD 4 23748 5936.9 5.474 0.0001 9879 415.91

CaxWD 1 12637 12637 58.148 0.0024 9852 719.38

CaxSD 4 7691.4 1922.8 17.729 0.0001 9832 143.7

WDxSD 4 6383.9 1596 14.715 0 .0097 9826 87.671

Co(CaxWD) 8 17434 2179.2 20.093 0.0001 9788 224.54

CaxWDxSD 4 8270.4 2067.6 19.064 0.0003 9832 337.04

Res 31 33621 1084.6 1084.6

Total 58 1.30E+05

PERMANOVA full 4-factor m odel te s t fo r differences b e tw e en canyon areas (Ca: Gollum and W hittard), w ate r d ep th s (WD: 700 and 1000 m), sed im en t d ep th s (SD: 1 ,2 ,3 , 4, 5 cm), cores (Co: 1-12), and interaction term s. Bold italic values indicate significant differences a t p < 0 .01 . Data w as s tandard ised and sq u are roo t transform ed; resem blance w as ca lcu la ted using Bray-Curtis. doi:10.1371 /journal.pone.0020094.t001

PLoS ONE I w ww.plosone.org May 2011 | Volume 6 | Issue 5 | e20094



T a b le 2 . Dominance list o f nematode genera per station.

G 1000 % G 700 % W 1 0 0 0 % W 7 0 0 %

Sabatieria 27.0 Sabatieria 16.2 Leptolaim us 14.4 Astom onem a 10.3

Halalaimus 4.9 Syringolaimus 8.2 M olgo la im us 13.5 Halalaimus 8.9

Acantho la im us 4.5 Halalaimus 6.8 Acantho la im us 9.1 Acantholaim us 7.0

Cervonema 3.6 Trefusia 4.5 Halalaimus 5.1 Sabatieria 4.5

Syringolaimus 3.5 Acantho la im us 3.8 M icrolaim us 5.0 Terschellingia 3.9

Viscosia 3.0 M etadesmolaimus 3.1 M onhystre lla 4.9 M etasphaerolaimus 3.4

Trefusia 2.6 Cervonema 2.8 Syringolaimus 4.5 M etadesmolaimus 2.7

Setosabatieria 2.3 M etadasynemella 2.2 Desmoscolex 3.7 aff. Nannolaim us 2.3

Actinonem a 2.3 Actinonem a 2.1 Sabatieria 3.5 Actinonem a 2.2

M olgolaim us 2.0 Desmodora 2.0 Cervonema 3.2 Cervonema 2.0

Setosabatieria 2.0 Campylaimus 2.5

G enera w ith ab u n d an ce -2%i. Mean values p resen ted here based on relative abundances correc ted for slice density differences. doi:10.1371 /journal.pone.0020094.t002

higher (1054-1426 ind. 10 c m " ', 89 .2 -276 .8 pg dwt) in the W hittard an d Gollum area com pared to values reported for areas such as the Iberian M argin or the M ed iterranean Sea a t similar depths (see Soltwedel [57] for review), bu t similar to values observed previously along the Celtic M argin [58,61], In contrast, at the G oban Spur, in betw een the Gollum and W hittard area, V anaverbeke et al. [62] found m eiofauna densities o f 612 and 619 ind. 10 cm ~ at 670 and 1034 m dep th respectively an d suggested that lower O M input was responsible for the lower densities com pared to neighbouring areas th a t receive h igher am ounts o f phytodetritus. O n a global scale, the Gollum an d W hittard stations exhibit m uch greater m eiobenthic standing stocks w hen com pared to trenches/canyons and similar dep th zones [60], E nhanced accum ulation o f O M in topographically enclosed areas o r can y o n / channel structures is a well-known phenom enon [5,6,60,63,64], and is often seen to result in benthic abundance an d biomass hotspots [11,60,65,66,67]. For the Gollum and W hittard area, the positive control o f food sources on m eiofauna standing stocks is illustrated by the strong correlations o f nem atode densities and

biomass with the quality and bioavailability o f the O M (Table 6) and the high contributions o f freshness (chl-a:phaeo) and C P E :T O C in explaining variation o f total nem atode densities and biomass (Table S6). How ever, the correlations were m ore frequent and stronger in the Gollum C hannel w here currents are capable o f producing significant turbidity com pared to the W hittard study area, w here hydrodynam ic d isturbance is less intense an d less frequent [26,33,35]. This suggests that the relation betw een nem atode abundance an d biomass, an d quantity , quality and bioavailability o f the present food is stronger when hydrodynam ic d isturbance is greater, providing that the disturbance is no t o f such proportions that it is able to preclude the establishm ent o f m eiobenthic com m unities. Interestingly, Bian- chelli et al. [12] suggested for the Bari C anyon (Adriatic Sea) that the trophic relations in canyon sediments are tightly connected with the hydrodynam ic conditions an d th a t such a relation becom es tighter as the hydrodynam ic forcing increases. Even though this hypothesis needs to be substantiated by m ore observations from different canyon and channel systems, it

T a b le 3. Results from analysis o f similarities and species contributions (SIMPER).

SIM PER s im ila r ity {% , c o n tr ib u tio n > 4 % )

Gollum 51.5 W hittard 53.8 700 m 49.9 1000 m 55.5

Sabatieria 14.5 Sabatieria 5.5 Sabatieria 9.7 Sabatieria 10.1

Trefusia 5.9 Syringolaimus 5.4 Astom onem a 5.2 Syringolaimus 5.9

Syringolaim us 5.0 M onhystre lla 4.8 Syringolaimus 4.4 M icrolaim us 5.5

Astom onem a 4.8 Trefusia 4.0 M onhystrella 5.1

Leptolaimus 4.7 Leptolaim us 4.5

M icrolaim us 4.5 Halalaimus 4.5

Halalaimus 4.5 Acantho la im us 4.4

Trefusia 4.3

SIM PER d is s im ila r ity (% , c o n tr ib u tio n > 3 % )

Gollum vs. W hittard 59.4 700 m vs. 1000 m 57.5

Sabatieria 4.9 Astom onem a 4.9

Astom onem a 4.6

Individual g en u s cu t-o ff level w as 4% fo r similarity analysis, 3% for dissimilarity doi:10.1371 /journal.pone.0020094.t003

analysis.

8 . PLoS ONE I w ww.plosone.org May 2011 | Volume 6 | Issue 5 | e20094



T a b le 4 . Results from univariate PERMANOVA analysis for differences in biomass.

S ource d f SS M S Pseudo-F P (p erm ) perm s

Ca 1 200.88 200.88 31.914 0 .0 2 7 261

WD 1 13.274 13.274 0.21089 0.605 258

CaxWD 1 30.117 30.117 47.848 0.096 263

Res 4 25.178 62.944

Total 7 257.5

PERMANOVA te s t for biom ass data (rig dw t) b e tw een canyon areas (Ca: Gollum and W hittard), w ate r d ep th s (WD: 700 and 1000 m), sed im en t d ep th s (SD: 1, 2, 3, 4, 5 cm), cores (Co: 1-12), and interaction term s. Data w as fourth -roo t transform ed; resem blance was calculated using Bray-Curtis. Bold values indicate significant differences a t p < 0 .05 . doi:10.1371 /journal.pone.0020094.t004

corroborates the Dynam ic E quilibrium M odel and there are o ther indications for its validity. D isturbance o f deep-sea, soft-bottom com m unities at an interm ediate level, i.e. w hen no t detrim ental to the sustenance o f the com m unity, can explain high benthic standing stocks because o f the replenishm ent o f available food sources th rough cycles o f resuspension an d deposition o f O M [68].

H igher rates o f hydrodynam ic d isturbance m ay therefore enhance food availability and intensify the positive relation betw een food sources in the sedim ent and benthic standing stocks. A nother possible explanation for the high m eiofauna num bers an d biomass at the Gollum and W hittard stations m ay be the w arm ing influence o f the M ed iterranean Outflow W ater (M O W , [33]) and its positive effect on bacterial biomass production . Since m any deep-sea nem atodes are specific bacteria feeders o r general m icrobial feeders, enhanced bacterial biom ass under the influence o f the w arm er and m ore saline M O W com pared to o ther locations at similar depths, m ay result in an increase o f suitable organic food sources an d lead to greater standing stocks [69].

T h e C :N ratio can give an indication of w hether the sedim entary O M is derived from a terrestrial source o r w hether it is o f a m arine origin. How ever, preferential rem oval o f labile, proteinaceous (i.e. N-rich) m arine O M during degradation processes m ay also lead to h igher C :N ratios. A t station G1000 and W 700, C :N values were very high, indicating the presence of either large am ounts o f terrigenous m aterial o r highly degraded O M . O u r sam pling sites are not located n ear a river system from w hich they could receive substantial fluvial export, as is the case for o ther canyon systems such as the N azaré an d Setúbal canyons at the Portuguese M argin. It is therefore very likely th a t enhanced degradation of the O M at these stations is responsible for the high

T a b le 5 . Structural and functional diversity indices per sediment layer and total per station.

Site S e d im e n t d e p th (cm ) T D ( ® -1 ) Ho H 1 h 2 Hinf EG (51)

G 700 0-1 3.67 86 67.0 52.3 20.2 39

G 700 1-2 3.20 63 43.4 29.5 9.3 24

G 700 2 -3 3.26 70 53.4 38.8 10.6 34

G 700 3 4 3.07 45 37.7 31.2 12.7 45

G 700 4 -5 2.83 45 36.1 29.2 11.8 33

G 700 0 -5 3 .7 8 127 4 6 .8 2 0 .8 6 .2 27

G1000 0-1 3.63 70 54.7 44.7 19.5 37

G1000 1-2 2.72 57 39.2 24.4 6.9 26

G1000 2-3 2.62 58 42.5 27.1 7.2 29

G1000 3 4 2.58 49 35.9 25.1 7.6 27

G1000 4 -5 3.20 52 38.8 29.3 10.9 25

G 1000 0 -5 3 .4 3 108 3 3 .5 11 .5 3 .7 2 6

G o llu m 3.6 7 152 4 3 .8 15 .9 4 .6 2 6

W700 0-1 3.39 73 57.7 45.0 16.0 39

W700 1-2 4.00 84 65.8 52.8 19.9 33

W700 2-3 4.21 62 46.4 33.8 11.5 32

W700 3 -4 3.58 42 29.7 20.4 7.4 26

W700 4 -5 2.96 37 22.1 13.5 4.9 14

W 7 0 0 0 -5 4 .3 7 112 5 0 .8 2 8 .2 9 .7 30

W1000 0-1 2.78 50 34.9 25.4 10.9 27

W1000 1-2 3.12 56 40.1 29.8 12.1 27

W1000 2-3 3.57 43 30.7 23.1 10.7 25

W1000 3 -4 3.70 40 33.0 28.2 12.8 30

W1000 4 -5 3.29 35 28.4 23.8 12.2 31

W 1 0 0 0 0 -5 3 .1 0 8 4 2 7 .6 15 .7 6 .9 21

W h itta rd 3 .7 5 132 4 8 .7 2 8 .3 1 2 .4 29

G o llu m -W h itta rd 3 .8 6 181 5 3 .9 2 6 .8 7.8 29

Com m unity ab u n d a n ce data w as averaged over replicates. For to tal values, relative abu n d an ces w ere first correc ted for density differences per sed im en t layer. doi:10.1371 /journal.pone.0020094.t005

PLoS ONE I w ww.plosone.org May 2011 | Volume 6 | Issue 5 | e20094



u nt¡ 0

<aos

-5

6.91.9

log(cnl-a:ahaeo+0.1)5.5

TOC'

log(chl-a:TOC+0.1)

0.5TN log(chl-a+0.1)

3.00.9

VW Mean

©

0 0Relative abundance: ^Astomonema vStandardise Samples by Total Transform: Square root Resemblance: Euclidean distance

-10 - 5 0

dbRDAl (40.4% of fitted, 26% of total variation)

10

Figure 5. Distance-based redundancy (dbRDA) bubble plot illustrating the DISTLM model based on the genera assemblage data and fitted environm ental variables with their vector (strength and direction of effect o f the variable on the ordination plot). Axis legends include % of variation explained by the fitted m odel and % of total variation explained by the axis. Relative abundance o f the 'chem osynthetic ' Astom onem a nem atodes is visualised by the size of the circles, including the value of their relative abundance in each sam ple. Chl-a: chlorophyll a, CPE: chloroplastic p igm ent equivalents, Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating 'freshness' of the phytodetrital OM, TN: total nitrogen con ten t, TOC: total organic carbon con ten t, C:N: molar carbon-nitrogen ratio, Chl-a:TOC: chlorophyll a divided by total organic carbon co n ten t indicating bioavailability o f the bulk OM, CPE:TOC: total p igm ent derived m atter relative to the bulk OM, VW Mean: volum e w eighted m ean grain size. doi:10.1371/journal.pone.0020094.g005

C :N ratios observed. E nhanced aerobic decay o f high organic loads m ay lead to reducing conditions and lim ited oxygen availability, and ultim ately lower infauna abundances and biomass beyond expected for the observed food input [23,70,71], bu t also change the com position o f the resident fauna. This is illustrated by the biogeochem ical conditions a t station W 700. Even though sediments at W 700 displayed greatest levels o f T O C and C PE and highest chl-a:phaeo and ch l-a :T O C values (Tig. 2, 3), m eiofauna densities an d nem atode biom ass were not h igher com pared to the o ther stations (Fig. 2), possibly due to oxygen lim itation in the sediment. Fu rther argum ents for this are given in the discussion of the nem atode com m unity assemblage o f station W 700.

C u rren ts p referen tia lly tran sp o rt finer g ra ined sedim ents, causing the rem ain in g sedim ent to becom e relatively coarser [72]. T h is is in ag reem en t w ith coarser sedim ents be ing p resen t a t the G ollum stations com p ared to the W h itta rd stations w hen considering cu rren t activity in these areas, an d reflects the G ollum C h an n e l’s role as p referen tia l condu it for the tran sp o rt o f sedim ent from shelf dep ths to the deep sea. Since cu rren ts in the G ollum C hannels a re able to tran sp o rt the ligh ter sedim ent fractions dow n canyon, the u p p e r reaches o f this ch annel system will be ch aracte rised by the coarser rem ain in g fraction , as illustra ted by the g ra in size differences betw een G 700 an d G 1000 (Fig. 3). F o r the G ollum stations, nem ato d e densities an d biom ass w ere negatively co rre la ted w ith sand co n ten t o f the sedim ents. Flowever, we expect th a t this is no t pu re ly a causal re la tion b u t ra th e r a result o f a decrease in g rain size w ith sedim ent d ep th coincid ing w ith decreasing n em atode a b u n dance a n d biom ass.

N e m a to d e c o m m u n i ty s t ru c tu re a n d func t ion , a n d the i r re la tion w ith e n v iro n m e n ta l co n d i t io n s

T h e generic assemblages were significantly different betw een the Gollum and W hittard stations, betw een 700 and 1000 m w ater depth, and betw een individual stations (Table 1). Furtherm ore, nem atode com position differed significantly betw een different sedim ent layers (Table 1), w hich was the m ost im portan t factor across canyons an d w ater depths as indicated by the estim ated com ponents o f variation (Table 1) and illustrated in the M D S plot (Fig. 4). T h e contrasting com m unity differences betw een sediment layers suggest that biogeochem ical gradients in the sedim ent have a greater im pact on nem atode com m unity structure than gradients related to canyon area, w ater depth, and station differences. M oreover, assem blage differences in the deeper sedim ent layers were m ore pronounced th an a t the surface, a p a tte rn th a t was also observed for the N azaré, Cascais an d Setúbal canyons and the open slope a t the Portuguese M argin [11,66]. T h e prevailing effect o f sedim ent dep th is m ost p robably related to the differential ability o f different genera to reside in o r penetrate different sedim ent layers and has been associated w ith oxygen availability in the sediments [23,70], Striking in this respect is the occurrence o f high num bers o f Astomonema southwardorum at station W 700 and G 1000, especially in the deeper sediment layers. N em atodes o f the genus Astomonema have a degenerated alim entary canal; they posses no m outh, no buccal cavity, nor pharynx, and they have a rud im entary gut containing m icro-organism s [73,74,75,76], ETntil now, -4. southwardorum h ad only been found in a m ethane seep pockm ark in the N orth Sea [73]. M outhless an d gutless nem atodes carrying endosym bionts have been observed previously in near-




T a b le 6 . Kendall-Tau correlation coefficients between biotic and abiotic variables.

P aram eter C hl-a CPE C h l-a :p h a eo TN TOC C:N Chl-a:TOC CPE:TOC C lay % S ilt % S a n d %

G ollu m

Nem aDens 0 .653 0 .835 0 .5 4 4 0 .6 0 0 0.378 0.244 0 .5 9 9 0.740 0.333 0 .511 - 0 . 5 5 6

C hem os. RA -0 .2 1 6 -0 .1 6 2 -0 .3 3 9 0.479 0 .6 3 0 0 .5 8 0 -0 .2 7 8 -0 .271 0.731 0.428 -0 .3 7 8

NemaBiom 0 .5 4 4 0.692 0 .5 4 4 0.467 0.244 0.111 0 .4 9 0 0.692 0.289 0.467 - 0 .5 1 1

Chem os.Biomass

0.031 0.054 -0 .0 9 3 0 .5 8 0 0.731 0 .5 8 0 -0 .031 -0 .0 5 4 0 .6 3 0 0 .6 3 0 - 0 .5 8 0

TD 0.435 0.358 0.435 -0 .0 6 7 0.067 0.022 0 .4 9 0 0.454 -0 .1 5 6 -0 .0 6 7 0.022

H0 0 .5 8 5 0 .6 3 5 0 .5 8 5 0.159 0.114 0.159 0.640 0.733 0.068 0.159 -0 .2 0 5

H, 0 .5 9 9 0.645 0 .5 9 9 0.111 0.067 0.111 0.653 0.740 0.022 0.111 -0 .1 5 6

EG (51) 0.109 0.072 0.000 -0 .2 0 0 -0 .2 4 4 -0 .0 2 2 0.054 0.072 -0 .3 7 8 -0 .2 0 0 0.156

W h ittard

N em aDens 0.368 0.405 0 .5 2 9 0.200 -0 .1 5 6 -0 .2 4 4 0 .5 5 2 0 .4 9 4 - 0 .6 4 4 -0 .0 2 2 0.022

Chem os. RA 0.420 0.290 0.162 0.406 0.835 0 .788 0.222 0.193 0.215 0 .501 - 0 .5 0 1

NemaBiom 0 .5 0 6 0 .5 3 9 0 .5 2 9 0.378 0.111 0.022 0 .5 0 6 0 .5 3 9 -0 .4 6 7 0.244 -0 .2 4 4

Chem os. 0.469 0.338 0.217 0.358 0.788 0 .740 0.272 0.145 0.263 0.454 - 0 .5 4 9Biomass

TD -0 .0 4 6 0.090 0.227 0.200 0.289 0.378 -0 .2 3 0 -0 .0 9 0 0.244 0.422 -0 .4 2 2

H „ 0.460 0 .4 9 4 0 .5 8 0 0.289 0.022 - 0.067 0.368 0.405 - 0.378 0.244 - 0.244

H i 0.368 0.405 0 .5 2 9 0.244 -0 .0 2 2 -0 .111 0.276 0.315 -0 .2 4 4 0.200 -0 .2 0 0

EG(51) 0.256 0.386 0 .5 1 0 0 .4 9 4 0.135 0.045 0.163 0.295 0.135 0.405 - 0.405

Chl-a: chlorophyll a, CPE: chloroplastic p igm en t equivalents, Chl-a:phaeo: chlorophyll a divided by its deg rada tion p roducts (phaeophytines) indicating 'freshness ' o f th e phytodetrita l OM, TN: to ta l n itrogen co n ten t, TOC: to tal organic carbon con ten t, C:N: m olar ca rbon-n itrogen ratio, Chl-a:TOC: chlorophyll a divided by to ta l organic carbon co n ten t indicating bioavailability o f th e bulk OM, CPE:TOC: to tal p igm en t derived m a tte r relative to th e bulk OM. NemaDens: n em a to d e abundance , Chem os. RA: relative ab u n d a n ce o f th e chem osyn thetic Astom onem a n em atodes , NemaBiom: to tal n em a to d e biom ass, C hem os. Biomass: to tal b iom ass o f th e chem osyn thetic Astom onem a n em atodes . Bold values d e n o te significance a t th e p < 0 .0 5 level. Bold, italic values d e n o te significance a t th e p <0.01 level. doi:10.1371 /journal.pone.0020094.t006

anoxic, deep-sea environm ents rich in O M [77], bu t this is the first tim e that Astomonema has been recovered from the deep-sea. T he genus Astomonema is m orphologically closely related to the genus Parastomonema, and has been classified in the same subfamily A stom onem atinae [78], T h e prokaryotic endosym bionts associated with A. southwardorum a re now recognised to be sulphur- oxidising chem olithoautotrophs, w hich reduce sulphur com pounds as an energy source to provide their nem atode hosts w ith nutrition [74,75,79]. T h e abundance a í Astomonema was positively correlated with T O C , T N , and C :N values an d the variation in its abundance and biomass was m ainly explained by T O C and T N content in the sediments (Table S6), suggesting that they are directly or indirectly dependent on the quantity an d level o f decom position o f O M . No correlations with any chl-a derived variable was observed (Table 6), nor were there im portan t contributions o f chl-a derived variables in explaining its abundance and biom ass variability (Table S6), indicating their independence o f fresh food sources. E nhanced aerobic respiration o f great am ounts o f O M m ay lead to reducing conditions w hereby anoxic m ineralisation leads to reduced byproducts such as sulphides. T h e presence of the ‘chem osynthetic’ Astomonema at W 700 and G 1000, w here O M levels and C :N ratios (as a m easure o f O M degradation) are greatest, supports the idea of a sulphidic biom e in reducing sedim ent conditions being present at these stations. T his is also attested by higher Astomonema abundances in deeper sedim ent layers, w here oxygen availability is likely lower an d the concentration of sulphidic com pounds higher th an a t the surface. Such environm ents constitute ideal habitats for Astomonema nem atodes to thrive in, as their endosym biotic, thiotrophic bacteria m ay provide them w ith a competitive advantage over o ther feeding types. M oreover, their filiform shape

is characteristic for nem atodes in oxygen-lim ited environm ents and m ay result in increased m obility, w hich enables these organism s to m igrate betw een anoxic patches and parts o f the sedim ent w here oxygen is available a n d /o r to bridge the gap betw een oxic and anoxic patches w ith their body [23]. T he presence of the ‘chem osynthetic’ Astomonema at stations w ith high organic input is in agreem ent w ith the T R O X m odel [22] in the sense that in eutrophic systems, the faunal com m unity seems oxygen driven as opposed to a food-driven com m unity in oligotrophic conditions. Astomonema is able to exploit the reduced environm ent w here patchy oxygen lim itation hinders the survival o f other, less tolerant genera and can use reduced com pounds as a food resource th rough an endosym biotic pathw ay. A com plem entary explanation for the prevalence of Astomonema in deeper sedim ent layers could be that in deeper layers they w ould be better pro tected against possible predators. Interestingly, m any o f the Astomonema individuals exhibited different stages o f w ound healing at their posterior end, an indication that these nem atodes are being preyed on, bu t have the ability to escape their attackers. Possibly, their long, flexible bodies enable them to avoid being killed th rough hind-body loss w hich m ay be facilitated by cuticle structure in o rder to survive p redation (Tchesunov, unpublished data). T h e same phenom enon of h ind-body w ound healing is found with typically very long Siphonolaimus individuals from the W hite Sea (Tchesunov, unpublished data).

D om inant genera differed greatly betw een stations (Table 2). W hilst Sabatieria consistently dom inated at the Gollum stations, the nem atode assemblage a t W 700 was controlled by the genus Astomonema and at W 1000 by the genera Leptolaimus and Molgolaimus (Table 2). M any Sabatieria species are adap ted to




exploit newly available habitats o r resources [13,80] and are typically found in unpredictable and variable environm ents [38,81]. T h e dom inance of Sabatieria a t the disturbed and h y d ro d y n a m ica ^ active Gollum stations is thus no t surprising. In contrast, Leptolaimus an d Molgolaimus, the dom inan t genera at W 1000, are usually no t regarded as opportunistic b u t ra ther as persisters (cf. definitions sensu Bongers e t al. [82]) and are not considered able to endure intense sedim ent d isturbance [80]. This suggests th a t sediments a t this station are m uch m ore stable and less d isturbed com pared to those a t the Gollum stations, w hich is confirm ed by our knowledge on hydrodynam ic activity in these areas [33]. O n the o ther hand, ex-situ colonisation experim ents investigating the ability o f nem atodes to actively colonise defaunated sediments revealed th a t Leptolaimus nem atodes m ay be considered active colonisers [83]. How ever, the Sabatieria and Leptolaimus species found to colonise the defaunated sediments [83] often occur in deeper, hypoxic to anoxic sedim ent layers [84], M oreover, the defaunated sediments used for the experim ent originated from deeper sedim ent layers (> 8 cm depth) and biogeochem ical characteristics such as oxygen availability m ay have been different com pared to natu ral surface and subsurface sediments. H ence, the preferential m igration o f Leptolaimus to the defaunated sediments m ay have been triggered by the biogeochem ical properties o f the substratum used.

G enus richness (Hq) at all four stations (84—127 genera) was very high in com parison with reported values by V anaverbeke et al. [62] for similar depths a t the G oban Spur (45-60). Even when com paring H q from the present study w ith values from a cold- w ater coral hab ita t in the Porcupine Seabight [85], w hich is globally considered one o f the m ost diverse nem atode habitats [86], genus richness is still m oderately h igher a t the Gollum and W hittard stations (91-106 and 84-127 for the cold-water coral hab ita t and the present study, respectively). For the Gollum area strong positive correlations were found betw een Hq, Hq and chl-a, CPE, chl- a:phaeo, and chl-a :T O C , indicating th a t structural diversity is related to the quality an d bioavailability o f the O M present (Table 6). In contrast, correlations in the W hittard area were m uch less p ronounced o r no t significant. This suggests th a t the underlying causes for structural nem atode diversity have a differential effect in these two study areas. M ore im portandy, it indicates th a t diversity in the Gollum C hannels m ay be m ore dependent on the trophic conditions (i.e. quantity, quality, and bioavailability o f the O M ) than is the case for the W hittard area. T his difference m ay reflect the differential hydrodynam ic regimes in the two areas. It suggests that increased hydrodynam ic activity, as in the Gollum Channels, enhances the relation betw een structural diversity an d food quality and availability, as was the case for nem atode abundance and biomass. F rom this perspective, the interm ediate disturbance hypothesis, w hich is generally applicable to deep-sea comm unities [68], can be evoked. An interm ediate disturbance frequency m ay prom ote high diversity, providing there is enough time betw een disturbance events for species to colonise and establish themselves in the new habitats. T oo m uch tim e betw een disturbance events, and competitive interactions an d exclusion m ay cause diversity to decline. C om bined with enhanced food availability because of cycles o f resuspension an d deposition caused by recurring hydrodynam ic disturbance, these factors m ay explain the tighter positive relation betw een nem atode structural diversity and trophic conditions in sediments a t the Gollum stations com pared to the W hittard C anyon.

N em atode trophic diversity (TD) was correlated with chl-a:T O C a t the Gollum stations and this was no t the case for the W hittard stations. This m ay imply th a t T D is differentially influenced by trophic conditions w hen com paring bo th areas. How ever, the lack

o f correlations betw een T D an d o ther trophic param eters suggests that, if such a contrast is even present, it occurs to a m uch lesser extent. Both structural and trophic diversity were lower a t the deeper stations than a t the shallower stations w ithin each study area, and bo th seemed to be controlled to some extent by trophic conditions (see D IS T L M results, T able S6), som ething w hich is also supported by the observation th a t the quality o f food a t the deeper stations is lower th an a t the shallower stations (Fig. 2, 3). O n the o ther hand, the high am ounts and freshness o f the O M , and the prevalence o f Astomonema a t W 700 (resulting in a high T D and indicative o f the use o f reduced com pounds as food source) agree with the principle o f m ore trophic levels being present w hen food input is greater [87]. This is also confirm ed by higher T D a t sites with higher organic input caused by greater relative abundances and biomass o f predators/scavengers [11]. W ith this in m ind, it can be argued that the ecofunctional im portance of Astomonema in the com m unity is transferring the energy contained in available sulphur com pounds into nem atode biomass so th a t the high trophic complexity o f the com m unity is sustained by providing a food source to nem atode predators/scavengers and m acrofauna. T he presence of Astomonema as a result o f local trophic conditions indicates the im portance o f small-scale heterogeneity and biogeochem ical patterns along the vertical sedim ent profile for structural and trophic diversity in canyon environm ents.

Supporting InformationFigure SI Draftsman plot or scatter plot matrix of environmental variables to assess collinearity between environmental variables. Chl-a: chlorophyll a, CPE: chloroplastic pigm ent equivalents, Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating ‘freshness’ o f the phytodetrital O M , T N : total n itrogen content, T O C : total organic carbon content, C:N: m olar carbon-nitrogen ratio, Chl- a :T O C : chlorophyll a divided by total organic carbon content indicating bioavailability o f the bulk O M , C P E :T O C : total pigm ent derived m atter relative to the bulk O M , V W M ean: volum e w eighted m ean grain size.(EPS)

Table SI Results from univariate PERMANOVA analyses for differences in sedimentary abiotic variables.PE R M A N O V A test for canyon areas (Ca: Gollum and W hittard), w ater depths (WD: 700 an d 1000 m), sedim ent depths (SD: 1, 2, 3, 4, 5 cm), cores (Co: 1-12), and interaction term s. Chl-a: chlorophyll a, CPE: chloroplastic pigm ent equivalents, Chl- a:phaeo: chlorophyll a divided by its degradation products (phaeopigments) indicating ‘freshness’ o f the phytodetrital O M , C hl-a:T O C : chlorophyll a divided by total organic carbon content indicating bioavailability o f the bulk O M ; T O C : total organic carbon, T N : total nitrogen, C:N: m olar ca rb o n /n itro g en ratio, M ean grain size: volum e weighted grain size. D a ta was norm alised; resem blance was calculated using Euclidean D istance. Bold values indicate significant differences a t p < 0 .0 5 , bold italic values indicate significant differences a t p < 0 .0 1 .(DO CX)

Table S2 Meiofauna taxa abundances per sediment layer. Values denote m eans o f replicates (ind. 10 cm -2 ). St. Dev.: standard deviation. G1000: Gollum C hannel a t 1090 m depth, G700: Gollum C hannel a t 755 m depth, W 1000: W hittard C anyon a t 1160 m depth, W 700: W hittard C anyon a t 762 m depth. (DO CX)

Table S3 Results from pair-wise multivariate PERMANOVA analyses for differences in nematode community




structure. PE R M A N O V A pair-wise com parisons w ithin each of the C a x W D xSD levels as p a rt o f the repeated m easures analysis, including M onte-C arlo perm utation p values (P(MC)). T h e p a irwise com parisons give an indication of the individual contribution to the significant three-w ay interaction term (C ax W D x S D ) with differences betw een canyon areas (Ca: Gollum and W hittard), w ater depths (WD: 700 and 1000 m), sedim ent depths (SD: 1, 2, 3, 4, 5 cm). D a ta was standardised an d square roo t transform ed; resem blance was calculated using Bray-Curtis. Bold values indicate significant differences a t p < 0 .0 5 .(DO CX)

Table S4 Spearman correlation coefficients between environmental variables. Chl-a: chlorophyll a, CPE: chloroplastic pigm ent equivalents, Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating ‘freshness’ o f the phytodetrital O M , T N : total n itrogen content, T O C : total organic carbon content, C :N: m olar carbon-nitrogen ratio, Chl- a :T O C : chlorophyll a divided by total organic carbon content indicating bioavailability o f the bulk O M , C P E :T O C : total pigm ent derived m atter relative to the bulk O M .(DO CX)

Table S5 Distance-based linear model (DISTLM) for genera assemblages and selected environmental variables. Variables: Selected environm ental variables used to calculate the optim um model. M arginal tests: explanation of variation for each variable taken alone. Sequential tests: conditional tests o f individual variables in constructing the model. Each test examines whether adding the variable contributes significandy to the explained variation. Selection procedure: step-wise, selection criterion: adjusted R 2. Prop.: % variation explained. Cumul.: cumulative variation explained. Chl-a: chlorophyll a, CPE: chloroplastic pigm ent equivalents, Chl-a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating ‘freshness’ o f the phytodetrital O M , TN : total nitrogen content, T O C : total organic carbon content, C:N: m olar carbon-nitrogen ratio, C hl-a:TO C : chlorophyll a divided by total organic carbon content indicating bioavailability o f the bulk O M , C PE :TO C : total pigm ent derived m atter relative to the bulk O M , M ean grain size: volume weighted m ean grain size.(DO CX)

References1. L a m p itt R S , R a in e R C T , B illett D S M , R ice A L (1995) M a te r ia l supp ly to the

E u ro p e a n co n tin e n ta l slope: A b u d g e t b a sed o n b e n th ic oxygen d e m a n d a n d o rg an ic supply. D e e p -S e a R e se a rc h P a r t I: O c e a n o g rap h ic R e se a rc h P a p e rs 42:1 8 6 5 -1 8 7 3 .

2. L o n g h u rs t A , S a th y e n d ra n a th S, P la tt T , C av erh ill C (1995) A n estim ate o f g lobal p r im a ry p ro d u c tio n in th e o c e an f ro m satellite rad io m ete r d a ta . J o u rn a l o f P la n k to n R e se a rc h 17: 1 2 4 5 -1 2 7 1 .

3. L a m p itt R S , A n tia A N (1997) P artic le flux in d eep seas: reg io n a l charac te ristics a n d tem p o ra l variab ility . D e e p -S e a R e se a rc h P a r t I: O c e a n o g rap h ic R esea rch P ap e rs 44: 1 3 7 7 -1 4 0 3 .

4. C u n n in g h a m M J, H o d g so n S, M asso n D G , P a rso n L M (2005) A n ev a lu a tio n o f along- a n d dow n-slope sed im en t t ra n sp o rt p rocesses b e tw een G o b a n S p u r a n d B re n o t S p u r o n th e C eltic M a rg in o f th e Bay o f Biscay. S ed im en tary G eology 179: 9 9 -1 1 6 .

5. A cco rn e ro A, P ico n P, d e Bovee F, C h a rrie re B, B uscail R (2003) O rg a n ic c a rb o n b u d g e t a t th e sed im en t-w a ter in te rface o n the G u lf o f L ions co n tin e n ta l m arg in . C o n tin e n ta l S h e lf R e se a rc h 23: 79—92.

6. T h o m se n L, v an W ee rin g T , G u s t G (2002) Processes in th e b e n th ic b o u n d a ry lay e r a t th e Ib e r ia n co n tin e n ta l m arg in a n d th e ir im p lica tio n fo r ca rb o n m in e raliza tio n . P rogress in O c e a n o g ra p h y 52: 315—329.

7. Ivanov V V , Shapiro G I, H u th n a n c e J M , A leynik D L , G olovin P N (2004) C ascades o f dense w ater a ro u n d the w o rld ocean . P rogress in O cean o g rap h y 60: 47—98.

8. H ill A E , S o u za A J, J o n e s K , S im p so n J H , S h ap iro G I, e t al. (1998) T h e M alin cascad e in w in ter 1996. J o u rn a l o f M a rin e R e se a rc h 56: 87—106.

9. B ianchelli S, G a m b i C , Z eppilli D , D a n o v a ro R (2010) M e ta z o a n m eio fau n a in deep -sea can y o n s a n d a d ja c e n t o p e n slopes: A large-sca le c o m p a riso n w ith focus o n th e ra re taxa. D eep -S ea R e se a rc h P a r t I: O c e a n o g ra p h ic R e se a rc h P ap ers 57: 4 2 0 ^ 3 3 .

Table S6 Distance-based linear models (DISTLM) - Sequential tests for best fitting model for each o f the univariate biotic parameters. Sequential tests: conditional tests o f individual variables in constructing the m odel. E ach test exam ines w hether add ing the variable contributes significandy to the explained variation. Selection procedure: step-wise, selection criterion: adjusted R 2. Prop.: % variation explained. Cum ul.: cum ulative variation explained. N em aD ens: nem atode abundance, Chem os. RA: relative abundance o f the chem osynthetic Astomonema nem atodes, Nem aBiom : total nem atode biomass, Chem os. Biomass: total biomass o f the chem osynthetic Astomonema nem atodes, T D : trophic diversity, H g: genus richness, H ¡\ H ill’s diversity index, EG(51): expected num ber o f genera. Chl-a: chlorophyll a, CPE: chloroplastic pigm ent equivalents, Chl- a:phaeo: chlorophyll a divided by its degradation products (phaeophytines) indicating ‘freshness’ o f the phytodetrital O M , T N : total n itrogen content, T O C : total organic carbon content, C:N: m olar carbon-nitrogen ratio, C h l-a:T O C : chlorophyll a divided by total organic carbon content indicating bioavailability o f the bulk O M , C P E :T O C : total pigm ent derived m atter relative to the bulk O M , M ean grain size: volum e w eighted m ean grain size.(DO CX)

AcknowledgmentsW e are very m uch indeb ted to co m m an d er a n d crew o f the R V Belgica for the ir excellent w ork du ring the Belgica 2 0 0 6 /1 3 cruise. T h e first au th o r w ould like to th an k the scientific crew fo r the ir m uch req u ired help du ring the sam pling cruise an d D . V a n Rooij fo r his help w ith m aps o f the study area. Special thanks go to K . R . C larke an d M . A nderson fo r their thoughtfu l com m ents an d suggestions on the statistical analyses perfo rm ed in this study. T w o anonym ous reviewers a re acknow ledged fo r th e ir helpful com m ents on an earlier version o f the m anuscrip t.

Author ContributionsA nalyzed the data: J I . C o n trib u ted reag en ts/m a te ria ls /an a ly sis tools: AV. W ro te the paper: J I A V T AV . Investigated all m eiofauna a n d nem atode specim ens: J I . Investigated nem atode specim ens characterised by endosym biotic bacteria: A V T .

10. G a rc ia R , K o h o K A , D e S tig ter H C , E p p in g E , K o n in g E , e t al. (2007) D is tr ib u tio n o f m e io b en th o s in th e N a z a re can y o n a n d a d ja c e n t slope {western Ib e r ia n M arg in ) in re la tio n to sed im en ta ry co m position . M a rin e Ecology- P rogress Series 340: 2 0 7 -2 2 0 .

11. Ingels J , K iriakou lak is K , W olff G A , V a n re u se l A (2009) N e m a to d e d iversity a n d its re la tio n to q u a n tity a n d q u a lity o f se d im e n ta ry o rg an ic m a tte r in th e N a z a ré C an y o n , W e ste rn Ib e r ian M arg in . D eep -S ea R e se a rc h P a r t I: O c e a n o g rap h ic R e se a rc h P a p e rs 56: 1 5 2 1 -1 5 3 9 .

12. B ianchelli S, G a m b i C , P u sced d u A , D a n o v a ro R (2008) T ro p h ic co n d itio n s a n d m eio fau n a l assem blages in th e B ari C a n y o n a n d th e a d ja c e n t o p e n slope {Adriatic Sea). C h e m is try a n d E cology 24: 10 1 -1 0 9 .

13. S c h ra tz b e rg e r M , L a m p ad a r io u N , S om erfie ld P , V a n d e p itte L, V a n d e n B erghe E {2009) T h e im p a c t o f seab ed d is tu rb an c e o n n e m a to d e com m unities: lin k in g field a n d lab o ra to ry observations. M a rin e Biology 156: 709—724.

14. S c h ra tz b e rg e r M , W arw ick R M {1998) Effects o f physical d is tu rb an c e o n n e m a to d e c o m m u n ities in san d a n d m u d : a m ic ro co sm ex p e rim en t. M arin e B iology 130: 6 4 3 -6 5 0 .

15. S h e rm a n K M , R e id e n au e r J A , T h isd e D {1981) T h e ro le o f d is tu rb an c e in a n assem blage o f m arin e free-liv ing n em ato d es. A m erican Zoo log ist 21: 1035—1035.

16. T h is tle D {1998) H a rp a c tico id c o p ep o d diversity a t tw o physically rew o rk ed sites in th e d eep sea. D eep -S ea R e se a rc h P a r t II: T o p ic a l S tudies in O c e a n o g ra p h y 45: 13 -2 4 .

17. T h is tle D , Levin L A {1998) T h e effect o f ex p e rim en ta lly in creased n e a r-b o tto m flow o n m e ta z o an m eio fau n a a t a deep -sea site, w ith c o m p a riso n d a ta o n m acro fau n a . D eep -S ea R e se a rc h P a r t I: O c e a n o g ra p h ic R e se a rc h P a p e rs 45:6 2 5 -6 3 8 .

18. T h is tle D , Levin LA , G o o d a y A J, P fan n k u ch e O , L a m b sh e a d P JD {1999) P hysical rew o rk in g b y n e a r-b o tto m flow a lte rs th e m e ta z o an m eio fau n a o f




F ieb erlin g G u y o t {northeast Pacific). D e e p -S e a R e se a rc h P a r t I: O c e a n o g rap h ic R e se a rc h P a p e rs 46: 2 0 4 1 -2 0 5 2 .

19. C o n n e ll J H (1978) D iversity in tro p ica l ra in forests a n d co ra l reefs - h igh d iversity o f trees a n d corals is m a in ta in e d only in a n o n -eq u ilib riu m state. Science 199: 1 3 0 2 -1 3 1 0 .

20. H u s to n M (1979) G e n e ra l H yp o th esis o f species diversity . A m e ric a n N a tu ra lis t 113: 8 1 -1 0 1 .

21. L ev in L A , E tte r R J, R e x M A , G o o d a y A J, S m ith G R , e t al. (2001) E n v iro n m e n ta l influences o n reg io n a l deep -sea species d iversity . A n n u a l R eview o f E co logy a n d S ystem atics 32: 51—93.

22. Jo ris se n FJ, d e S tig ter H C , W id m a rk J G V (1995) A c o n c ep tu a l m o d el ex p la in in g b e n th ic fo ram in ife ra! m ic ro h ab ita ts . M a rin e M ic ro p a leo n to lo g y 26: 3—15.

23. S o e tae rt K , M u th u m b i A , H e ip C (2002) Size a n d sh ap e o f o cean m arg in n e m a to d e s: m o rp h o lo g ic a l d iv ers ity a n d d e p th - r e la te d p a tte rn s . M a rin e Eco logy-P rogress Series 242: 179 -1 7 9 3 .

24. T y le r P , A m aro T , A rzo la R , G u n h a M R , d e S d g ter H , e t al. (2009) E u ro p e ’s G ra n d C a n y o n N a z a re S u b m a rin e C an y o n . O c e a n o g ra p h y 22: 46—57.

25. R e id G S , H a m ilto n D (1990) A R eco n n a issan ce Survey o f the W h it ta rd Sea F an , S o u th w este rn A p p ro ach es, British-Isles. M a rin e G eo logy 92: 69—86.

26. W h ite M (2008) H E R M E S D eliv erab le 118: H y d ro g ra p h y a t th e can y o n s o f the N E A d a n tic /I r ish M a rg in - a synthesis (W P5 - C a n y o n systems).

27. D u in ev e ld G , L avaleye M , B erghu is E , de W ild e P (2001) A ctiv ity a n d co m p o sitio n o f th e b e n th ic fau n a in th e W h it ta rd C a n y o n a n d th e a d ja c e n t co n tin e n ta l slope {NE A tlantic). O ce a n o lo g ic a A c ta 24: 69—83.

28. D u ro s P , F o n ta n ie r C , M e tzg e r E , P u sced d u A , C e sb ro n F , e t al. {2011) Live {stained) b e n th ic fo ram in ife ra in th e W h it ta rd C an y o n , C eltic m arg in {NE A tlantic). D eep -S ea R esea rch P a r t I: O c e a n o g rap h ic R e se a rc h P ap e rs 58: 1 2 8 -1 4 6 .

29. Ingels J , V a n R oo ij D , V a n re u se l A {2006) H E R M E S R V Belgica 2 0 0 6 /1 3 Biology cru ise r e p o r t {23—29 J u n e 2006): G o llu m C h a n n e ls a n d W h it ta rd C an y o n .

30. K e n y o n N H {1987) M ass-w asting fea tu res o n th e co n tin e n ta l slope o f n o rth w es t E u ro p e . M a rin e G eology 74: 5 7 -7 7 .

31. K e n y o n N H , B elderson R H , S tride A H {1978) C h an n e ls , can y o n s a n d slum p folds o n co n tin e n ta l slope b e tw een so u thw est I re lan d a n d Spain . O cean o lo g icaA c ta 1: 369—380.

32. V a n R oo ij D {2004) A n in te g ra te d study o f Q u a te rn a ry sed im en ta ry processes o n th e ea ste rn slope o f th e P o rc u p in e S eab ig h t, S W o f Ire lan d . G h en t: G h e n t U niversity . 327 p.

33. W h ite M {2006) H E R M E S D e liv e ra b le 44: P re lim in a ry r e p o r t o n the h y d ro g ra p h y o f th e P o rc u p in e S eab ig h t (W P5 - C a n y o n systems).

34. L ev in LA , G o o d ay A J {2003) T h e d eep A tlan tic O c e a n . In : T y le r PA , ed. E co sy stem s o f th e d eep o c e an s E cosystem s o f th e W o rld . A m s te rd a m , N e th e rlan d s: E lsevier, p p 111—178.

35. P in g ree R D , L e c a n n B {1989) C eltic a n d A rm o ric an slope a n d sh e lf residua l cu rren ts . P rogress in O c e a n o g ra p h y 23: 303—338.

36. T h ie l H {1978) B en thos in u pw elling reg ions. In : Boje R , T o m c za k M , eds. U p w ellin g Ecosystem s. Berlin: Springer, p p 124—138.

37. W rig h t SW , J e ffrey SW {1997) H ig h -re so lu tio n H P L G system for ch lo rophy lls a n d ca ro ten o id s o f m arin e p h y to p lan k to n . In : J e ffrey SW , M a n to u ra R F C , W rig h t SW , eds. P h y to p lan k to n p ig m en ts in o cean o g rap h y : gu idelines to m o d e rn m ethods. Paris: U n ite d N a tio n s E d u ca tio n a l, Scientific a n d C u ltu ra l o rg an iza tio n , p p 327—341.

38. V in cx M {1996) M e io fa u n a in m a rin e a n d fresh w a te r sed im ents. In : H a ll G S, ed. M e th o d s fo r th e ex a m in a tio n o f o rg an ism a l d iversity in sils a n d sed im en ts, C A B In te rn a tio n a l, U n iversity Press, C am b rid g e , p p 214—248.

39. H e ip G , V in cx M , V ra n k e n G {1985) T h e E co logy o f M a rin e N em ato d es . O c e a n o g ra p h y a n d M a rin e Biology 23: 399—489.

40. H igg ins R P , T h ie l H {1988) In tro d u c tio n to th e study o f M eio fau n a. L on d o n : S m ith so n ian In stitu tio n Press.

41. W arw ick R M , P la tt H M , S om erfie ld P {1998) F ree-liv ing m arin e n e m a to d e s p a r t I I I - M o n hyste rids. P ic to ria l key to w o rld g e n e ra a n d n o tes fo r th e iden tification o f B ritish species B arnes D K A , C r o th e r s J H , eds. L o n d o n : T h e L in n e an Society o f L o n d o n .

42. D e p re z T , S tey ae rt M , V an a v e rb ek e J , S p ey b ro eck J , R aes M , e t al. {2005) N eM ys. W o rld W id e W eb e lec tro n ic pu b lica tio n , nem y s.u g en t.b e. D e p a r tm e n t o f M a rin e Biology, G h e n t U niversity .

43. W ieser W {1953) B ezieh u n g en zw ischen M u n d h ö h le n g e s ta lt, E rn äh ru n g sw eise u n d V o rk o m m e n b e i fre ileb en d en m a rin e n N e m a to d e n . A rkiv F ö r Z oo log i 2: 4 3 9 M 8 4 .

44. A nd rassy I {1956) T h e d e te rm in a tio n o f vo lu m e a n d w e ig h t o f n e m ato d es. A cta Z o o lo g ica { H ungarian A cad em y o f Science) 2: 1—15.

45. A n d e rso n M {2005) P E R M A N O V A : P e rm u ta tio n a l m u ltiv aria te analysis o f v a rian ce. A uck land : D e p a r tm e n t o f Statistics.

46. A n d e rso n M J, G orley R N , C la rk e K R {2007) P E R M A N O V A + fo r P R IM E R : gu ide to softw are a n d statistical m eth o d s. P R IM E R -E , P lym outh .

47 . H il l M O {1973) D iv e rs ity a n d ev en n ess: a u n ify in g n o ta t io n a n d its consequences. Ecology 54: 42 7 —432.

48. H u r lb e r t SH {1971) T h e n o n c o n c ep t o f species diversity: a c ritiq u e a n d a lte rn a tiv e p a ra m e te rs . Ecology 52: 57 7 —586.

49. S an d e rs H L {1968) M a rin e b e n th ic diversity: a c o m p ara tiv e study. A m erican N a tu ra lis t 102: 2 4 3 -2 8 2 .

50. H e ip C , H e rm a n P, S o e tae rt K {1998) Ind ices o f D iversity a n d Evenness. O cean is 24: 6 1 -8 7 .

51. A n d e rso n M J, G o rley R N , C la rk e K R {2008) P E R M A N O V A + fo r P R IM E R : g u ide to softw are a n d sta tistical m ethods. P lym outh : P R IM E R -E L td . 214 p.

52. C la rk e K R , G orley R N {2006) P R IM E R v6: U se r m a n u a l/tu to r ia l . P ly m o u th , U K : P R IM E R -E . 192 p.

53. J o in t I, W ollas t R , C h o u L, B a tten S, E lskens M , e t al. {2001) Pelagic p ro d u c tio n a t the C e ltic S ea sh e lf b reak . D eep -S ea R e se a rc h P a r t II: T o p ic a l S tudies in O c e a n o g ra p h y 48: 3 0 4 9 -3 0 8 1 .

54 . B ille tt D S M , L a m p it t R S , R ic e A L , M a n to u r a R F C {1983) S e a so n a l se d im e n ta tio n o f p h y to p la n k to n to the deep -sea b en th o s. N a tu re 302: 52 0 —522.

55. T h ie l H {1983) M eio b en th o s a n d n a n o b e n th o s o f th e d eep sea. In : R o w e G T , ed. D eep -sea biology. N e w Y ork: W iley, p p 167—230.

56. V in cx M , B ett BJ, D in e t A , F e rre ro T , G o o d ay A J, e t al. {1994) M e io b e n th o s o f th e D eep N o rth e a s t A d an tic . A d v an ces in M a rin e Biology, V o l 30. L on d o n : A cad em ic P ress L td . p p 1—88.

57. Soltw edel T {2000) M e ta z o a n m eio b en th o s a lo n g co n tin e n ta l m arg ins: a review . P rogress in O c e a n o g ra p h y 46: 5 9 —84.

58. P fan n k u ch e O {1985) T h e D eep -S ea M eio fau n a o f th e P o rc u p in e S eab ig h t a n d A byssa l-P la in {Ne A d an tic ) - P o p u la d o n -S tru c tu re , D is tr ib u tio n , S ta n d in g Stocks. O ce a n o lo g ic a A c ta 8: 343—353.

59. P fan n k u ch e O , S o m m er S, K a h le r A {2000) C o u p lin g b e tw een p h y to d e tr itu s d ep o sitio n a n d th e sm all-sized b e n th ic b io ta in th e d eep A ra b ia n Sea: analyses o f b io g en ic sed im en t co m p o u n d s. D eep -S ea R e se a rc h P a r t II: T o p ic a l S tud ies in O c e a n o g ra p h y 47: 2 8 0 5 -2 8 3 3 .

60. M okievsky V O , U d a lo v AA , Azovskii A I {2007) Q u a n d ta d v e d istr ib u d o n o f m eio b en th o s in d eep -w a te r zones o f th e W o rld O c e a n . O c ean o lo g y 47: 7 9 7 -8 1 3 .

61. V a n re u se l A , V in cx M , S ch ram D , V an g an sb ek e D {1995) O n th e V ertica l- D is tr ib u tio n o f th e M e ta z o a n M eio fau n a in S h e lf B reak a n d U p p e r Slope H a b ita ts o f th e N E A d an tic . In te rn a tio n a le R ev u e D e r G esam ten H ydro b io lo g ie80: 3 1 3 -3 2 6 .

62. V an a v e rb ek e J , S o e tae r t K , H e ip C , V a n re u se l A {1997) T h e m e ta z o an m eio b en th o s a lo n g th e co n tin e n ta l slope o f th e G o b a n S p u r {NE A dantic).

J o u rn a l o f Sea R e se a rc h 38: 93—107.63. O liv e ira A , S an tos A I, R o d rig u es A , V ito rin o J {2007) S ed im en tary partic le

d istr ib u tio n a n d d ynam ics o n the N a z a re , can y o n system a n d a d ja c e n t shelf {Portugal). M a rin e G eo logy 246: 1 0 5 -1 2 2 .

64. S c h m id t S, de S tig ter H C , v a n W e erin g T C E {2001) E n h a n c e d sh o rt- te rm sed im en t dep o sitio n w ith in th e N a z a re C an y o n , N o rth -E a s t A d an tic . M arin e G eology 173: 5 5 -6 7 .

65. D e L eo FG , S m ith G R , R o w d e n AA , B ow den D A , C la rk M R {2010) S u b m a rin e canyons: ho tsp o ts o f b e n th ic b io m ass a n d p ro d u ctiv ity in th e d eep sea. P ro ceed in g s o f th e R o y a l S ociety B: B io logical Sciences.

66. Ingels J , B illett D S M , W o lff G , K iriakou lak is K , V a n re u se l A {In Press) S tru c tu ra l a n d fu n cd o n a l diversity o f N e m a to d a in re la d o n w ith e n v iro n m en ta l variab les in th e S e tú b a l a n d C ascais canyons, W e ste rn Ib e r ia n M arg in . D eep - Sea R e se a rc h P a r t II: T o p ic a l S tud ies in O c e a n o g rap h y .

67. Soltw edel T , M ilju d n a M , M okievsky V , T h isd e D , V o p e l K {2003) T h e m eio b en th o s o f the m olloy d eep {5 600 m ), F ra m S tra it, A rcd c O c e a n . V ie E t M ilieu-L ife a n d E n v iro n m e n t 53: 1—13.

68. G a g e J D , T y le r PA {1992) D eep -S ea Biology: a N a tu ra l H is to ry o f o rgan ism s a t th e D eep -S ea floor. C am b rid g e : C a m b rid g e U n iv ers ity Press.

69. D a n o v a ro R , D e ll’A n n o A , F a b ian o M , P u sced d u A, T selep ides A {2001) D eep- sea ecosystem response to c lim ate changes: th e ea ste rn M e d ite rra n e a n case s tudy. T re n d s in Ecology & E v o lu tio n 16: 5 0 5 —510.

70. S o e tae rt K , F ran co M , L a m p ad a r io u N , M u th u m b i A , S tey ae rt M , e t al. {2009) F ac to rs affecd n g n e m a to d e b io m ass , len g th a n d w id th f ro m th e shelf to th e d eep sea. M a rin e E cology-Progress Series 392: 123—132.

71. S o e tae rt K , H e ip C {1995) N e m a to d e A ssem blages o f D eep -S ea a n d S he lf B reak Sites in th e N o rth -A d a n tic a n d M e d ite rra n ea n -S e a . M a rin e E cology-Progress Series 125: 1 7 1 -1 8 3 .

72. M c L a re n P , Bowles D {1985) T h e effects o f sed im en t t ra n sp o rt o n grain-size d istribu tions. J o u rn a l o f S ed im en tary R e se a rc h 55: 4 5 7 M 7 0 .

73. A u sten M C , W arw ick R M , R y a n K P {1993) A s to m o n e m a -S o u th w a rd o ru m Sp- N o v , a G udess N e m a to d e D o m in a n t in a M e th a n e Seep A re a in th e N o rth -S ea . J o u rn a l o f th e M a rin e B io logical A ssociation o f the U n ited K in g d o m 73: 6 2 7 -6 3 4 .

74. G ie re O , W in d o ffer R , S o u th w a rd E C {1995) th e b a c te r ia l endosym biosis o f the gudess n e m a to d e , Astomonema southwardorum - u ltra s tru c tu ra l aspects. J o u rn a l o f th e M a rin e B iological A sso c iad o n o f th e U n ited K in g d o m 75: 153—164.

75. M u sa t N , G ie re O , G ieseke A , T h ie rm a n n F, A m a n n R , e t al. {2007) M o lecu lar a n d m o rp h o lo g ic a l c h a ra c te r iz a t io n o f th e asso c ia tio n b e tw e en b a c te r ia l en d o sy m b io n ts a n d the m arin e n e m a to d e A sto m o n e m a sp f ro m th e B aham as. E n v iro n m e n ta l M ic rob io logy 9: 1345—1353.

76. V id ak o v ic J , B o u ch e r G {1987) G u dess m a rin e n e m ato d es o f th e genus Astomonema O t t e t al., 1982. C ah ie rs D e B iologie M a rin e 28: 11 1 -1 2 0 .

77. H o p e W D {1977) G u d ess n e m a to d e s o f th e d e e p sea. M ik ro fau n a des M eeresb o d en s 61: 3 0 7 -3 0 8 .

78. K ito K , A ry u th a k a C {2006) N e w m o u th le ss n e m a to d e o f th e gen u s P a ras to m o n e m a K ito , 1989 {N em atoda : S ipho n o la im id ae) fro m a m an g ro v e fo re s t o n th e c o a s t o f T h a i la n d , a n d e re c t io n o f th e n e w su b fam ily A sto m o n e m a tin a e w ith in th e S ip h o n o la im id ae. Z o o tax a , p p 39M 9.

79. O t t J , B rig h t M , B u lgheresi S {2004) S ym bioses b e tw een m arin e n e m ato d es a n d su lfu r-ox id iz ing c h e m o a u to tro p h ic b a c te ria . Sym biosis 36: 103—126.




80. L ee H J , G erd es D , V a n h o v e S, V in c x M (2001) M eio fau n a resp o n se to iceberg d is tu rb an c e o n th e A n ta rc tic co n tin e n ta l shelf a t K a p p N o rv eg ia (W eddell Sea). P o la r B iology 24: 9 2 6 -9 3 3 .

81. H e ip C , H uys R , V in cx M , V an reu se l A, Sm ol N , e t al. (1990) C o m p o sitio n , D is tr ib u tio n , B iom ass a n d P ro d u c tio n o f N o rth -S e a M eio fau n a. N e th e rla n d s

J o u rn a l o f Sea R esea rch 26: 3 3 3 -3 4 2 .82. B ongers T , A lk em ad e R , Y eates G W (1991) In te rp re ta tio n o f d istu rb an ce -

in d u ce d m a tu r ity dec rease in m a rin e n e m a to d e assem blages b y m ean s o f the m atu r ity index . M a rin e Eco logy-P rogress Series 76: 135—142.

83. G allucc i F , M o en s T , V a n re u se l A , Fonseca G (2008) A ctive co lon isa tion o f d is tu rb ed sed im en ts b y deep -sea nem ato d es: ev idence fo r th e p a tc h m osaic m odel. M a rin e E co logy-P rogress Series 367: 173—183.

84. H o s te E (2006) T e m p o ra l a n d sp a tia l variab ility in deep -sea m eio b en th ic c o m m u n ities fro m th e A rc tic M a rg in a l le e Z one . G h en t: G h e n t U niversity .

85. R a e s M , V a n re u se l A (2006) M ic ro h a b ita t type d e te rm in es th e c o m p o sitio n o f n e m a to d e co m m u n ities a ssoc ia ted w ith sed im en t-c lo g g ed co ld -w a te r co ra l fram ew o rk in th e P o rc u p in e S eab ig h t {NE A tlantic). D eep -S ea R e se a rc h P a r t I: O c e a n o g ra p h ic R e se a rc h P a p e rs 53: 1880—1894.

86. V a n re u se l A , F o n seca G , D a n o v a ro R , d a Silva M C , Esteves A M , e t al. (2010) T h e im p o rta n c e o f deep -sea m a c ro h a b ita t h e te ro g en eity for g lobal n e m a to d e d iversity . M a rin e E co logy 31: 6—20.

87. H u s to n M (1994) B iological diversity: th e coexistence o f species o n c h an g in g lan d scap es C a m b rid g e U n iv Pr.



meiofauna in the gollum channels and the whittard canyon ... · gollum channel system were sampled...

Documents