Mesh-Size Matters in Epibenthic Surveys

J. Mar. Biol. Ass. U. K. 12002), 82,1^8 Printed in the United Kingdom

Mesh-size matters in epibenthic surveys

Ruth Callaway*P, Simon JenningsP, John Lancaster* and John CotterO *University ofWales Swansea, School ofBiological Sciences, Singleton Park, Swansea, Wales, SA2 8PP. OCentre for Environment, Fisheries and Aquaculture Science, Lowestoft Laboratory, Lowestoft, Norfolk, NR33 0HT. PCorresponding author 1ne¨ eR.Zu« hlke): e-mail: [email protected]

This study aimed to identify the e¡ects of di¡erent sieve mesh-sizes on processing time, the number of species retained, diversity measures and multivariate community analysis in the North Sea. Samples were collected at 63 sites throughout the North Sea and washed through two successive sieves, 10-mm and 5-mm mesh respectively. Processing time for whole samples 15- and 10-mm fraction) averaged 91 Æ25 min compared with 55 Æ16 min for the 10-mm mesh fraction. Altogether 40% of free-living species and 9% of attached species were recorded exclusively in the 5-mm fraction. The majority of these species were rare. Spatial gradients of species diversity and community structure were identical, independent ofthe mesh-size used. Multivariate community analysis showed no signi¢cant di¡erence between descriptions ofcommunity structure based on fauna from 10-mm or 5-mm mesh. The use ofcoarser sieving mesh would save time and money, ifthe aims ofan epibenthic survey were to describe broad patterns ofcommunity structure and relative diversity. It would be possible to process approximately 50% more samples, ifthe time saved with 10-mm mesh were allocated to additional sampling. However, ifinformation on single species is required, then sorting with the ¢ner sieve mesh will yield crucial information. It was decided to employ a 5-mm mesh for epibenthic monitoring of the North Sea.

INTRODUCTION or temporal replication. The aspiration to use ¢nancial resources and time as productively as possible initiated Currently ¢ve European countries 1Germany, England, research on e¡ects ofmesh-size on infauna studies 1Reish, Norway, the Netherlands and Denmark) contribute to an 1959; Rees, 1984; Bachelet, 1990; James et al., 1995; epibenthic monitoring programme in the North Sea Schlacher & Wooldridge, 1996). Most studies assessed 1Jennings et al., 1999; Zu« hlke et al., 2001). The aim is to 1-mm and 0.5-mm mesh while Schlacher & Wooldridge describe spatial diversity patterns based on univariate and 11996) and Reish 11959) evaluated e¡ects ofeven smaller multivariate data analysis, to analyse the distribution of sizes. individual species and to provide a baseline study for Although studies ofepibenthic invertebrates were future analysis oftemporal changes ofthese patterns. Two among the ¢rst faunal investigations in the North Sea hundred and forty-one stations in 143 ICES rectangles 1e.g. Peterson, 1914; Peterson, 1918), sampling methods 1International Council for the Exploration of the Sea, have not been widely standardized 1Dyer et al., 1983; boxes 0.58 latitudeÂ18 longitude) were sampled in 1999. Reise & Bartsch, 1990; Basford & Ra¡aelli, 1990; Clearly, on such a large, multi-participant survey, a stand- Bergmann & Hup, 1992). Beside the variation in sampling ardized sampling protocol that speeds sample processing equipment 1dredges and several di¡erent types oftrawls), without losing the required biological information can processing ofsamples di¡ered considerably between the provide crucial cost and time savings. Such a concern studies. Some samples were not sieved 1Basford et al., prompted the research described in this paper. 1989; Hosten, 2000), while others were sieved with Studies ofthe distribution and diversity ofmarine meshes from 5-mm 1Rees et al., 1999) to 14-mm benthic invertebrates generally require that samples are 1Frauenheim et al., 1989). collected and sieved through mesh screens. The choice of The aims ofthis study were to determine how 10-mm mesh-size depends on the objectives ofa study, the compo- mesh, in comparison to 5-mm mesh, a¡ects: 1i) the time nent of fauna investigated 1meiofauna, macrofauna), the required to sort and identify epibenthic invertebrate coarseness ofthe sediment and the importance of samples; 1ii) the range ofspecies sampled; and 1iii) esti- collecting juveniles or smaller species. Finer meshes will mations ofspecies diversity and community structure. retain a larger number ofindividuals and species, providing more complete information of the community and populations. However, they will increase the time MATERIALS AND METHODS and cost ofprocessing 1Rees, 1984; Schlacher & Sampling Wooldridge, 1996). Larger meshes should allow more samples to be processed per unit time, which releases time Epibenthic invertebrates were sampled at 63 stations and ¢nancial resources that can be used to increase spatial throughout the North Sea 1528N^618N) from August to

Journal of the Marine Biological Association of the United Kingdom 2002) 2 R. Callaway et al. Mesh-size matters in epibenthic surveys

Table 1. Time needed to sort and identify samples sieved September 1996. At each station one sample was collected through two hierarchically ordered mesh sizes 10- and 5-mm). with a 2-m-beam trawl, equipped with a 2-mm mesh liner. Times are given as mean minutes ÆSE, N63. The precise locations ofstations and the beam trawl design were described in Jennings et al. 11999). 10-mm 5-mm Total sample Each catch was sieved through two successive metal fraction fraction 15- and 10-mm) sieves, the ¢rst with 10-mm square mesh, the second with 5-mm mesh 1internal measurement). The material Free-living fauna 34 Æ724Æ956Æ15 retained on each sieve, referred to as 5-mm and 10-mm Attached fauna 20 Æ10 3 Æ425Æ13 fractions respectively, was treated separately. All Total fauna 55 Æ16 32 Æ12 91 Æ25 epibenthic species were sorted from the retained material

Table 2. Species unique to 5-mm fraction. Ubiquity indicates the number of sites a species was found. Abundance is the total number of individuals found throughout the study. For attached species abundance was not recorded.

Ubiquity Abundance

Attached species Hydrozoa Obelia longissima 1Pallas, 1766) 1 ^ Halecium baenii 1Johnston, 1838) 1 ^ Amphisbetia operculata 1Linnaeus, 1758) 1 ^ Diphasia pinaster Hincks, 1861 1 ^ Bryozoa Amphiblestrum auritum 1Hincks, 1877) 1 ^ Amphiblestrum £emingii 1Busk, 1854) 1 ^ Palmiskenea lorea 1Alder, 1864) 1 ^ Hexacorallia Caryophyllia smithii var. clavus Stokes & Broderip, 1828 2 ^

Free-living species Polychaeta Laetmonice ¢licornis Kinberg, 1855 1 1 Gattyana cirrosa 1Pallas, 1766) 1 1 Lepidonotus squamata 1Linnaeus, 1758) 1 1 Amphipoda Lophogaster typicus M. Sars, 1857 1 3 Epimeria cornigera 1Fabricius, 1779) 2 2 Tmetonyx cicada 1Fabricius, 1779) 3 7 Ampelisca macrocephala Liljeborg, 1852 2 8 Ampelisca spinipes Boeck, 1861 2 2 Maera loveni 1Brucellius, 1859) 1 1 Melita dentata 1KrÖyer, 1842) 2 3 Isopoda Rocinela damnoniensis Leach, 1815 1 1 Astacilla longicornis 1Sowerby, 1806) 2 2 Decapoda Pandalina brevirostris 1Rathke, 1837) 3 3 Philoceras bispinosus 1Hailstone, 1835) 2 5 Philoceras echinulatus 1M. Sars, 1861) 1 1 Philoceras trispinosus 1Hailstone, 1835) 2 15 Philoceras sculptus 1Bell, 1847) 1 1 Anapagurus chiroacanthus 1Liljeborg, 1856) 1 2 Anapagurus hyndmanni 1Bell, 1845) 2 2 Munida rugosa 1Fabricius, 1775) 1 1 Ebalia cranchii Leach, 1817 7 12 Gastropoda Emarginula ¢ssura 1Linnaeus, 1758) 1 1 Trivia arctica 1Pulteney, 1799) 1 1 Velutina velutina 1O.F. Mu« ller, 1776) 4 4 Polinices pulchellus 1Risso, 1826) 1 1 Polinices fuscus 1de Blainville, 1825) 2 5 Polinices montagui 1Forbes, 1838) 1 1 Opisthobranchia Archidoris pseudoargus 1Rapp, 1827) 1 1 Aeolidia papillosa 1Linnaeus, 1761) 1 1 Bivalvia Nucula nitidosa Winckworth, 1930 4 28 Palliolum tigerinum 1O.F. Mu« ller, 1776) 1 1 Palliolum striatum 1O.F. Mu« ller, 1776) 1 1 Turtonia minuta 1Fabricius, 1780) 1 1 Corbula gibba 1Olivi, 1792) 3 56 Cuspidaria cuspidata 1Olivi, 1792) 1 3 Cephalopoda Sepiola atlantica Orbigny, 1840 5 7 Rossia macrosoma 1delle Chiaje, 1830) 1 1 Ophiuroidae Amphiura brachiata 1Montagu, 1804) 1 1 Amphiura chiajei Forbes, 1845 2 21 Amphiura ¢liformis 1O.F. Mu« ller 1776) 1 1 Holothuriidae Leptopentacta elongata 1Dueben & Koren, 1844) 1 2 Ocnus lacteus 1Forbes & Goodsir, 1839) 1 1 Pseudothyone raphanus 1Dueben & Koren, 1845) 1 1

Journal of the Marine Biological Association of the United Kingdom 2002) Mesh-size matters in epibenthic surveys R. Callaway et al. 3 onboard. The majority ofspecies were identi¢ed un- samples. Time to remove formalin and wash the sample preserved, but some were kept in 4% formalin for sub- was not included. sequent identi¢cation in the laboratory. Species were classi¢ed as `free-living' or `attached', the latter being recorded as present/absent only. Lists ofspecies assigned Diversity indices to free-living and attached categories are given in Jennings Diversity offree-living epibenthos was measured using et al. 11999).Todetermine species composition and abund- the Hill's indices N0 and N1 1Hill, 1973), where ance for 5-mm mesh, data for the 5-mm and the 10-mm N0number ofspecies 1species richness) and N1 exp fractions were combined. 1H), where H is Shannon^Wiener diversity. Hill's N0 is the total number ofspecies in a sample, while N1 incorpo- Sorting and identi¢cation time rates individual abundance. Analysis ofvariance The time required to sort and identify free-living and 1ANOVA) was applied to test for signi¢cant di¡erences attached species in each fraction was recorded for all between latitudes.

Table 3. Presence and abundance of species using a 10-mm or a 5-mm mesh. Ubiquity indicates the number of sites a species was found. Abundance is the total number of individuals found throughout the study.

Ubiquity Abundance

10 -mm 5- and 10 -mm 10-mm 5- and 10-mm

Cirolana borealis Liljeborg, 1851 1 7 1 29 Crangon crangon 1Linneaus, 1758) 29 45 303 1637 Pandalus montagui Leach, 1814 13 19 87 387 Spirontocaris lilljeborgi 1Daniellsen, 1859) 11 16 96 414 Pontophilus spinosus 1Leach, 1815) 3 7 4 15 Pagurus bernhardus 1Linnaeus, 1758) 44 50 630 977 Anapagurus laevis 1Bell, 1845) 19 25 250 464 Pagurus pubescens KrÖyer, 1838 16 19 86 161 Hyas coarctatus Leach, 1815 30 35 95 185 Galathea dispersa Bate, 1859 3 6 4 18 Macropodia rostrata 1Linnaeus, 1761) 5 8 12 20 Buccinum undatum Linnaeus, 1758 20 23 64 72 Ophiura a¤nis Luetken, 1858 2 6 2 8

Figure 1. Cumulative species richness based on all 63 samples taken during the North Sea survey. Samples were randomly ordered 125 randomizations) and additional species found in added sample were estimated. Mean numbers of species 1 ÆSD) for the10-mmsievefractionandthe5-&10-mmfractionaregiven.

Journal of the Marine Biological Association of the United Kingdom 2002) 4 R. Callaway et al. Mesh-size matters in epibenthic surveys

Figure 2. Diversity of free-living and sessile fauna, as well as individual abundance, in relation to latitude for two di¡erent mesh sizes: 5- & 10-mm and 10-mm, means ÆSE, the number ofstations per latitude 1N) varied between 3 and 9. 1A) Hill's N0 1species richness) for free-living fauna; 1B) Hill's N1 1exp 1Shannon-index)) for free-living fauna; 1C) Hill's N0 1species richness) for sessile fauna; 1D) individual abundances of free-living species.

Figure 3. Multi-dimensional scaling 1MDS) ordination for epibenthic fauna of the northern and southern North Sea. All species were included and numbers were transformed as presence/absence. White squares, fauna retained by 5- and 10-mm sieve; grey squares, fauna retained by 10-mm sieve. Numbers indicate corresponding stations. Northern North Sea, sampling stations north of 568N, stress 0.23; Southern North Sea, sampling stations south of56 8N, stress 0.20. The two groups, 10-mm fraction and 5- & 10- mm fraction, were not signi¢cantly di¡erent in both geographical regions 1ANOSIM, P40.05). Several stations are hidden by overlapping positions in the MDS plot.

Journal of the Marine Biological Association of the United Kingdom 2002) Mesh-size matters in epibenthic surveys R. Callaway et al. 5

Analysis of community structure fraction were also assessed calculating the Spearman rank Multivariate statistical techniques were used to correlation coe¤cient [r] between the respective describe similarities between epibenthic communities similarity matrices. This was carried out by the PRIMER from di¡erent stations. Analysis was carried out with programme RELATE. the PRIMER statistical package 1Clarke & Warwick, Analysis ofsimilarity 1ANOSIM) was used to test for 1994). signi¢cant di¡erences between samples sieved with 10-mm The term `community' is used here for assemblages of and 5- & 10-mm mesh 1Clarke & Warwick, 1994). similar sessile or free-living species, characteristic for Samples from the northern and southern North Sea were certain areas. Ecological interactions are not implied. analysed separately to reduce some ofthe spatial variation For multivariate analysis, free-living and sessile fauna in the data. were combined and abundance transformed to presence/ absence. Additionally, free-living fauna was analysed separately, which allowed the incorporation ofabundance. RESULTS Numbers were 4th root transformed, which down- Sorting and identi¢cation time weighted abundant species. The Bray^Curtis similarity index was calculated The sorting and identi¢cation time increased by 60% between each possible pair ofsamples and organized in when the whole sample 15- and 10-mm fraction), rather similarity matrices. These were the bases for non- than the 10-mm fraction was analysed 1Table 1). Process- parametric multi-dimensional scaling 1MDS). Stress ing offree-living fauna generally took longer than levels of0.19 to 0.23 were relatively high forall MDS or- attached fauna. The 5-mm fraction prolonged the process- dinations, so that too much reliance should not be placed ing time for free-living fauna by 80% and for attached on the detail ofthe plots 1Clarke & Warwick, 1994). The fauna by 10%. The increase for attached fauna was di¡erences between the two sets ofmultivariate data relatively small because most species were retained by the attained from the 10-mm and the 5- & 10-mm sieve 10-mm mesh.

Figure 4. Multi-dimensional scaling 1MDS) ordination for epibenthic fauna. Each point represents one of 63 stations. 1A) Sessile and free-living species, 10-mm sieve fraction, presence/absence data transformation, stress 0.21; 1B) sessile and free-living species, pp5- and 10-mm sieve fraction, presence/absence data transformation, stress 0.20; 1C) free-livingpp fauna, 10-mm sieve fraction, data transformation, stress 0.20; 1D) free-living fauna, 5- and 10-mm sieve fraction, data transformation, stress 0.19. Position ofsampling stations, *,60^628N; &,58^608N; ^,56^588N; &,54^568N; *,51^548N.

Journal of the Marine Biological Association of the United Kingdom 2002) 6 R. Callaway et al. Mesh-size matters in epibenthic surveys

Mesh-size e¡ects at species level attached fauna 110%). The 5-mm mesh retained smaller Nearly 40% ofall free-living species were exclusively and rarer free-living species which often required specia- found in the 5-mm fraction 143 out of a total of 113 lized keys for identi¢cation. This contributed considerably species).They were representative ofa wide range oftaxo- to the increase in processing time. Most ofthe attached nomic groups 1Table 2). More than 80% ofthem were species were retained by the 10-mm sieve as a consequence found at one or two stations only and could be classi¢ed ofthe size ofshells and stones they settled on. Hence, as rare species. However, some occurred at up to seven processing the 5-mm fraction did not add much to stations and were found in considerable numbers, for sorting or identi¢cation time for attached species. Alto- example Ebalia cranchii, Corbula gibba and Sepiola atlantica. gether the time saved by sorting with 10-mm mesh Only 9% ofattached species were uniquely found in the rather than 5-mm mesh would allow 50% more samples 5-mm fraction 19 of 97 species). Of those, all bar one to be processed in the same time. were recorded at a single station only. Other studies also acknowledge the considerable As expected, the recorded abundance ofspecies was increase in processing time when using smaller sieve mesh lower in the 10-mm fraction than in the whole samples 1Reish, 1959; James et al., 1995). James et al. 11995) found 15- and 10-mm fraction). However, abundances of several that sorting time for infauna doubled or quadrupled, species were over 40% lower 1Table 3) and species like depending on location, when using 0.5-mm instead of Crangon crangon, Pandalus montagui and Pagurus bernhardus 1.0-mm mesh. were recorded at 35% fewer stations when samples were sieved with 10-mm mesh. Species level Almost 40% ofall species were foundsolely in the Diversity indices and total abundance 5-mm fraction. For most species this was a consequence of A total of210 species was identi¢ed when using the addi- their small size, but for others it was a result of their parti- tional 5-mm sieve, 158 species when using just the 10- cular body shape or texture. The sieve mesh is clearly a mm mesh 1Figure 1). The average number ofnew species rather crude approach for size fractionation of the recorded with every additional sample was 3.0 1 Æ2.4) for sample, since the probability ofindividuals passing the 5- and 10-mm mesh and 2.3 1 Æ1.9) for the 10-mm through the mesh will depend on their orientation and mesh. morphology. Shrimps, such as Crangon allmanni may be Diversity measures and relative individual abundance well over 10 mm long, but can be washed away because were generally lower for the 10-mm fraction than for the they ¢t through the mesh lengthways. Other species may whole sample 15- and 10-mm fraction) 1Figure 2). squeeze through the 10-mm mesh due to their soft bodies, However, the same spatial trends in species diversity were such as 3 cm long Sepiola atlantica 1personal observation). detected by analysing the 10-mm fraction and the whole Rees 11984) noted that the proportion ofmaterial retained sample 15- and 10-mm fraction). Signi¢cant increases on the sieve was also dependent on the degree offragment- 1ANOVA, P50.05) in species richness 1Hill's N0) and ation during the ¢eld sampling, which varied with the care diversity 1Hill's N1) from the southern to the northern applied in the washing down ofsamples. On the other North Sea were revealed by both the 10-mm fraction and hand he found that the sieve retained species such as the 5- & 10-mm fraction. No clear spatial trend was found tube-dwelling polychaetes, which would have slipped for individual abundance, independent of the sample through the mesh without their comparatively large tubes. fraction. The considerable number ofspecies that were unique to the 5-mm fraction emphasizes the value of smaller mesh- sizes when assessing species composition. Given that many Community analysis ofthe species retained on the 5-mm mesh were relatively No signi¢cant di¡erence was found between epibenthic `rare', their presence may provide important information community structure derived from the 10-mm fraction on geographical distribution, even ifsome ofthem argu- and the 5- & 10-mm fraction 1Figure 3). Similar to ably belong to infauna 1e.g. Nucula nitidosa) and were species diversity, multivariate data analysis showed a certainly not sampled quantitatively. Shifts of distribution north^south gradient for species communities of the boundaries and occurrence ofimmigrating or imported 10-mm mesh as well as 5- and 10-mm mesh 1Figure 4). species are more likely to be detected ifsmaller mesh- This was found for all species, analysed as present or sizes are used and rarer species identi¢ed. Reise & absent, as well as for free-living fauna taking abundances Bartsch 11990) pointed out that rare species may often be into account 1Figure 4). Spearman rank correlation was habitat specialists, which contribute to higher regional high between similarity matrices. It was 0.91 for the simi- heterogeneity. However, ecological studies ofrare species larity matrices based on all species 1sessile and free-living are challenging 1Chapman, 1999), since such species species; presence/absence data transformation) and 0.90 cannot be sampled to a speci¢ed precision, irrespective of for free-living species 14th root data transformation). the size ofmesh used 1Schlacher & Wooldridge, 1996). Rare species are, by de¢nition, not abundant and this DISCUSSION may violate assumptions ofunivariate statistics, ruling out comparison ofabundance at species level 1James et The overall time needed to sort and identify epibenthic al., 1995). The authors suggest that they be best incorpor- samples increased by about 60% ifthey were sieved ated into multivariate methods, although the results ofour through a mesh of5-mm rather than 10-mm. The increase study suggest that rare species had only a minor e¡ect on in time was greater for free living fauna 180%) than for the output ofmultivariate analyses.

Journal of the Marine Biological Association of the United Kingdom 2002) Mesh-size matters in epibenthic surveys R. Callaway et al. 7

Diversity indices data for such analysis 1Reish, 1959; Schlacher & As expected, more species were recorded at any one Wooldridge, 1996) and it was therefore decided to use the station when using 5-mm mesh than 10-mm mesh. 5-mm mesh in the epibenthic monitoring programme that However, the spatial trend in species richness 1Hill's N0) prompted this mesh-size study. offree-living or sessile fauna, as well as species diversity 1Hill's N1), was almost identical for both mesh sizes. Absolute abundance was considerably lower inthe10-mm We wish to thank the o¤cers and crew ofRV `Cirolana'; Barrie fraction for several species, but information on abundances Horton, Mike Kaiser and Hubert Rees for assistance with beam- from beam-trawl samples may be of limited value. Beam- trawl and sieve design; Nina Goodyear, Andrew Woolmer, Pete Cadman, Pete Hayward and Nathalie Yonow for taxonomic trawls are generally regarded as a semi-quantitative advice. This study was funded by the European Community sampling device 1Basford et al., 1989) because of their 1EC FAIR CT95-0817). variable performance on di¡erent types of seabed 1Rees, 1999). REFERENCES Community structure Bachelet, G., 1990. The choice ofa sieving mesh-size in the quantitative assessment ofmarine macrobenthos: a necessary Multivariate analysis based whole samples 15- and compromise between aims and constraints. Marine 10-mm fraction) and 10-mm fractions revealed similar Environmental Research, 30,21^35. spatial patterns ofspecies assemblages. Multi-dimensional Basford, D.A.E. & Ra¡aelli, D., 1990. The infauna and epifauna scaling 1MDS) showed a north^south gradient, indepen- ofthe northern North Sea. Netherlands Journal of Sea Research, dent ofthe mesh-size used. The same spatial pattern had 25,165^173. been found by other authors 1Dyer et al., 1983; Basford, D.J., Eleftheriou, A. & Ra¡aelli, D., 1989. The epifauna Frauenheim et al., 1989; Heip et al., 1992; Rees, 1999). ofthe northern North Sea 156 ^618N). Journal of the Marine Corresponding geographical trends were also found for Biological Association of the United Kingdom, 69,387^407. other benthic groups like meiofauna 1Huys et al., 1990) Bergmann, M.J.N. & Hup, M., 1992. Direct e¡ects ofbeam- and macroinfauna 1Ursin, 1960 [echinoderms]; trawling on macrofauna in a sandy sediment in the southern Kirkegaard, 1969 [polychaetes]; Heip & Craeymeersch, North Sea. ICES Journal of Marine Science, 49,5^11. Chapmann, M.G., 1999. Are there adequate data to assess how 1995). well theories ofrarity apply to marine invertebrates? In contrast to our results, Schlacher & Wooldridge Biodiversity and Conservation, 8,1295^1318. 11996) concluded that the interpretation ofcommunity Clarke, K.R., 1993. Non-parametric multivariate analysis of structure changed markedly with mesh-size. However, changes in community structure. Australian Journal of Ecology, their assertion was based on the retention e¤ciency 18,117^143. results ofmajor taxonomic groups ofinfauna rather than Clarke, K.R. & Warwick, R.M., 1994. Change in marine commu- multivariate statistical analysis. Applying MDS analysis to nities: an approach to statistical analysis and interpretation. 4th root transformed infauna-data, James et al. 11995) Plymouth: Natural Environment Research Council. showed that neither mesh-size nor taxonomic resolution Dyer, M.F, Frey,W.G., Frey, P.D. & Cranmer, G.J., 1983. Benthic during species identi¢cation had major e¡ects on results regions within the North Sea. Journal of the Marine Biological Association of the United Kingdom, 63,683 693. ofspatial patterns ofinfauna species assemblages. In ^ Frauenheim, K., Neumann,V.,Thiel, H. & Tu« rkay, M., 1989. The terms ofmesh-size this coincides with our results on distribution oflarger epifauna during summer and winter in epibenthic communities. the North Sea and its suitability for environmental monitoring. Senckenbergiana Maritima, 20,101^118. CONCLUSIONS Gray, J.S., Clarke, K.R., Warwick, R.M. & Hobbs G., 1990. Detection ofinitial e¡ects ofpollution on marine benthos: an Our study highlights the importance ofde¢ning the example from the Eko¢sk and Eld¢sk oil¢elds, North Sea. objectives ofa study before the size ofsieve is determined. Marine Ecology Progress Series, 66, 285^299. Major e¤ciency gains, in terms oftime, money or Heip, C. et al., 1992. Trends in biomass, density and diversity increased replication are possible iflarger mesh can be ofNorth Sea macrofauna. ICES Journal of Marine Science, 49, used. Large mesh is appropriate for broad descriptions of 13^22. Heip, C. & Craeymeersch, J.A., 1995. Benthic community assemblage structure based on multivariate analyses. structure in the North Sea. Helgola« nder Meeresuntersuchungen, 49, Multivariate data analysis has been proven to be sensitive 313^328. to variation in the structure ofmacrofaunal assemblages Hill, M.O., 1973. Diversity and evenness: a unifying notation and 1Clarke, 1993) and is regarded as being particularly its consequences. Ecology, 54,427^432. useful for environmental impact assessments and routine Hosten, K., 2000. Spatial patterns and seasonality in the marine monitoring work 1Gray, 1990; Warwick, 1993). epibenthic communities ofthe Westerschelde 1Southern Bight However, such monitoring programmes not only describe ofthe North Sea). Journal of the Marine Biological Association of the the current situation, but often aspire to detecting United Kingdom, 80,27^36. temporal changes, which can be subtle and may be Huys, R., Herman, P.M.J., Heip, C.H.R. & Soetaert, K., 1992. restricted to a few species. This study indicates that The meiobenthos ofthe North Sea: density, biomass trends and distribution ofcopepod communities. ICES Journal of multivariate techniques might not be sensitive enough to Marine Science, 49,23^44. detect such changes. In our monitoring programme it was James, R.J., Lincoln-Smith, M.P. & Fairweather, P.G., 1995. decided not to rely on multivariate data analysis alone, but Sieve mesh-size and taxonomic resolution needed to describe place importance on the analysis ofspatial patterns of natural spatial variation ofmarine macrofauna. Marine single species. Smaller sieve mesh-sizes provide essential Ecology Progress Series, 118,187^198.

Journal of the Marine Biological Association of the United Kingdom 2002) 8 R. Callaway et al. Mesh-size matters in epibenthic surveys

Jennings, S., Lancaster, J., Woolmer, A. & Cotter, J., 1999. Reish, D.J., 1959. A discussion ofthe importance ofthe screen Distribution, diversity and abundance ofepibenthic fauna in size in washing quantitative marine bottom samples. Ecology, the North Sea. Journal of the Marine Biological Association of the 40,307^309. United Kingdom, 79,385^399. Schlacher, T.A. & Wooldridge, T.H., 1996. How sieve mesh-size Kirkegaard, J.B., 1969. A quantitative investigation ofthe central a¡ects sample estimates ofestuarine benthic macrofauna. North sea: Polychaeta. Spolina Zoologica Musei Hauniensis, 29, Journal of Experimantal Marine Biology and Ecology, 201,159^171. 1^197. Ursin, E., 1960. A quantitative investigation ofthe echinoderm Petersen, C.G.J., 1914. Valuation ofthe sea. II. The animal fauna of the central North Sea. Meddelelser fra Danmarks communities ofthe sea bottom and their importance for Fiskfri-og Havundersogelser, 2, 1^204. marine zoography. Report of the Danish Biological Station, 21,1^44. Warwick, R.M., 1993. Environmental impact studies on marine Petersen, C.G.J., 1918. The sea bottom and its production ofsea- communities: pragmatical considerations. Australian Journal of food. Report of the Danish Biological Station,1^62. Ecology, 18,63^80. Rees, H.L., 1984. A note on mesh selection and sampling Zu« hlke, R. et al., 2001. Epibenthic diversity in the North Sea. e¤ciency inbenthic studies. MarinePollutionBulletin,15,225^229. In Burning issues of North Sea ecology. Proceedings of the Fourteenth Rees, H.L., Pendle, M.A., Waldrock, R., Limpenny, D.S. & International Senckenberg Conference North Sea 2000 1ed. Boyd, S.E., 1999. A comparison ofbenthic biodiversity in the I. Kro« ncke et al.). Senckenbergiana Maritima, 31,inpress. North Sea, English Channel, and Celtic Seas. ICES Journal of Marine Science, 56, 228^246. Reise, K. & Bartsch, I., 1990. Inshore and o¡shore diversity of epibenthos dredged in the North Sea. NetherlandsJournal of Sea Research, 25,175^179. Submitted 4 December 2000. Accepted 13 November 2001.

Journal of the Marine Biological Association of the United Kingdom 2002)