The Structuring Role of Microhabitat Type in Coral Degradation Zones: a Case Study with Marine Nematodes from Kenya and Zanzibar

Coral Reefs (2007) 26:113–126 DOI 10.1007/s00338-006-0184-8

REPORT

The structuring role of microhabitat type in coral degradation zones: a case study with marine nematodes from Kenya and Zanzibar

M. Raes · M. De Troch · S. G. M. Ndaro · A. Muthumbi · K. Guilini · A. Vanreusel

Received: 31 July 2006 / Accepted: 20 November 2006 / Published online: 7 February 2007 © Springer-Verlag 2007

Abstract Nematode genus assemblages were identi- three-dimensional structure of coral fragments was Wed from four locations in coral degradation zones found. DiVerences between nematode assemblages in (CDZs) along the African east coast: Watamu and the coralline sediment and on coral fragments were Tiwi Beach (Kenya) and Matemwe and Makunduchi attributed to the exposed nature of the latter habitat, (Zanzibar). Three microhabitat types were distin- its large surface area and its microbial or algal cover. guished: coralline sediment, coral gravel and coral frag- DiVerences in available food sources were reXected in ments. Nematode community composition was nematode trophic composition. comparable to that of other studies dealing with the same habitat. The presence of a common genus pool in Keywords Coral degradation zones · Nematodes · CDZs was reXected in the considerable similarities Microhabitats · Spatial turnover · Indian Ocean between samples. The addition of coral fragments as a habitat for nematodes resulted in an increased importance of taxa typical for coarse sediments and large Introduction substrata. Local and regional turnover were of the same order of magnitude. The structuring eVect of Numerous studies have investigated the major factors microhabitat type clearly overrode the eVect on a local that structure nematode community composition in and regional scale. DiVerences in sediment characteris- coral reef associated sediments (Alongi 1986; Boucher tics were more important in structuring the nematode and Gourbault 1990; Gourbault and Renaud-Mornant assemblages than diVerences between the coralline 1990; Tietjen 1991; Ólafsson et al. 1995; Boucher 1997; sediment and coral fragments. No eVect related to the Ndaro and Ólafsson 1999; Netto et al. 1999; Kotta and Boucher 2001; de Jesús-Navarrete 2003). These eVorts have produced a list of variables that are, at least Communicated by Ecology Editor P.J. Mumby. potentially, determinant: (1) mean sediment grain size, (2) sediment clay-silt content, (3) sediment sorting, (4) M. Raes (&) · M. De Troch · K. Guilini · A. Vanreusel Marine Biology Section, Biology Department, sediment oxygen content, (5) position of redox poten- Ghent University, Sterre complex - S8, tial discontinuity (RPD) layer in the sediment, (6) Krijgslaan 281, 9000 Gent, Belgium organic content of the sediment, (7) extent of bioturba- e-mail: [email protected] tion by macrobenthos, (8) macroXoral biomass, (9) S. G. M. Ndaro water depth, (10) longitude, (11) latitude and (12) sam- Department of Aquatic Environment and Conservation, pling scale. The vast majority of these factors is related University of Dar Es Salaam, P.O. Box 35064, to sediment characteristics. Dar Es Salaam, Tanzania Coral reef associated sediments can be found in all X A. Muthumbi functional zones of the reef, e.g. the reef crest, reef at, Department of Zoology, University of Nairobi, outer reef, reef pools, reef lagoon, etc. Most of the P.O. Box 30197, Nairobi, Kenya studies mentioned above have however focused on 123 114 Coral Reefs (2007) 26:113–126 lagoons, which separate the reef platform from the structure compares to the changes due to turnover on a adjacent land mass. In this zone, dead coral material local and regional scale. The key questions here are: coming from the main reef or from scattered coral (1) do CDZs harbour a typical nematode community? thickets in the lagoon itself (e.g. from patch reefs), is (2) how strong is turnover in taxonomic composition progressively degraded into smaller pieces (coral operating at local and regional scales? And (3) is gravel) until it becomes coralline sediment. Because of microhabitat structure an additional source for varia- the dynamic character of this functional zone, the term tion in nematode community composition? Coral Degradation Zone (CDZ) will be used through- out the text when referring to the lagoon between the reef and the coast. As a result of this gradual process of Materials and methods degradation, the CDZ sediment will obviously be mixed with coral fragments of diVerent sizes and Sampling sites, procedure and microhabitats shapes. The presence of coral degradation products (dead coral fragments, coral gravel) on and in the sedi- Meiofauna samples were collected in the CDZ of the ment may have a profound eVect on nematode assem- fringing reefs stretching along the south coast of Kenya blage structure on a local scale. So far, these coral and along the east coast of Zanzibar Island (Unguja, degradation products have not yet been considered as Tanzania) (Fig. 1). In Kenya, samples were taken on an ecologically valuable habitat in tropical CDZs. two locations: Watamu, the northernmost location The present study contributes to the knowledge of (03°23ЈS, 40°00ЈE; 27th February 2002), and Tiwi nematode assemblages in CDZs of the Indian Ocean, a Beach, more to the south (4°14ЈS, 39°36ЈE; 22nd–23rd subject investigated in only a limited number of studies February 2002). In Zanzibar, samples were taken in [Thomassin et al. 1976 (Madagascar); Ólafsson et al. Matemwe, located in the north of the island (5°52ЈS, 1995 (Zanzibar); Ndaro and Ólafsson 1999 (Zanzibar)]. 39°21ЈE; 24th August 2004 and 31st August 2004), and Here, three distinct microhabitat types (i.e. coralline Makunduchi (6°28ЈS, 39°32ЈE and 6°25ЈS, 39°34ЈE; sediment, coral gravel and coral fragments) are distin- 15th August 2004 and 25th August 2004), located in the guished. This unique approach makes it possible to south of the island. Distance between both Kenyan investigate how changes in community composition locations is 104 km, between Tiwi Beach and Matemwe resulting from the structuring role of microhabitat 183.5 km and between both Zanzibar locations 70 km.

Fig. 1 Map of the study area; 38˚ 0' E 39˚ 0' E 40˚ 0' E 41˚ 0' E 42˚ 0' E location of sampling sites is indicated. The northernmost island is Pemba, the southern- Watamu N most island is Unguja (Zanzibar Island) m -500

4˚ 0' S 0m Tiwi Beach -100

200 5˚ 0' S -

m 0m 0 0 0 0 20 - -3

-100 m Matemwe 6˚ 0' S

Makunduchi km

0 50 7˚ 0' S 123 Coral Reefs (2007) 26:113–126 115

On each location, coralline sediment (six samples in sample 200 nematodes (or all nematodes when less both Kenya locations; two samples in both Zanzibar than 200 individuals were present in the examined locations) and coral fragments (two samples in Wat- sample) were randomly picked out. They were subse- amu, three samples in Tiwi Beach and six samples in quently mounted onto slides using the formalin-etha- both Zanzibar locations) were collected. Coral gravel nol-glycerol technique of Seinhorst (1959) and Vincx was collected in Watamu, Matemwe and Makunduchi (1996), and identiWed up to the genus level, using (two samples in all three locations), but not in Tiwi Lorenzen (1994), Warwick et al. (1998), the Desmo- Beach due to the absence of this microhabitat in the doridae key in NeMys (Deprez et al. 2004), and origi- sampled area. Coral gravel can easily be distinguished nal descriptions. The trophic composition of the from coralline sediment because small pieces of coral nematode community was analysed according to the can still be recognised in this microhabitat, whereas classiWcation of Wieser (1953). this is no longer true for the sediment (Fig. 2). The dead coral fragments were either compact or branched, Statistical analyses unaVected or eroded. In Zanzibar, diVerent coral morphotypes were collected, which were identiWed as Fun- The PRIMER5 software (Plymouth Marine Labora- gia, Goniastrea, Pocillopora, Tubipora and Porites/ tory; Clarke and Gorley 2001) was used to calculate Stylophora. Porites and Sylophora were considered as Bray-Curtis (dis)similarities between all samples. Sam- the same morphotype in this study, based on their ples were grouped together in three ways, concor- robustness and branching morphology. All sampling dantly with three spatial scales: (1) the diVerent regions was carried out under water. Sediment samples were (Kenya/Zanzibar: regional scale), (2) the diVerent sam- taken with a Perspex sediment core (10 cm²), coral pling locations (Watamu/Tiwi Beach/Makunduchi/ gravel was gently scooped out with a spoon or sedi- Matemwe: local scale) and (3) the diVerent habitats ment core and coral fragments were taken out manu- (coralline sediment/coral gravel/coral fragments: ally. Coral gravel and coral fragments were put directly microhabitat scale). The obtained similarity matrix was into Wrm plastic bags or buckets. used to produce non-metric multidimensional scaling two-dimensional plots (MDS). The stress value gives a Laboratory analyses measure for goodness-of-Wt of the MDS ordination: a low-stress value (<0.2) indicates a good ordination with Meiofauna extraction from coralline sediment and no real prospect for a misleading interpretation coral gravel was done by decantation with Wltered (Clarke 1993). One-way, two-way crossed and two-way seawater over a 1 mm and a 32 m sieve and subse- nested Analysis of Similarities (ANOSIM) were car- quent centrifugation. Coral fragments were rinsed oV ried out to test for signiWcant diVerences in the nema- thoroughly, also with Wltered seawater, over the same tode community structure between diVerent groups sieves. The material collected on the 32 m sieve was that were delimited beforehand. Similarity of Percent- then subjected to density gradient centrifugation, ages (SIMPER) was used to investigate which genera using Ludox (a colloidal silica polymer; speciWc grav- were responsible for these diVerences. Due to diVer- ity 1.18) as a Xotation medium (Heip et al. 1985; ences in sample size, relative data were used per sam- Vincx 1996). All material was Wxed with 4% buVered ple and these data were Log (x + 1)-transformed prior formalin and stained with Rose Bengal. From each to the analysis.

Fig. 2 The three microhabitats: a coralline sediment; b coral gravel; c coral fragment (Porites/Stylophora is given as an example). Scale bars 5cm 123 116 Coral Reefs (2007) 26:113–126

The Pcord4 software (McCune and MeVord 1999) (R = 0.484; p = 0.001) and pairwise tests revealed sig- was applied to perform an Indicator Species Analysis niWcant diVerences between sediment and coral sam- (ISA) on the dominant genera (i.e. with a relative ples (R = 0.648; p = 0.001) and between coral gravel abundance >0.5%). Calculated indicator values were and coral samples (R = 0.406; p = 0.01). SigniWcant tested for statistical signiWcance using a Monte Carlo diVerences were also found between communities from test (Dufrêne and Legendre 1997). The same software diVerent microhabitats irrespective of potential diVer- was also used for TWINSPAN analysis, which is a divi- ences between locations (R = 0.552; p = 0.001). In this sive classiWcation method (Hill 1979; Gauch Jr and analysis, pairwise tests again showed signiWcant diVer- Whittaker 1981). Calculated cut levels were 0.0; 0.5; ences between sediment and coral samples (R = 0.713; 1.2; 3.0; 6.0; 53.0. For both ISA and TWINSPAN, rela- p = 0.001) and between coral gravel and coral samples tive data were used without transformation. (R = 0.479; p = 0.028). There were no signiWcant diVer- Parametric (one-way ANOVA) and non-parametric ences between sediment and gravel samples. The (Kruskal–Wallis ANOVA by ranks) analysis of vari- gravel samples are also found scattered between sam- ance was performed using the STATISTICA6 soft- ples from both other microhabitats in the MDS plot ware. Bartlett’s and Cochran’s test were used to verify (Fig. 3a). Therefore, the gravel samples were omitted the homogeneity of variances prior to the analysis. in further analyses. On the MDS plot, the coral sam- Turnover on a regional, local and microhabitat scale ples were clearly clustered more closely together than is visualised in a ternary plot (KoleV et al. 2003). The the sediment samples. Furthermore, it is clear from values of a’, b’ and c’ (i.e. the percentage of shared spe- Fig. 3a that the two sediment samples in the top left cies a, the percentage of species exclusively present in corner of the plot, which originate from Matemwe the neighbouring sample b and the percentage of spe- (Zanzibar), have a community composition diVerent cies exclusively present in the focal sample c) are plot- from that of the other sediment samples. ted against a background of sim-values (Lennon et al. No clear separation of groups of samples from the 2001). This technique has been used before to unravel four diVerent sampling locations could be observed nematode community turnover (G. Fonseca, personal from the plot (not shown). In agreement with this communication). observation, two-way nested ANOSIM showed no clear separation of communities from diVerent locations within each region (R = 0.142), although the p- Results level was signiWcant (p =0.018). The eVect of locations on community structure, irrespective of the potential A total of 7,087 nematodes belonging to 149 diVerent eVect of microhabitats, was calculated per region to genera and 35 diVerent families were included in the rule out the regional eVect, and will be discussed analysis. The coral gravel, coral fragments and coral- below. line sediment yielded 89, 108 and 127 genera, respec- To deWne the most important structuring factor for tively. A list of all encountered genera is provided in the nematode communities, the eVects of (1) regional the Appendix. forces within both microhabitats, (2) local forces within both regions, (3) local forces within both microhabitats Are there characteristic nematode communities for the per region, (4) microhabitat structure within both diVerent regions, locations and microhabitats? What is regions and (5) microhabitat structure within all loca- the structuring role of microhabitat structure? tions were investigated with one-way (1, 3, 5) and two- way crossed (2, 4) ANOSIM (Table 1; Fig. 3b–e). Most The MDS biplot of all samples (Fig. 3a) shows both a eVects were found to be signiWcant. Especially the separation of the samples from the two regions (dashed eVect of microhabitat structure yielded well-separated line) and a separation of the coral samples from the groups. However, R-values indicated the occurrence of sediment samples (dotted line), although the separa- clearly diVerent groups on both microhabitat, local and tion between regions is not as clear-cut as the separa- regional scales. There was no signiWcant separation tion between the two microhabitats. Two-way crossed between the coral samples from Watamu and those ANOSIM calculated signiWcant assemblage diVer- from Tiwi Beach (R = ¡0.3), which is also obvious ences between samples from Kenya and Zanzibar, irre- from Fig. 3b. A TWINSPAN dendrogram (Fig. 4) spective of the microhabitat type (R =0.595; revealed that at a Wrst level the two sediment samples p = 0.001). DiVerences between the nematode commu- from Matemwe (TWIN group 1) branch oV from the nities in diVerent microhabitats, irrespective of the other samples. Richtersia was speciWed as a TWIN- region where they occur, were also signiWcant SPAN indicator genus for group 1. At a second level, 123 Coral Reefs (2007) 26:113–126 117

(a) Stress: 0,17

Coral Kenya Gravel Kenya Sediment Kenya Coral Zanzibar Gravel Zanzibar Sediment Zanzibar

(b)Stress: 0,16 (c) Stress: 0,14

Matemwe Makunduchi Tiwi Beach Watamu Matemwe Makunduchi Tiwi Beach Watamu

(d) Stress: 0,17 (e) Stress: 0,07

Coral fragments Sediment Coral fragments Sediment

Fig. 3 Multidimensional scaling two-dimensional ordination dashed lines separate samples from diVerent regions (Kenya– plots. Stress values are indicated: a all samples; b coral samples; c Zanzibar); the dotted line separates sediment samples from coral sediment samples; d Kenya samples; e Zanzibar samples. The samples the other sediment samples (TWIN group 3) are sepa- are, however, clearly of the same order of magnitude, rated from the coral samples (TWIN group 2). Atro- with a sim-value around 0.2. chromadora, Chromadora and Chromadorina were An MDS plot (not shown) and subsequent two-way speciWed as TWINSPAN indicator genera for group 2, crossed ANOSIM analyses demonstrated the absence whereas Neochromadora was speciWed as a TWIN- of an eVect of the three-dimensional build-up of diVer- SPAN indicator genus for group 3. ent coral morphotypes on the nematode community The ternary graph (Fig. 5) shows that turnover ( sim) composition. between coral and sediment within the same location (grey squares) is generally higher than the turnover How unique and speciWc are the nematode communi- between locations within the same region (black cir- ties in the diVerent regions and microhabitats? cles) and between locations from diVerent regions (white triangles). Due to a lower value for the turnover Even though the communities inhabiting coral frag- between the microhabitats in Watamu, these diVer- ments and coralline sediment in both regions are sig- ences were not signiWcant. Local and regional turnover niWcantly diVerent from each other, they do not make 123 118 Coral Reefs (2007) 26:113–126

Table 1 R-values and signiWcance levels of the eVects on diVerent spatial scales, as calculated by one-way and two-way crossed ANO- SIM (see text for details) Selected samples EVect

Regions Locations Microhabitats

Regions Kenya R = 0.316; p = 0.007 R = 0.482; p = 0.009 Zanzibar R = 0.386; p = 0.002 R = 0.964; p = 0.001 Locations Watamu R = 0.188; p = 0.357 Tiwi Beach R = 0.747; p = 0.012 Matemwe R = 1.000; p = 0.036 Makunduchi R = 0.927; p = 0.036 Microhabitats Coral fragments R = 0.461; p = 0.004 Coralline sediment R = 0.729; p = 0.001 Microhabitats per region Kenya coral fragments R=¡0.333; p = 0.900 Kenya coralline sediment R = 0.391; p = 0.002 Zanzibar coral fragments R = 0.354; p = 0.004 Zanzibar coralline sediment R = 1.000; p = 0.333 Bold italics stands for well-separated groups (R > 0.75), italics underline for overlapping but clearly diVerent groups (0.75 ¸ R >0.5), and italics double underline for the absence of clear groups (R · 0.5)

Fig. 4 TWINSPAN Richtersia (1) dendrogram based on relative abundances of nematode Atrochromadora (4) Chromadora (4) genera in each sample. Only Chromadorina (4) Neochromadora (3) sediment and coral fragment (1) samples are considered. The star Wgure indicates a mis- matched sediment sample in (2) (3) the coral samples TWIN group. TWINSPAN indicator genera for each TWIN group are indicated with their signs

Coral samples Watamu, Tiwi Beach Matemwe and Makunduchi sediment samples sediment samples

Fig. 5 Ternary plot a' representing species turnover 0 100 β between coral and sediment sim within the same location 1 (turnover between 0,8 microhabitats), between 0,6 25 75 locations within the same 0,4 region (local turnover) and 0,2 between locations from diVerent regions (regional turnover). Shading visualizes 50 50 Regional turnover the values of sim Local turnover Turnover between microhabitats

75 25

100 0 0255075100 c' b' up distinct, clear-cut groups (Fig. 3a). Moreover, each region. The lowest number of shared genera, as derived group has at least half of its genera in common with from the surface of the radar chart, was found for the other groups (Fig. 6). This eVect is independent of the sediment in Zanzibar, which was also characterised by

123 Coral Reefs (2007) 26:113–126 119

(a) Zanzibar sediment (72) (b) Zanzibar coral (85) Zcoral (85) Zsediment (72) 75 75

50 52 50 52

25 25

0 0

Ksediment (115)58 49 K coral (87) Ksediment (115) 70 62 K coral (87)

25 25

0 0

Zcoral (85) K coral (87) Zcoral (85) Ksediment (115) 70 67 62 67

Fig. 6 Radar charts depicting the number of nematode genera crohabitats in both regions (Z Zanzibar, K Kenya) on the other shared between a certain microhabitat type in one of the regions hand. The total number of genera in each microhabitat is indi- (indicated above each graph) on the one hand and the other mi- cated between brackets the lowest number of genera (72) (Fig. 6a). The highest The speciWcity of the nematode communities in the number of shared genera was found for the sediment in same four groups was evaluated in terms of uniqueness Kenya (Fig. 6c), which was characterised by the highest of genera, i.e. whether and how many genera are number of genera (115). Average dissimilarity (SIM- restricted to a certain microhabitat or region (Fig. 7). PER analysis) between the four groups varied between Although more stations were sampled for both Kenya 53.2% (Kenya coral-Zanzibar coral) and 82.8% (Kenya sediment (12) and Zanzibar coral (12) (Fig. 7a), this coral-Zanzibar sediment). Coral samples from both was only reXected in a higher number of unique genera regions are relatively comparable in terms of associ- in the sediment from Kenya (Fig. 7b). The detailed dis- ated nematode communities, whereas sediment sam- tribution shows that most of the unique genera are ples from both regions are much more dissimilar from restricted to 1 or, to a lesser extent, 2–3 samples within each other (average dissimilarity: 75.4%). On the other a group (Fig. 7c). Moreover, the number of unique hand, the average similarity of samples within each genera corresponds well with the distribution of single- group is relatively low: between 38.1% (Zanzibar sedi- tons (i.e. unique genera found in only one sample). ment) and 52.9% (Zanzibar coral). Overall, average There were no genera unique for Kenya and only three similarity of coral samples (50.2%) was higher than for genera unique for Zanzibar. Within Kenya, 11 and 24 sediment samples (33.9%). genera were unique for coral fragments and sediment,

30 30 (a)(b) 20 (c)

25 ra 25

en le 15

g p 20 20

les

n 15 15 am 10

r genera restricted

10 10 to

b u 5

5 u 5

n number of 0 0 0 12345 microhabitat microhabitat number of samples x Kenya coral Kenya sediment Zanzibar coral Zanzibar sediment

Fig. 7 Stacked columns depicting the number of nematode genera unique for a certain microhabitat type in one of the regions. a Com- parison of sampling intensity; b Number of unique genera and c Detailed distribution of unique genera

123 120 Coral Reefs (2007) 26:113–126 respectively, whilst in Zanzibar 5 and 7 genera were ingly, the Wve most abundant families for both micro- restricted to coral fragments and sediment, respec- habitats are the same: Chromadoridae, Cyatholaimidae, tively. Desmodoridae, Epsilonematidae and Xyalidae. Chro- madoridae is the dominant family on corals, with Des- Characterisation of the nematode communities modoridae the second most abundant. The opposite is the case in the coralline sediment. An overview of the most abundant genera characteris- All genera exhibiting signiWcant indicator values are tic for coral and sediment samples from both regions is listed in Table 2. The highest indicator values and high- given in Fig. 8a, b, respectively. For the coral samples, est signiWcance levels for the coral fragments are found all genera with a relative abundance >2%, calculated in representatives of the family Chromadoridae (Atro- over all coral samples, and occurring in at least 75% of chromadora, Chromadorina, Chromadora) and for the the coral samples were selected. For the sediment sam- coralline sediment they are found in representatives of ples, all genera with a relative abundance of >2%, cal- both Desmodoridae (Eubostrichus, Metachromadora, culated over all sediment samples, and occurring in at Bolbonema) and Chromadoridae (Neochromadora). least 50% of the sediment samples were selected. This For coral fragments, the same results were found diVerence in procedure is due to the low abundances of within each region. Nine of ten genera featured in the dominant genera in the sediment samples. In this Fig. 8a are also indicator genera for coral fragments. way 46.5, 50.5, 60.5 and 66.5% of the Zanzibar sedi- This correspondence is not clear for the sediment sam- ment, Kenya sediment, Zanzibar coral and Kenya coral ples. The four indicator genera for coralline sediment communities is shown in the stack bars, respectively. in Kenya belong to four diVerent families, but none The three most abundant genera in the coral samples belongs to the Desmodoridae. belong to the families Chromadoridae (Atrochromadora, The list of genera that explain most of the average Chromadora) and Epsilonematidae (Epsilonema), similarity within each of these four groups, as pointed whereas those in the sediment samples are representa- out by a SIMPER analysis (not shown), corresponds tives of the families Desmodoridae (Chromaspirina, well with the list of indicator genera (Table 2) and the Spirinia) and Chromadoridae (Neochromadora). Strik- genera provided in Fig. 8. Only for the overall coralline

(a) (b) 70

60 60

50 Paradraconema 50 Bolbonema Theristus Euchromadora Spilophorella Dracognomus Metachromadora 40 Syringolaimus %) 40 Chromadora Acanthonchus Epsilonema Chromadorina Perepsilonema 30 Euchromadora 30 Metadesmolaimus Epsilonema

Relative abundance (%) Chromadora Eubostrichus 20 20 Theristus Atrochromadora

Relative abundance ( Neochromadora Spirinia 10 10 Chromaspirina

0 0 Kenya coral Zanzibar coral Kenya sediment Zanzibar sediment Microhabitats Microhabitats

Fig. 8 Dominant nematode genera in coral and sediment sam- a relative abundance >2% of the total sediment community and ples from both regions. a Coral fragments: genera with a relative occurring in at least 50% of the sediment samples. Indicator gen- abundance >2% of the total coral community and occurring in at era for either coral fragments or coralline sediment, as speciWed least 75% of the coral samples; b Sediment samples: genera with by an indicator species analysis, are printed in bold type

123 Coral Reefs (2007) 26:113–126 121

Table 2 Indicator genera for each separate microhabitat and for abundance of epistratum feeders (Wieser group 2a) was each microhabitat within a region, as speciWed by an indicator signiWcantly higher on coral fragments (p = 0.005). species analysis Coral fragments Coralline sediment

Indicator genus Indicator Indicator genus Indicator Discussion value value Do coral degradation zones harbour a typical Atrochromadora 89.2*** Eubostrichus 69.5** nematode community? Chromadorina 82.2*** Neochromadora 69.0** Chromadora 79.8*** Metachromadora 56.7** Paradraconema 72.6*** Bolbonema 53.3** Desmodoridae, Chromadoridae, Xyalidae and Cyath- Daptonema 59.5*** Perepsilonema 56.1* olaimidae dominated both the sediment and coral frag- Euchromadora 74.7** Chromadorita 56.0* ments in the study area. This is consistent with most Acanthonchus 72.8** Ptycholaimellus 52.0* Epsilonema 72.5** Dracognomus 46.0* studies in tropical, reef-associated sediments (Grelet Spilophorella 69.6** Paracomesoma 43.4* 1984; Renaud-Mornant and Gourbault 1984; Gourba- Calomicrolaimus 54.8** ult and Renaud-Mornant 1990; Boucher and Gourba- Paracanthonchus 53.7** ult 1990; Tietjen 1991; Boucher 1997; Ndaro and Syringolaimus 61.6* Halalaimus 48.3* Ólafsson 1999; Kotta and Boucher 2001). This general trend is also reXected in the genus composition of these Kenya coral Kenya sediment sediments. A comparison with complete genus lists in Chromadora 68.6*** Theristus 49.6** similar environments (Alongi 1986; Gourbault and Neochromadora 69.6* Renaud-Mornant 1990; de Jesús-Navarrete 2003) has Perepsilonema 59.0* shown that, respectively, 90, 80 and 77% of the genera Dracognomus 56.0* encountered in Australia, French Polynesia and the Zanzibar coral Zanzibar sediment Caribbean were also found along the East African coast. The total number of genera in these studies was Chromadorina 64.0*** Chromaspirina 71.1*** Atrochromadora 58.4** Paracomesoma 94.3** rather low (35, 43 and 56). Nevertheless, these high Daptonema 56.9** Marylynnia 71.1** percentages suggest similar (iso-) communities in Halalaimus 51.7* Metadesmolaimus 63.5* CDZs all over the world. Moreover, the nematode Epsilonema 47.1* Molgolaimus 52.1* communities in the sediments of lagoonal seagrass Spirinia 51.5* meadows (Ndaro and Ólafsson 1999; Fisher 2003; Only genera with a signiWcant microhabitat preference are listed. Fisher and Sheaves 2003) and on seagrass blades (Hop- Indicator values and signiWcance levels are provided per and Meyers 1967) were also found to be compara- ***p · 0.001 ble with the communities in the coralline sediment and **0.001 < p · 0.01 on coral fragments of the present study, respectively. *0.01 < p · 0.05 On the other hand, most taxa in lagoonal sediments belong to the same families and genera as those in most sediment group, some considerable diVerences with the temperate, sublittoral sands (Boucher 1997), and espe- list of indicator genera were observed. cially Chromadoridae, Desmodoridae and Xyalidae Marylynnia, Metadesmolaimus, Paracomesoma and become increasingly more important in gradually Molgolaimus are the signiWcant indicator genera for coarser sediments (Heip et al. 1985). Furthermore, the sediment in Matemwe (Zanzibar). The importance very coarse sands yield high abundances of taxa of these genera in the distinction between the Mate- belonging to the families Epsilonematidae and/or Dra- mwe sediment samples and all other samples is con- conematidae (Willems et al. 1982; Ndaro and Ólafsson Wrmed by a SIMPER analysis. 1999). The dominant families in the present study are Epistratum feeders were the dominant trophic group thus explained solely by grain size and are not speciWc in each microhabitat (65.6% in coralline sediment; for this particular habitat. 75.2% on coral fragments). No obvious structuring Communities associated with the coral fragments in eVect on either regional, local or microhabitat scale was CDZs are considered for the Wrst time in our study. found. However, some signiWcant eVects on the individ- This resulted in an increased importance of typical ual trophic groups were detected with an analysis of coarse sand/coarse substratum taxa such as Chromado- variance. For example, the relative abundance of non- ridae and Epsilonematidae. selective deposit feeders (Wieser group 1b) was signiW- There were no new families or genera found in our cantly higher in the sediment (p = 0.01) and the relative samples. This is consistent with the observation by 123 122 Coral Reefs (2007) 26:113–126

Inglis (1968) that in coral reef sediment samples from Matemwe sediment have been recognised as typical New Caledonia, all species were new to science, taxa for Wne, silty sediments (Marylynnia, Comesomati- whereas all genera and families were already dae: Boucher and Gourbault 1990; Wieser and Hopper described. This was explained by the fact that genera at 1967) and oxygen depleted conditions (Molgolaimus: least tend to be cosmopolitan while species do not. Boucher and Gourbault 1990). This observation has been conWrmed in the studies of Furthermore, the diVerences between the communi- Boucher and Gourbault (1990), Gourbault and ties in the coralline sediment and those on the coral Renaud-Mornant (1990), Tietjen (1991) and Boucher fragments were proven to be signiWcant. Taking into (1997). It can be concluded that CDZs do not harbour account the TWINSPAN dendrogram, it can be con- a typical community in terms of taxa restricted to this cluded that there is (1) a principal distinction between system or in terms of new taxa above species level. Wne/medium sediment communities and coarse habitat communities and (2) a distinction between coarse sedi- Is microhabitat structure an additional source ment communities and coral fragment communities on for variation in nematode community composition? a secondary level. There are no signiWcant diVerences between sediment and gravel samples and coral gravel Microhabitat structure is the main factor structuring is also signiWcantly diVerent from coral fragments. the nematode assemblages in CDZs along the east It has been observed that the coral samples cluster coast of Kenya and Zanzibar, as its eVect on the nema- more closely together in the MDS biplots than the sedi- tode community structure overrides that of local or ment samples and that average similarity is higher regional turnover. The nematode communities in between coral samples, whereas average dissimilarity is CDZs seem to have a patchy distribution, determined higher between sediment samples. This could either be by small-scale diVerences in microhabitat structure. explained by (1) diVerences between sediment samples The assemblages are even more aVected by changes in due to variation in grain size or (2) a lack of structuring sediment grain size than by the structural diVerences eVect of the three-dimensional build-up of the coral between sediment and coral fragments, which was evi- fragments. The latter explanation has been conWrmed in denced by the separation of the Matemwe sediment the present study despite of the considerable diVerences samples from both the coral samples and the coarser in surface structure of the Wnely branched Pocillopora sediment samples in the TWINSPAN dendrogram. A compared to e.g. the solid surface of Goniastrea, the granulometric analysis of the Zanzibar sediment sam- grooved surface of Fungia or the complex tubular habi- ples revealed that the sediment in Matemwe had a tus of Tubipora. Govaere et al. (1980) and Vanaverbeke much smaller coarse sand fraction and a larger medium et al. (2002) have demonstrated that slight diVerences in and Wne sand fraction than the coarser Makunduchi grain size, even within the same size class, can funda- sediment. Several studies have indeed shown that nem- mentally inXuence nematode community composition. atode assemblages in CDZs are mainly determined by Moreover, as the majority of nematodes are typically sediment characteristics (Alongi 1986; Boucher and slender, sediment-dwelling organisms (Giere 1993), Gourbault 1990; Gourbault and Renaud-Mornant 1990; which live in the interstitia between the sand grains, Tietjen 1991; Ólafsson et al. 1995; Boucher 1997; Ndaro they are more prone to changes in sediment composi- and Ólafsson 1999; Netto et al. 1999; Kotta and Bou- tion than to changes in the three-dimensional build-up cher 2001; de Jesús-Navarrete 2003). The explanation of the substratum they are associated with. for the separation of the Wner sediments at Matemwe The diVerences between coral associated communi- may be related to food availability, oxygen availability ties and (coarse) sediment associated communities can and hydrodynamics: Wner sediment is found in calm, be attributed to (1) the more exposed nature of the undisturbed conditions, which are characterised by coral microhabitat, (2) diVerences in available surface higher abundances of deposited food and a higher RPD area for epifaunal taxa and (3) the presence of a micro- layer, whereas coarser sediment is typically found in bial bioWlm and algal cover on the dead coral’s surface. conditions characterised by strong hydrodynamic The fauna living on the surface and/or between the stress, resulting in removal of the phytodetritus on the branches of the coral fragments, which lie relatively sediment surface and better oxygenation. These three unprotected on the bottom and protrude from the sedi- variables are known to inXuence nematode community ment, is much more exposed to physical erosion by composition in CDZs (Boucher and Gourbault 1990; current activity than the fauna in the sediment. Gourbault and Renaud-Mornant 1990; Tietjen 1991; Hydrodynamic stress was indeed considerable on most Boucher 1997; Ndaro and Ólafsson 1999; Netto et al. sampling locations (M. Raes, personal observation). 1999). In addition, the indicator genera for Wne/medium As a result, dead coral fragments are to be considered 123 Coral Reefs (2007) 26:113–126 123 preferable habitats only for those nematodes that are genera living in the sediment are also found on corals, able to withstand the current’s eroding eVect, such as even between diVerent regions. The Wve most abun- the epifaunal Epsilonematidae and Draconematidae. dant families are also the same in both microhabitats. Representatives of these two nematode families are As already discussed above, this background commu- morphologically and ecologically well-adapted to phys- nity on family and genus level is typical for coarse, sub- ical disturbance (Willems et al. 1982; Raes and Van- tidal sands. The relatively low number of unique reusel 2006). They are able to walk over diVerent types genera in each microhabitat also supports this idea. At of substratum like inchworms (StauVer 1924; Lorenzen least part of the similarity between coral fragments and 1973), attaching themselves to the surface with specia- the coralline sediment can be explained by sediment- lised setae, adhesive tubes and/or caudal glands with trapping between the coral branches. It is clear that the specially adapted outlets (Raes et al. 2006). In accor- diVerent communities associated with corals and coral- dance with this hypothesis, the genera Paradraconema line sediments from Kenya and Zanzibar, respectively, (Draconematidae) and Epsilonema (Epsilonematidae) are based in particular on diVerent contributions of the were recognised as indicator genera for coral frag- genera that are present and not on the presence of ments and both genera are also among the dominant unique, very speciWc genera restricted to a particular genera on corals. Moreover, short, fat and heavily region or microhabitat. cuticularised nematodes such as Epsilonematidae are also more able to withstand diVerent types of distur- How strong is the turnover in taxonomic composition bance (Soetaert et al. 2002; Vanaverbeke et al. 2004). operating at local and regional scales? The shift in dominance from Chromadoridae on corals to Desmodoridae in the sediment is at present not The extent of spatial turnover on a local and regional well understood, although the larger body size of scale appears to be very much comparable, notwith- desmodorids might be a limiting factor on coral frag- standing the separation of Zanzibar Island from the ments as larger animals tend to be washed away more African mainland by the Zanzibar Channel. The only easily from the coral surface. Trophic segregation indication that a regional eVect may be more important might not play a role here as both Chromadoridae and than a local eVect lies in the absence of clear-cut groups most Desmodoridae are epistratum feeders according in Fig. 3b and the low R-values for local eVects within to Wieser (1953). The desmodorid Eubostrichus (sub- regions (Table 1). Nevertheless, diVerences in nema- family Stilbonematinae), which is the most pronounced tode community structure between regions are rela- indicator genus for the coralline sediment, is known as tively small, given the high number of shared genera a sediment-dweller carrying ectosymbiotic, sulphide- between the microhabitats of both regions and the oxidising chemoautotrophic bacteria on its cuticle in absence or very low number of unique genera for order to survive in the deeper, oxygen depleted layers Kenya or Zanzibar, respectively. This could be related of the sediment (Ott 1995; Ott et al. 2005). to the cosmopolitan nature of nematode genera (see Epifaunal diatoms or an organic coating covers cal- above). Considering the limited structuring eVect of cium carbonate structures such as coral fragments after localities within regions, the low average similarities the death of the living tissue (Suess 1968). The signiW- between samples within the same group (Zanzibar sed- cantly higher relative abundance of epistratum feeders iment, Kenya sediment, Zanzibar coral and Kenya on coral fragments indicates the importance of these coral) are attributed to patchiness. food sources on the coral surface, whereas the signiW- High turnover on a regional scale has been observed cantly lower abundance of non-selective deposit feed- by Kotta and Boucher (2001), with spatially closer ers in this microhabitat is attributed to the low amount regions having higher generic aYnity. These regions of detrital material on the exposed coral fragments due (New Caledonia, Fiji, Moorea, Japan, Great Barrier to removal and resuspension by hydrodynamic activity. Reef, Davies Reef, Guadeloupe, Indian Ocean and Red Epigrowth feeders and/or non-selective deposit feeders Sea) were however much more geographically distant are generally the dominant trophic groups in subtidal from each other than the two regions in the present coralline sediments (Alongi 1986; Gourbault and study. Turnover on a local (km) scale may have either a Renaud-Mornant 1989, 1990; Tietjen 1991; Ólafsson negligible (Heip et al. 1979) or a signiWcant (Li et al. et al. 1995; Boucher 1997; Ndaro and Ólafsson 1999). 1997; Netto et al. 2003) eVect on nematode assemblage Next to the diVerences between communities in both structure. Other studies also provide evidence for signiW- microhabitats, the similarities between these assem- cant turnover between diVerent areas or functional blages are also considerable. The analysis of the num- zones of a reef, i.e. on a scale of hundreds of meters to ber of shared genera has shown that at least 50% of the kilometers (Alongi 1986; Netto et al. 1999, 2003). These 123 124 Coral Reefs (2007) 26:113–126 diVerences were however attributed to diVerent environ- Table 3 continued mental conditions in the lagoon, the reef Xat, reef crest, outer reef, reef pools and tidal Xats. Kotta and Boucher Ammotheristus (Lorenzen 1977) (2001) also found that environmental variables such as Anticoma (Bastian 1865) grain size, silt content and water depth contributed most Aphelenchoides (Fischer 1894) Araeolaimoidea sp. 1 to the variability of nematode assemblages on a local Araeolaimus (de Man 1888) scale, whereas variability on a regional scale was mainly Atrochromadora (Wieser 1959) determined by the geographical position of the sampling Axonolaimus (de Man 1889) stations. Boucher (1997) and Kotta and Boucher (2001) Bathyepsilonema (Steiner 1931) Bolbolaimus (Cobb 1920) observed that the extent of turnover between replicates Bolbonema (Cobb 1920) often exceeded variability between regions and con- Calomicrolaimus (Lorenzen 1976) cluded that the pattern of nematode distribution is rela- Calyptronema (Marion 1870) tively homogeneous over tens of kilometers. This Camacolaimus (de Man 1889) W Cephalobidae gen. 1 signi es that the variation in community composition Ceramonema (Cobb 1920) within locations, due to small-scale diVerences in sedi- Cervonema (Wieser 1954) ment characteristics and environmental conditions, cf. Aegialoalaimus exceeds the variation between localities on this scale. cf. Rhynchonema cf. Rotylenchulus The present survey showed that variations in the Cheironchus (Cobb 1917) structure of the microhabitat and diVerences in envi- Chitwoodia (Gerlach 1956) ronmental conditions occur on very small spatial scales Chromadora (Bastian 1865) and that these small-scale diVerences are the predomi- Chromadorella (Filipjev 1918) Chromadorina (Filipjev 1918) nant factors determining the structure of nematode Chromadorita (Filipjev 1922) assemblages in CDZs. Chromaspirina (Filipjev 1918) Comesomoides (Gourbault 1980) Acknowledgments The authors wish to thank Prof Dr. Kenneth Cricolaimus (Southern 1914) Mavuti (UONBI, Kenya), Dr. Alfonse Dubi, Dr. Desiderius C. P. Croconema (Cobb 1920) Masalu and the people at the Institute of marine Sciences in Zan- Cyartonema (Cobb 1920) zibar (UDSM, Tanzania). Dr. Ann Dewicke, Drs. Tom Gee- Cyatholaimus (Bastian 1865) rinckx, Drs. Hendrik Gheerardyn and Msc Ruth Teerlynck Daptonema (Cobb 1920) helped during sampling. Special thanks go to two anonymous Dasynemoides (Chitwood 1936) reviewers and the editor for critically reading the manuscript and Desmodora (de Man 1889) for providing many constructive remarks. The authors thank Dr. Desmodorella (Cobb 1933) David Obura for identifying the coral fragments; Msc SoWe Dery- Desmoscolex (Claparède 1863) cke and Dr. Wim Bert identiWed the rhabditids and tylenchids. Dichromadora (Kreis 1929) Special thanks go to Annick Van Kenhove, Danny Peelaers, Bart Didelta (Cobb 1920) Beuselinck and Daniëlle Schram for technical support. Gustavo Diodontolaimus (Southern 1914) Fonseca is much acknowledged for his help in constructing the Diplopeltis (Cobb in Stiles and Hassal 1905) ternary graph. The Wrst and second author acknowledge an aspi- Diplopeltula (Gerlach 1950) rant and a postdoctoral fellow grant, respectively, from the Fund Dolicholaimus (de Man 1888) for ScientiWc Research (FWO-Flanders, Belgium). This research Dracognomus (Allen and NoVsinger 1978) was conducted within the framework of the FWO research pro- Dracograllus (Allen and NoVsinger 1978) ject G.0199.03. Draconema (Cobb 1913) Eleutherolaimus (Filipjev 1922) Enoplida gen. n. 1 Enoplida sp. 1 Appendix Enoploides (Ssaweljev 1912) Enoplolaimus (de Man 1893) Table 3 Enoplus (Dujardin 1845) Epacanthion (Wieser 1953) Epsilonema (Steiner 1927) Table 3 List of identiWed genera. Taxonomy after Lorenzen Eubostrichus (GreeV 1869) (1994) and original genus descriptions Euchromadora (de Man 1886) Eurystomina (Filipjev 1921) Acanthonchus (Cobb 1920) Gammanema (Cobb 1920) Acanthopharynx (Marion 1870) Gomphionchus (Platt 1982) Actinonema (Cobb 1920) Halalaimus (de Man 1888) Aegialoalaimus (de Man 1907) Halichoanolaimus (de Man 1886) aV. Leptosomatum Innocuonema (Inglis 1969) aV. Nannolaimoides Laimella (Cobb 1920) Alaimella (Cobb 1920) Latronema (Wieser 1954)

123 Coral Reefs (2007) 26:113–126 125

Table 3 continued Table 3 continued

Leptepsilonema (Clasing 1983) Stylotheristus (Lorenzen 1977) Leptolaimoides (Vitiello 1971) Symplocostoma (Bastian 1865) Leptolaimus (de Man 1876) Synonema (Cobb 1920) Leptonemella (Cobb 1920) Syringolaimus (de Man 1888) Linhomoeus (Bastian 1865) Terschellingia (de Man 1888) Litinium (Cobb 1920) Thalassironus (de Man 1889) Longicyatholaimus (Micoletzky 1924) Thalassoalaimus (de Man 1893) Marylynnia (Hopper 1977) Theristus (Bastian 1865) Mesacanthion (Filipjev 1927) Trefusia (de man 1893) Metachromadora (Filipjev 1918) Trichotheristus (Wieser 1956) Metacyatholaimus (Stekhoven 1942) Tricoma (Cobb 1893) Metadesmolaimus (Stekhoven 1935) Trissonchulus (de Man 1889) Metalinhomoeus (de Man 1907) Trochamus (Boucher and Bovée 1972) Metepsilonema (Steiner 1927) Tubolaimoides (Gerlach 1963) Metoncholaimus (Filipjev 1918) Viscosia (de Man 1890) Microlaimus (de Man 1980) Zalonema (Cobb 1920) Molgolaimus (Ditlevsen 1921) Monhystera (Bastian 1865) Monhystrella (Cobb 1918) Nannolaimus (Cobb 1920) References Neochromadora (Micoletzky 1924) Odontophora (Bütschli 1874) Alongi DM (1986) Population structure and trophic composition Odontophoroides (Boucher and Helléouet 1977) of the free-living nematodes inhabiting carbonate sands of Omicronema (Cobb 1920) Davies Reef, Great Barrier Reef, Australia. Aust J Mar Onchium (Cobb 1920) Freshw Res 37:609–619 Oncholaimus (Dujardin 1845) Boucher G (1997) Structure and biodiversity of nematode assem- Onyx (Cobb 1891) blage in the SW Lagoon of New Caledonia. Coral Reefs Oxystomina (Filipjev 1921) 16:177–186 Papillonema (Verschelde Muthumbi and Vincx 1995) Boucher G, Gourbault N (1990) Sublittoral meiofauna and diver- Paracanthonchus (Micoletzky 1924) sity of nematode assemblages oV Guadeloupe Islands Paracomesoma Home and Murphy 1972) (French West Indies). Bull Mar Sci 47:448–463 Paracyatholaimoides (Gerlach 1953) Clarke KR (1993) Non-parametric multivariate analyses of Paracyatholaimus (Micoletzky 1922) changes in community structure. Austr J Ecol 18:117–143 Paradraconema (Allen and NoVsinger 1978) Clarke KR, Gorley RN (2001) PRIMER v5: User Manual/Tuto- Paralinhomoeus (de Man 1907) rial. PRIMER-E, Plymouth Paramesacanthion (Wieser 1953) de Jesús-Navarrete A (2003) Diversity of Nematoda in a Carri- Paramonohystera (Steiner 1916) bean Atoll. Banco Chinchorro, Mexico. Bull Mar Sci 73:47–56 Pareurystomina (Micoletzky 1930) Deprez T, Vanden Berghe E, Vincx M (2004) NeMys: a multidis- Parodontophora (Timm 1963) ciplinary biological information system. In: Vanden Berghe Paroxystomina (Micoletzky 1924) E, Brown M, Costello MJ, Heip C, Levitus S, Pissierssens P Perepsilonema (Lorenzen 1973) (eds) Proceedings of ‘the colour of ocean data’ Symposium, Phanoderma (Bastian 1865) Brussels, 25–27th November 2002, IOC Workshop Report Polkepsilonema (Verschelde and Vincx 1992) 188, UNESCO, Paris, pp 57–63 Polygastrophora (de Man 1922) Dufrêne M, Legendre P (1997) Species assemblages and indicator Praeacanthonchus (Micoletzky 1924) species: the need for a Xexible asymmetrical approach. Ecol Procamacolaimus (Gerlach 1954) Monogr 67:356–366 Prochromadora (Filipjev 1922) Fisher R (2003) Spatial and temporal variations in nematode Prochromadorella (Micoletzky 1924) assemblages in tropical seagrass sediments. Hydrobiologia Promonhystera (Wieser 1956) 493:43–63 Pseudochromadora (Daday 1899) Fisher R, Sheaves MJ (2003) Community structure and spatial Pseudonchus (Cobb 1920) variability of marine nematodes in tropical Australian pio- Pternepsilonema (Verschelde and Vincx 1992) neer seagrass meadows. Hydrobiologia 495:143–158 Ptycholaimellus (Cobb 1920) Gauch HG Jr, Whittaker RH (1981) Hierarchical classiWcation of Rhabditis (Dujardin 1845) community data. J Ecol 69:135–152 Rhinema (Cobb 1920) Giere O (1993) Meiobenthology: the microscopic fauna in aquat- Rhips (Cobb 1920) ic sediments. Springer, Berlin Heidelberg New York Rhynchonema (Cobb 1920) Gourbault N, Renaud-Mornant J (1989) Distribution, assemblages Richtersia (Steiner 1916) et stratégies trophiques des micro-méiofaunes d’un atoll semi- Sabatieria (Rouville 1903) fermé (Tuamotu Est). CR Acad Sci Paris 309:69–75 Southerniella (Allgén 1932) Gourbault N, Renaud-Mornant J (1990) Micro-meiofaunal com- Spilophorella (Filipjev 1917) munity structure and nematode diversity in a lagoonal eco- Spirinia (Gerlach 1963) system (Fangataufa, Eastern Tuamotu Archipelago). PSZNI Steineria (Micoletzky 1922) Mar Ecol 11:173–189

123 126 Coral Reefs (2007) 26:113–126

Govaere JC, Van Damme D, Heip C, De Coninck LAP (1980) Ott JA (1995) Sulphide symbioses in shallow sands. In: Elefthe- Benthic communities in the Southern Bight of the North Sea riou A, Ansell A, Smith C (eds) Biology and ecology of shal- and their use in ecological monitoring. Helgol Wiss Meere- low coastal waters. Olsen & Olsen, Fredensborg, Denmark, sunters 33:507–521 pp 143–147 Grelet Y (1984) Peuplements meiobenthiques et structure de la Ott J, Bright M, Bulgheresi S (2005) Marine microbial thiotrophic nematofaune du Golfe d’Aqaba (Jordanie-Mer Rouge). ectosymbioses. Oceanogr Mar Biol 42:95–118 Ph.D. thesis, University of Aix-Marseille, France, p140 Raes M, Vanreusel A (2006) Microhabitat type determines the Heip C, Herman R, Bisschop G, Govaere JCR, Holvoet M, Van composition of nematode communities associated with sedi- Damme D, Vanosmael C, Willems KA, De Coninck LAP ment-clogged cold-water coral framework in the Porcupine (1979) Benthic studies of the Southern Bight of the North Sea Seabight (NE Atlantic). Deep-Sea Res Part I 53:1880–1894 and its adjacent continental estuaries. Coordinated research Raes M, Decraemer W, Vanreusel A (2006) Postembryonic mor- actions interuniversitary actions oceanology: symposium re- phology in epsilonematidae, with a discussion on the vari- ports. Programmatie van het Wetenschapsbeleid, Brussels, ability of Caudal Gland Outlets. J Nematol 38(1):97–118 Belgium, Progress Report I, ICES, CM/L9, pp 133–163 Renaud-Mornant J, Gourbault N (1984) Premières prospections Heip C, Vincx M, Vranken G (1985) The ecology of marine nem- meiofaunistiques en Guadeloupe II: Communautés des sa- atodes. Oceanogr Mar Biol Annu Rev 23:399–489 bles littoraux. Hydrobiologia 118:113–118 Hill MO (1979) TWINSPAN—A FORTRAN program for Seinhorst JW (1959) A rapid method for the transfer of nema- arranging multivariate data in an ordered two-way table by todes from Wxative to unhydrous glycerine. Nematologica classiWcation of the individuals and attributes. Ecology and 4:67–69 Systematics, Cornell University, Ithaca, NY Soetaert K, Muthumbi A, Heip C (2002) Size and shape of ocean Hopper BE, Meyers SP (1967) Population studies on benthic margin nematodes: morphological diversity and depth-relat- nematodes within a subtropical seagrass community. Mar ed patterns. Mar Ecol Prog Ser 242:179–193 Biol 1:85–96 StauVer H (1924) Die Lokomotion der Nematoden. Beiträge zur Inglis WG (1968) Interstitial nematodes from St. Vincent Bay, Kausalmorphologie der Fadenwürmer. Zool Jahrb 49:1–118 New Caledonia. In: Expédition Française sur les récifs coral- Suess E (1968) Calcium carbonate interaction with organic com- lines de la Nouvelle-Calédonie. Singer-Polignac, Paris, Edit pounds. Ph.D. thesis, Lehigh University, USA, p153 2:29–74 Thomassin BA, Vivier MH, Vitiello P (1976) Distribution de la KoleV P, Gaston KJ, Lennon JJ (2003) Measuring beta diversity méiofaune des sables corallines de la retenue d’eau épirécif- for presence-absence data. J Anim Ecol 72:367–382 ale du Grand Récif de Tuléar (Madagascar). J Exp Mar Biol Kotta J, Boucher G (2001) Interregional variation of free-living Ecol 22:31–53 nematode assemblages in tropical coral sands. Cah Biol Mar Tietjen JH (1991) Ecology of free-living Nematodes from the 42:315–326 Continental Shelf of the Central Great Barrier Reef Prov- Lennon JJ, KoleV P, Greenwood JJD, Gaston KJ (2001) The geo- ince. Estuar Coast Shelf Sci 32:421–438 graphical structure of British bird distributions: diversity, Vanaverbeke J, Gheskiere T, Steyaert M, Vincx M (2002) Nema- spatial turnover and scale. J Anim Ecol 70:966–979 tode assemblages from subtidal sandbanks in the Southern Li J, Vincx M, Herman PMJ, Heip C (1997) Monitoring meioben- Bight of the North Sea: eVect of small sedimentological thos using cm-, m- and km-Scales in the Southern Bight of diVerences. J Sea Res 48:197–207 the North Sea. Mar Environ Res 43:265–278 Vanaverbeke J, Soetaert K, Vincx M (2004) Changes in morpho- Lorenzen S (1973) Die familie Epsilonematidae (Nematodes). metric characteristics of nematode communities during a Mikrofauna des Meeresbodens 25:1–86 spring phytoplankton bloom deposition. Mar Ecol Prog Ser Lorenzen S (1994) The phylogenetic systematics of freeliving 273:139–146 nematodes. The Ray Society, London Vincx M (1996) Meiofauna in marine and freshwater sediments. McCune B, MeVord MJ (1999) PC-ORD. Multivariate analysis of In: Hall GS (ed) Methods for the examination of organismal ecological data, Version 4.32. MjM Software. Gleneden diversity in soils and sediments. Cab International, Walling- Beach, OR, USA ford, UK, pp 187–195 Ndaro SGM, Ólafsson E (1999) Soft-bottom fauna with emphasis Warwick RM, Platt HM, SomerWeld PJ (1998) Free-living marine on nematode assemblage structure in a tropical intertidal la- Nematodes Part III Monhysterids. In: Barnes RSK, Crother goon in Zanzibar, eastern Africa: I. spatial variability. Hyd- JH (eds) Synopses of the British Fauna (New Series). The robiologia 405:133–148 Linnean Society of London, Shrewsbury, pp 1–296 Netto SA, Warwick RM, Attrill MJ (1999) Meiobenthic and mac- Wieser W (1953) Die Beziehung zwischen Mundhöhlengestalt, robenthic community structure in carbonate sediments of Ro- Ernährungsweise und Vorkommen bei freilebenden mari- cas Atoll (North-east, Brazil). Estuar Coast Shelf Sci 48:39–50 nen Nematoden. Eine ökologisch—morphologische Studie. Netto SA, Attrill MJ, Warwick RM (2003) The relationship be- Arkiv für Zoologie 4:439–483 tween benthic fauna, carbonate sediments and reef morphol- Wieser W, Hopper BE (1967) Marine nematodes of the east coast ogy in reef-Xat tidal pools of Rocas Atoll (north-east Brazil). of North America I Florida. Bull Mus Comp Zool 135:239– J Mar Biol Ass UK 83:425–432 344 Ólafsson E, Johnstone RW, Ndaro SGM (1995) EVects of inten- Willems KA, Vincx M, Claeys D, Vanosmael C, Heip C (1982) sive seaweed farming on the meiobenthos in a tropical Meiobenthos of a sublittoral sandbank in the southern bight lagoon. J Exp Mar Biol Ecol 191:101–117 of the North Sea. J Mar Biol Assoc UK 62:535–548

123