Hydrobiologia 389: 21–28, 1998. © 1998 Kluwer Academic Publishers. Printed in the Netherlands.
21
The effect of sieve mesh size on the abundance and composition of macrophyte-associated macrofaunal assemblages Marcel Okamoto Tanaka1,∗ & Fosca Pedini Pereira Leite2 1 Programa
de P´os-Graduação em Ecologia, Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, CEP 13083-970, Campinas, SP, Brazil E-mail:
[email protected] 2 Departamento de Zoologia, Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, CEP 13083-970, Campinas, SP, Brazil (∗ author for correspondence) Received 2 June 1998; in revised form 14 October 1998; accepted 23 October 1998
Key words: epifauna, macrofauna, Sargassum, sampling
Abstract Sampling efficiency of several sieve mesh sizes (2, 1, 0.5, 0.2 and 0.1 mm) was tested in invertebrate assemblages associated to Sargassum stenophyllum. Samples of this brown algae were collected in southeastern Brazil to determine whether different sampling devices resulted in differences on the abundance and composition of the macrofauna. The 1 and 2 mm mesh sizes sampled less organisms, resulting in smaller abundances, densities, number of species and diversity, when compared to the other three mesh sizes, that achieved similar results. The most efficient sieve was the 0.5 mm mesh, sampling 85.8% and 94.5% of gammaridean and caprellidean amphipods, respectively, and 93.1% of gastropods. However, polychaetes and isopods were best sampled with the 0.2 mm sieve. Our results suggest that caution should be exercised when deciding on which sieve to use, as mesh sizes commonly used in benthic studies (1 and 0.5 mm) may result in very different estimates of diversity and abundance, as well as community structure patterns.
Introduction The patterns of distribution and abundance of benthic communities can be strongly influenced by the sampling methods, including the number of samples, sieve sizes used to process the samples and the taxonomic resolution used in the analysis (Bachelet, 1990; James et al., 1995; Somerfield & Clarke, 1995; Schlacher & Wooldridge, 1996). In these communities, the macrofauna is defined as all organisms retained in a 0.5 mm sieve, although there is some spread around the size modes (Schwinghamer, 1981). However, the fraction of juvenile individuals of the macrofauna may eventually be lost when 1 or 0.5 mm sieves are used, compromising studies of population dynamics (Bachelet, 1990). Thus, the sampling methods should consider both the precision required to the objectives of the study, and the costs involved to attain such precision
(Kingston & Riddle, 1989; Schlacher & Wooldridge, 1996). Comparative studies on the influence of different sieve sizes on the faunal composition have been conducted mainly on infaunal benthic communities (Bachelet, 1990; James et al., 1995; Schlacher & Wooldridge, 1996). In general, the comparison of 0.5 and 1 mm sieves indicate that a substantial part of the community is lost when the latter sieve is used (Bachelet, 1990). There are no critical studies of the effect of different sieve sizes on the estimates of composition and abundance of macrophyte-associated invertebrate assemblages, despite the great number and diversity of papers published on the subject. Marine macrophytes as algae and seagrasses shelter large numbers of mobile animals, frequently dominated by gastropods, polychaetes and small crustaceans as amphipods, isopods and others (Edgar &
22 Moore, 1986; Taylor & Cole, 1994). Due to its relative ease of manipulation, this system has been used to test several hypotheses on the structure of ecological communities, including the effects of habitat complexity (Heck & Wetstone, 1977; Hacker & Steneck, 1990; Gee & Warwick, 1994), island biogeography (Gunnill, 1982a, b; Stoner, 1985), and community stability (Virnstein & Curran, 1986; Martin-Smith, 1994). The influence of environmental factors and biological interactions on community structure have also been addressed, commonly through comparative studies (Fenwick, 1976; Edgar, 1983; Taylor & Cole, 1994). However, the sieve mesh sizes used to process the samples varied between 0.163 mm (Duffy, 1990) and 1 mm (Nelson, 1979; Gunnill, 1982a, b; Taylor & Cole, 1994) and, due to a lack of critical evaluation of the relative efficiency of different sieve sizes, there are no clues about which fraction of the community was used to test the hypotheses. Thus, the conclusions of these studies could differ if a greater or smaller fraction of the assemblage was sampled (Bachelet, 1990). In this study, we ask whether the use of different sieve mesh sizes (0.1, 0.2, 0.5, 1 and 2 mm) influence the conclusions about (1) the abundance of the associated mobile fauna, and (2) the composition of gammaridean amphipods associated to the brown algae Sargassum stenophyllum (Mertens) Martius in two shores located in southeastern Brazil.
Study areas Sampling was carried out during september 1997, in two shores located on the northern coast of São Paulo State: Lázaro (23◦ 310 S – 45◦ 080 W) and Tabatinga (23◦ 340 S – 44◦ 160 W). Sampling sites in each shore were located in the sublittoral fringe, and all samples were collected in the morning, during low tides. Lázaro is a semi-exposed shore located in Ubatuba district. The community is dominated by Sargassum stenophyllum, also occurring other species of brown algae as Dictyopteris delicatula Lamouroux and D. plagiogramma (Montagne) Vickers (Eston & Bussab, 1990). Tabatinga is a protected shore in Caraguatatuba district. Sampling sites were located 200 m beyond the mouth of Tabatinga River, and were dominated by Sargassum stenophyllum and Galaxaura marginata (Ellis & Solander) Lamouroux.
Materials and methods In each site, ten Sargassum individuals were collected. Each individual plant was detached from the rock with a knife and enclosed in plastic bags underwater. All samples were later fixed with 4% formaldehyde. In the laboratory, the sample was transferred to a bucket with water and the alga vigorously shaken; the contents was filtered through a series of sieves with the following mesh sizes: 2, 1, 0.5, 0.2 and 0.1 mm. The protocol was repeated 3 times for each sample, to guarantee the complete removal of mobile invertebrates. The material retained in each sieve was preserved in alcohol 70%, and later analysed with a stereoscopical microscope. To calculate faunal densities, the wet weight of Sargassum was considered an estimate of plant size; thus, each individual plant was weighted after removal of excess water by manual centrifugation. Comparisons of the densities of taxonomic groups retained in different sieves were made with 2-factor ANOVAs, considering sampling sites as blocks to remove the spatial variance (Box et al., 1978). A preliminary analysis including the interaction term showed that p > 0.83 for this term in all analyses, except for polychaetes (p = 0.64). Sokal & Rohlf (1995) suggested that if there is no knowledge a priori on the importance of the interaction, the interaction mean square can be pooled with the error mean square when p > 0.75. Comparisons of selected species of gammaridean amphipods were made with a two-factor ANOVA, with sieve size and sampling sites as fixed effects. To achieve homogeneity of variances, densities were transformed to log10 (x+1; after each analysis, residuals were checked graphically to assess the effectiveness of the transformations and to search for possible trends (Box et al., 1978). To verify whether different sieve sizes influenced faunal composition, all gammaridean amphipods were identified to species. Densities were transformed to square-root to reduce the influence of rare species, and species with only one record in all samples were eliminated from the analysis. The similarity between pairs of samples was calculated using Bray-Curtis distance and all samples compared with cluster analysis using UPGMA, for each mesh size considered. The clustering process forces the formation of hierarchical clusters (Clarke, 1993), a property desirable since our replicates were nested within sites.
23
Figure 1. Cumulative contribution (in percentage) of each faunal group in different sieve sizes to the assemblage associated to Sargassum stenophyllum in Tabatinga and L´azaro.
Table 1. Retention efficiency (in percentage) of different sieve sizes for faunal groups associated to Sargassum stenophyllum
Gammaridea Caprelidea Isopoda Gastropoda Polychaeta Total fauna
Sieve size (mm) 0.5 0.2
2
1
11.2 45.6 3.6 47.0 2.3 10.5
24.6 55.5 5.9 68.2 5.1 16.8
85.8 94.5 46.6 93.1 47.0 58.7
99.3 99.7 93.1 99.3 84.1 86.0
0.1 100 100 100 100 100 100
Results A total of 8,959 individuals was collected, 4,912 in Lázaro and 3,547 in Tabatinga. The relative contribution of each taxonomic group to the assemblage differed depending on the mesh size used, and was dominated by gammaridean amphipods in almost all sieve sizes. Large numbers of gastropods and caprellidean amphipods occurred in the larger sieves (1 and 2 mm), while polychaetes and isopods occurred mainly in the smaller ones (Figure 1). Other organisms present in the larger sieves included ophiuroids, mysids and decapods, but in small numbers. In the
smaller sieves, we also recorded a great amount of copepods, members of the meiofauna. Sampling efficiency of each sieve depended on the taxon considered (Table 1), and is here defined as the proportion of individuals retained by a particular mesh relative to the total number of individuals sampled (Schlacher & Wooldridge, 1996). For the amphipods, the 0.5 mm sieve was very efficient, retaining 85.8% of the gammarideans and 94.5% of the caprellideans. On the other hand, isopods and polychaetes presented great differences among sieve sizes, with about 50% of all individuals sampled with the 0.5 mm sieve, and greater proportions recorded as smaller mesh sizes were used. Gastropods presented a pattern similar to that of caprellideans, with 93.1% sampled in the 0.5 mm sieve. When all organisms were pooled, different proportions were retained in each sieve, sampling distinct faunal groups that dominated each size class. These trends were maintained when cumulative densities were considered, but with weaker discrimination of different sieves. There was no difference in gastropod densities relative to sieve sizes, while for all other taxonomic groups the 3 smaller sieves resulted in similar densities (Figure 2). The precision of the estimates – here defined as SE/mean following Andrew & Mapstone (1987) – increased as finer meshes were used. The largest difference on precision among
24
Figure 2. Cumulative mean densities (± SE) of different faunal groups retained in each sieve. Similar superscripts indicate that the means are not different, according to the Tukey HSD test (p > 0.05). The F-tests computed had 4 and 94 degrees of freedom. ∗ = p < 0.05, ∗∗ = p < 0.001.
25 the 0.5, 0.2 and 0.1 mm meshes was 0.03, while the largest difference among the larger meshes (2 and 1 mm) in relation to the 0.1 mm mesh was 0.46. Thus, the three smaller meshes achieved density estimates with similar accuracy and precision. The most abundant gammaridean species in both sites were Hyale media (Dana) (sensu Leite & Wakabara, 1989) and Sunampithoe pelagica (M. E.), followed by Amphilocus neapolitanus Della Valle in Tabatinga and Stenothoe valida Dana in Lázaro (Table 2). Sampling with different mesh sizes was very selective in species retention, with great differences between the 1 and 2 mm sieves relative to the smaller ones, which had similar densities. Young individuals constituted the exception to this pattern (Table 2). In Lázaro, the greatest differences in density estimates between the 1 and 0.5 mm mesh sizes were recorded for S. valida (7.2 and 41.6 individuals 25 g−1 Sargassum, respectively), Jassa falcata (Montagu) (1.7 and 12.6) and Corophium acherusicum Costa (0.5 and 7.8); in Tabatinga, A. neapolitanus (1.5 and 29.5) and Batea catharinensis Muller (1.3 and 9.7) presented the greatest variation. When all sieves were compared, the 1 and 2 mm mesh sizes sampled significantly smaller densities for the most abundant species, H. media (F1,4 = 15.6, p < 0.001), S. pelagica (F1,4 = 10.2, p < 0.001), S. valida (F4,45 = 9.01, p < 0.001) and A. neapolitanus (F4,45 = 47.03, p < 0.001), while densities were similar in the smaller sieves (Tukey’s HSD tests, p > 0.05). Thus, if only the results from the 1 or 2 mm sieves were analysed, the assemblages would be completely dominated by H. media and S. pelagica, while the results from the smaller sieves suggest a more complex relationship in the distribution of species abundances, with greater diversity (Figure 3A, ANOVA: F4,90 = 9.8, p < 0.001). Species richness also increased significantly with the reduction of sieve size, staying constant from beyond the 0.5 mm mesh (Figure 3B, ANOVA: F4,90 = 41.8, p < 0.001). The cluster analysis indicated that the pattern of community structure is very similar when using mesh sizes between 0.1 and 0.5 mm (Figure 4). There is low discrimination between both sites when mesh sizes of 1 and 2 mm are used, while the other three present a similar pattern of site discrimination.
Discussion This study demonstrated that, as with benthic infaunal
Figure 3. Mean values (± SE) of species richness (A) and Shannon’s diversity index (B) for the gammaridean fauna associated to Sargassum stenophyllum in Tabatinga (closed circles) and L´azaro (open circles).
communities, samples processed with different mesh sizes can result in distinct estimates of parameters from the populations and assemblages studied. More precisely, the use of 1 and 2 mm mesh sizes underestimated the abundance of all taxonomic groups associated to Sargassum stenophyllum (Table 1), and retained a maximum of 16.8% of all macrofauna. On the other hand, the 0.5 mm sieve sampled much of the dominant groups, with the exception of isopods (46.6%) and polychaetes (47%). In relation to macrofaunal density, estimates made with sieves of 0.5, 0.2 and 0.1 were similar for all taxonomic groups (Figure 2). Thus, the usage of the 0.5 mm mesh is recommended due to the smaller time needed to process the samples, although the 1 mm mesh is commonly used in the study of macrophyte-associated invertebrate assemblages, as with seagrasses and macroalgae (Nelson, 1979; Gunnill, 1982a,b; Taylor & Cole, 1994). In these studies, several tests of different hypotheses were made through experimental and/or comparative
26 Table 2. Cumulative mean densities (± SE) of gammaridean species retained in several sieve sizes, found in association with Sargassum stenophyllum in L´azaro and Tabatinga
Sieve size (mm):
2
1
L´azaro 0.5
0.2
0.1
Hyale media 22.2 (8.88) 43.3 (21.84) 76.2 (32.57) 79.0 (33.64) 79.0 (33.64) Sunampithoe pelagica 7.5 (2.29) 14.3 (4.19) 33.4 (9.97) 34.0 (9.86) 34.0 (9.86) Cymadusa filosa – 0.2 (0.19) 0.2 (0.19) 0.2 (0.19) 0.2 (0.19) Ampithoe ramondi 0.7 (0.57) 0.7 (0.57) 1.2 (0.67) 1.2 (0.67) 1.6 (0.64) young Ampithoiidae – – 6.5 (2.86) 20.3 (6.78) 21.2 (7.18) Stenothoe valida 2.1 (0.88) 7.2 (3.00) 41.6 (16.72) 63.5 (24.83) 64.7 (25.10) Amphilocus neapolitanus – 0.1 (0.09) 2.1 (1.00) 2.7 (1.03) 2.9 (0.99) Jassa falcata 0.9 (0.74) 1.7 (1.31) 12.0 (6.23) 12.6 (6.23) 12.6 (6.23) Batea catharinensis – – 0.3 (0.26) 0.3 (0.26) 0.3 (0.26) Ericthonius brasiliensis – 0.2 (0.19) 1.1 (0.69) 1.5 (0.71) 1.5 (0.71) Corophium acherusicum – 0.5 (0.31) 7.8 (3.46) 11.6 (4.90) 11.9 (5.23) Gammaropsis sp. – – 0.2 (0.22) 0.2 (0.22) 0.2 (0.22) young Ericthonius 0.1 (0.09) 0.1 (0.09) 0.1 (0.09) 0.1 (0.09) 0.1 (0.09) Lysianassa sp. – – – – – young Elasmopus 0.1 (0.09) 0.2 (0.17) 0.5 (0.26) 0.5 (0.26) 0.5 (0.26) Podocerus brasiliensis – – – – – Podocerus fissipes – 0.2 (0.17) 0.3 (0.26) 0.3 (0.26) 0.3 (0.26)
studies under the assumption that, as all samples were processed with the same sieve size, the patterns recorded were representative of the system studied. However, as a smaller fraction of the community is retained, selectively sampling greater species and individuals, the real patterns of the system could be different. Representatives of the benthic macrofauna have been described as those retained in 0.5 or 1 mm mesh sizes (see discussion in Schwinghamer, 1981). In both cases, comparative studies on the retention efficiency of 1 and 0.5 mm sieves indicate that population estimates are greatly influenced by sieve mesh size, due to the loss of young individuals (Bachelet, 1990). Our results corroborate these observations, since for the 4 most abundant species, significantly greater proportions of individuals were obtained with the 0.5 mm mesh (Table 2). Besides, many of the amphipods retained in the smaller sieves were individuals that were in the marsupium, and dropped in the water during sample processing. Some studies suggest that, for comparisons of different communities in larger scales, the mesh size used would not have a great influence, as both sizes (1 and 0.5 mm) would sample the species composition in a similar way (James et al., 1995; Schlacher & Wooldridge, 1996). Our study showed that significant differences in species richness and diversity can be found between samples processed with 0.5 and 1 mm
2
1
Tabatinga 0.5
0.2
0.1
4.4 (2.61) 15.9 (9.46) 58.4 (22.13) 62.9 (23.16) 63.0 (23.14) 7.6 (2.62) 13.6 (4.24) 28.6 (8.17) 30.1 (9.44) 30.1 (9.44) 1.8 (0.72) 3.5 (1.45) 4.9 (1.76) 4.9 (1.76) 4.9 (1.76) 0.9 (0.45) 1.0 (0.44) 2.9 (1.14) 2.9 (1.14) 2.9 (1.14) – – 2.6 (1.12) 4.3 (1.05) 4.4 (1.04) – – – – – – 1.5 (0.71) 29.5 (8.34) 39.1 (10.82) 40.0 (10.88) – – 0.3 (0.29) 0.3 (0.29) 0.3 (0.29) – 1.3 (0.63) 9.7 (3.76) 9.8 (3.77) 9.8 (3.77) 0.1 (0.13) 0.1 (0.13) 0.1 (0.34) 0.5 (0.34) 0.5 (0.34) – – – 0.1 (0.13) 0.1 (0.13) – – – – – – – – – – 0.6 (0.40) 1.8 (0.82) 4.0 (2.17) 4.0 (2.17) 4.0 (2.17) – 0.3 (0.26) 0.6 (0.44) 0.6 (0.44) 0.6 (0.44) 0.1 (0.13) 0.1 (0.13) 0.1 (0.13) 0.1 (0.13) 0.1 (0.13) – – 0.1 (0.13) 0.1 (0.13) 0.1 (0.13)
sieve sizes, indicating that this is not a general trend. Even if the communities compared were similar, the conclusion arising from the usage of the 1 mm mesh would be that both communities had a smaller species number, with greater dominance patterns; on the other hand, the distribution of species abundances sampled with the 0.5 mm mesh would result in greater diversity (Figure 3). Although the 1 mm mesh would sample a smaller number of species, its usage could allow a greater number of samples to be taken, thus increasing the sampling area and, in this way, the number of species recorded. However, there is no guarantee that the diversity patterns will not change when using the coarser mesh in large scale studies. Also, cluster analysis showed that although a great variation within site is recorded, processing the samples with 1 or 0.5 mm sieves result in different discrimination patterns of both sites (Figure 4). Thus, for phytal communities, estimates on faunal composition and abundance were very accurate and precise when samples were processed with a 0.5 mm sieve. When smaller mesh sizes were used, there was a greater contribution of sediment and meiofaunal organisms to the samples, increasing considerably the processing time; nevertheless, the results were similar to those found with the 0.5 mm mesh (e.g. Schlacher & Wooldridge, 1996). Meiofaunal organisms might eventually be found in this sieve, but commonly occur between 0.063 and 0.5 mm (Gee & Warwick, 1994).
27 Thus, there is an optimal mesh size to sample the macrofauna, including adult individuals and a great part of the young ones, although studies on population dynamics and recruitment of individual species would benefit from the usage of smaller meshes, to sample adequately the juvenile fraction (e.g. Bachelet, 1990). Pilot studies are necessary to sample benthic communities, searching for the optimization of time and resources according to the questions being addressed; besides, critical evaluations of the available methods can contribute significantly to the selection of the most adequate to the objectives of the study.
Acknowledgements Special thanks to A. L. T. Souza for numerous suggestions and help with field work. E.J. de Paula kindly identified the algal species. M. O. T. had a FAPESP fellowship during this work (proc. 97/04340-6). Two anonymous referees greatly improved an earlier version of the manuscript.
References
Figure 4. Cluster analysis using Bray-Curtis distances of square-root transformed gammaridean species densities. Results are for the cummulative fauna retained in sieves of 2 mm (A), 1 mm (B), 0.5 mm (C), 0.2 mm (D) and 0.1 mm (E), sampled in Tabatinga (closed circles) and L´azaro (open circles).
Andrew, N. L. & B. D. Mapstone, 1987. Sampling and the description of spatial pattern in marine ecology. Oceanogr. mar. Biol. A. Rev. 25: 39–90. Bachelet, G., 1990. The choice of a sieving mesh size in the quantitative assessment of marine macrobenthos: a necessary compromise between aims and constraints. Mar. environ. Res. 30: 21–35. Box, G. E. P., W. G. Hunter & J. S. Hunter, 1978. Statistics for experimenters: an introduction to design, data analysis and model building. Wiley & Sons, New York, 653 pp. Clarke, K. R., 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18: 117–143. Duffy, J. E., 1990. Amphipods on seaweeds: partners or pests? Oecologia 83: 267–276. Edgar, G. J. & P. G. Moore, 1986. Macro-algae as habitats for motile macrofauna. Monogr. Biol. 4: 255–277. Edgar, G. J., 1983. The ecology of south-east Tasmanian phytal animal communities. I. Spatial organization on a local scale. J. exp. mar. Biol. Ecol. 70: 181–203. Eston, V. R. & W. O. Bussab, 1990. An experimental analysis of ecological dominance in a rocky subtidal macroalgal community. J. exp. mar. Biol. Ecol. 136: 170–195. Fenwick, G. D., 1976. The effect of wave exposure on the amphipod fauna of the alga Caulerpa brownii. J. exp. mar. Biol. Ecol. 25: 1–18. Gee, J. M. & R. M. Warwick, 1994. Metazoan community structure in relation to the fractal dimensions of marine macroalgae. Mar. Ecol. Prog. Ser. 103: 141–150. Gunnill, F. C., 1982a. Effects of plant size and distribution on the numbers of invertebrate species and individuals inhabiting the brown alga Pelvetia fastigiata. Mar. Biol. 69: 263–280.
28 Gunnill, F. C., 1982b. Macroalgae as habitat patch islands for Scutellidium lamellipes (Copepoda: Harpacticoida) and Ampithoe tea (Amphipoda: Gammaridae). Mar. Biol. 69: 103–116. Hacker, S. D. & R. S. Steneck, 1990. Habitat architecture and the abundance and body-size dependent habitat selection of a phytal amphipod. Ecology 71: 2269–2285. Heck, K. L. Jr & G. S. Wetstone, 1977. Habitat complexity and invertebrate species richness and abundance in tropical seagrass meadows. J. Biogeogr. 4: 135–142. James, R. J., M. P. L. Smith & P. G. Fairweather, 1995. Sieve meshsize and taxonomic resolution needed to describe natural spatial variation of marine macrofauna. Mar. Ecol. Prog. Ser. 118: 187– 198. Kingston, P. F. & M. J. Riddle, 1989. Cost effectiveness of benthic faunal monitoring. Mar. Poll. Bull. 20: 490–496. Leite, F. P. P. & Y. Wakabara, 1989. Aspects of marsupial and postmarsupial development of Hyale media (Dana) 1853 (Hyalidae, Amphipoda). Bull. mar. Sci. 45: 85–97. Martin-Smith, K. M., 1994. Short-term dynamics of tropical macroalgal epifauna: patterns and processes in recolonisation of Sargassum fissifolium. Mar. Ecol. Prog. Ser. 110: 177–185.
Nelson, W. G., 1979. An analysis of structural pattern in an eelgrass (Zostera marina L.) amphipod community. J. exp. mar. Biol. Ecol. 39: 231–264. Schlacher, T. A. & T. H. Wooldridge, 1996. How sieve mesh size affects sample estimates of estuarine benthic macrofauna. J. exp. mar. Biol. Ecol. 201: 159–171. Schwinghamer, P., 1981. Characteristic size distributions of integral benthic communties. Can. J. Fish. aquat. Sci. 38: 1255–1263. Sokal, R. R & F.J. Rohlf, 1995. Biometry. W. H. Freeman, New York. Somerfield, P. J. & K. R. Clarke, 1995. Taxonomic levels, in marine community studies, revised. Mar. Ecol. Prog. Ser. 127: 113–119. Stoner, A. W., 1985. Penicillus capitatus: an algal island for macrocrustaceans. Mar. Ecol. Prog. Ser. 26: 279–287. Taylor, R. B. & R. G. Cole, 1994. Mobile epifauna on subtidal brown seaweeds in northeastern New Zealand. Mar. Ecol. Prog. Ser. 115: 271–282. Virnstein, R. W. & M. C. Curran, 1986. Colonization of artificial seagrass versus time and distance from source. Mar. Ecol. Prog. Ser. 29: 279–288.
