Assessment of long-term temporal changes in the macrobenthic communities south of Peel Island, Moreton Bay, Queensland Peter J.F. Davie Queensland Museum, PO Box 3300, South Brisbane, Queensland, 4101. Email: [email protected] Ian W. Brown Agri-Science Queensland, EcoSciences Centre, 41 Boggo Rd, QLD 4102 David G. Mayer Agri-Science Queensland, EcoSciences Centre, 41 Boggo Rd, QLD 4102 Citation: Davie, P.J.F., Brown, I.W. & Mayer, D.G. 2010 12 30. Assessment of long-term temporal changes in the macrobenthic communities south of Peel Island, Moreton Bay, Queensland. In, Davie, P.J.F. & Phillips, J.A. (Eds), Proceedings of the Thirteenth International Marine Biological Workshop, the Marine Fauna and Flora of Moreton Bay, Queensland. Memoirs of the Queensland Museum — Nature 54(3): 401-435. Brisbane. ISSN 0079-8835 ABSTRACT A grab sampling survey of 15 sites in a relatively small geographic area to the south of Peel Island, Moreton Bay, Queensland, was undertaken in February 2005. This broader area is already known as a biodiversity hot-spot within the Bay. The sampling was designed to replicate a longer-term survey undertaken 35 years earlier between March 1970 and December 1971. The new study was intended to assess changes to species composition of those earlier communities after so many years, and provide a yardstick on the present ecological health of the system. The sediments, and hydrographic features such as depth and currents appear not to have changed significantly. There have however been some minor changes in site groupings based on species presence, and a marked change in the species characterising the site groupings. In particular, in 2005 there was an absence of benthic tunicates that had been an important component at some sites in the earlier sun/ey; and secondly, there has been the development of significant Trichomya mussel aggregations that had not been noted from this area in the past. An analysis of community trophic structure found essentially the same site classification as the simple species x sites analysis, and as found in other studies, deposit feeders predominate in the muddiest site-groups. Overall, the species richness was very high (564 species), and this was greater than the 394 species found earlier. It is believed the earlier survey had under-estimated the number of species present. There is every indication that the present communities in this area are healthy and resilient. (cid:9633) ecology, marine, macrobenthos, Moreton Bay, communities, sediments, trophic structure, biodiversity. Around the world, human impacts on estuarine from an increasingly complex and unpredictable and coastal environments have been dramatic arsenal of chemicals, and wholesale marine since the advent of the industrial revolution (Lotze resource over-exploitation. Moreton Bay (Fig. 1), et cil. 2006), but arguably have accelerated is a sheltered, coastal embayment, with one of significantly over the last 50 years. This has Australia's largest cities on its foreshores. Over been a time of massive human population the last 30 years, the Brisbane-Gold Coast explosion with concomitant encroachment and corridor has become one of the fastest growing destruction of coastal environments, pollution human population centres in the developed Memoirs of the Queensland Museum — Nature • 2010 • 54(3) • www.qm.qld.gov.au 401 Davie, Brown & Mayer FIG. 1. Map of Moreton Bay showing the sampling area (dark gray square) just to the south of Peel Island, in the southern part of the Bay. world, and thus has a major impact potential on have been made of the soft bottom macro¬ the Moreton Bay region. benthos of Moreton Bay (e.g. Raphael 1974; Stephenson et al 1970, 1974, 1976, 1977, 1978; Moreton Bay is .protected from the South Poiner 1979a, b; Stephenson & Cook 1977,1979; Pacific Ocean by the large sand islands of North Stephenson 1980a, b, c; Stephenson & Sadacharan and South Stradbroke, Moreton and Bribie. To 1983; Lorz & Bamber 2010, this volume). Trawl the north and east it is primarily oceanic while studies were conducted by Jones (1973), Stephen¬ to the south and west the Bay becomes a son & Dredge (1976), Quinn (1979,1980), Burgess complex estuarine system, with numerous islands (1980), Stephenson & Burgess (1980), and Stephen¬ and muddy banks in its southern portion. son et al. (1982a, b). Dredged macrobenthos Interestingly it lies at the subtropical/temperate near the mouth of the Brisbane River was biogeographic transition zone (see Davie & reported on by Hailstone (1972, 1976), Boesch Hooper 1998), and has an extraordinary mix of (1975) and Park (1979). Campbell et al. (1974) southern Australian endemic species and wide¬ studied nine estuaries in southeastern Queens¬ spread tropical Indo-West Pacific and Great land, most of which feed into Moreton Bay, and Barrier Reef species. Campbell et al. (1977), Stephenson & Campbell Our understanding of Moreton Bay's faunal (1977) and Davie (1986), reported on the sublittoral composition and ecological processes have macrobenthic fauna of Serpentine and Jackson's grown considerably over the last 30-40 years Creeks. Young & Wadley (1979) also examined (see review by Skilleter 1998). Extensive studies the distribution of shallow water epibenthic 402 Memoirs of the Queensland Museum - Nature • 2010 • 54(3) Macrobenthic communities south of Peel Island macrofauna in the Bay. Most recently Stevens & whole Bay, most likely because of its well Connolly (2005) mapped and classified developed coral reefs and a mix of consolidated macrobenthic habitat types by using a compact hard and muddy-sand bottoms. video array at 78 sites spaced 5 km apart. They The study reported here was undertaken recognised nine habitats, with only one being during the Thirteenth International Marine on hard substrate. These included previously Biological Workshop, held in Moreton Bay. The unreported deep-water algal and soft-coral intention was to repeat the earlier survey of reefs, and new areas of seagrasses. Broader Stephenson et al (1974) of 15 sites in a relatively ecological work on understanding nutrient cycling small geographic area to the south of Peel and the impact of sewage on the western Bay Island. As already noted this region has been was also undertaken during the 1990s, and this identified as one of the biodiversity hot-spots in has also contributed significantly to our under¬ the bay. The work of Stephenson et al (1974) standing of the ecological dynamics of the Bay involved 8 sampling times over three-monthly and the Brisbane River (Dennison & Abal 1999). intervals for two years from March 1970 to Conservation of biodiversity is a major priority December 1971, thus also providing them with for the continued healthy functioning of com¬ patterns of seasonal change. While the new munities, but marine biodiversity issues have sampling was a once-off snapshot of the area 35 not received the attention currently given to years later, we hoped that this would provide terrestrial systems, perhaps because they are some interesting insights into the state and less easily studied, impacts are less conspicuous, composition of these communities after so many and taxonomic difficulties are immense. Davie years, and provide an indication of the system's & Hooper (1998) examined the species richness current ecological health. and distributional patterns of the fauna inside Moreton Bay and identified two major biodiverse MATERIALS AND METHODS regions — an inshore estuarine-dominated region, and an eastern marine-dominated region. This POSITIONS AND DEPTHS OF SITES latter region, including the northern end of The present work was carried out towards Stradbroke Island, and Peel, Bird and Goat the northern end of the southern half of Islands had the highest species richness in the Moreton Bay immediately south of Peel Island. Table 1. Station details for grab samples taken south of Peel Island, southern Moreton Bay, February 2005. Salinities were not taken consistently, but were around 28%o at all sites during the period of sampling. Station No. Latitude Longitude Depth Date 1 27°31.25' 153°22.00' 6.5 m 17.02.2005 2 27°31.25' 153°21.85' 6.6 m 17.02.2005 3 27°31.25/ 153°21.65' 6.4 m 18.02.2005 4 27°31.53' 153°21.44' 5.9 m 20.02.2005 5 27°31.53' 153°21.70' 6.5 m 20.02.2005 6 27°31.55' 153°20.80' 8.5 m 18.02.2005 7 27°31.48' 153°20.72' 9.0 m 18.02.2005 8 27°31.48/ 153°20.48' 9.2 m 18.02.2005 9 27°31.6r 153°20.38' 7.6 m 20.02.2005 10 27°31.68' 153°20.54' 8.4 m 18.02.2005 11 27°32.39# 153°20.80' 4.2 m 17.02.2005 12 27°32.20' 153°20.75' 3.9 m 20.02.2005 13 27°31.98' 153°20.62' 7.0 m 20.02.2005 14 27°32.29' 153°20.42' 5.1m 20.02.2005 15 27°32.61' 153°20.42' 4.4 m 17.02.2005 Memoirs of the Queensland Museum — Nature • 2010 • 54(3) 403 Davie, Brown & Mayer FIG. 2. Positions and groupings of the sample sites. Specific coordinates are given in Table 1. The explanation for the site-groupings is explained later in the text. In a landmark dredge study of the macro¬ ous fixed points (data filed in archives of Queens¬ benthos of Moreton Bay, Stephenson et al (1970) land Museum)', we were unable to locate these established this area as having a rich fauna and data in the Queensland Museum, and instead a small scale patterning of 'communities'. Their we interpolated the positions using modern work was subsequently followed-up (Stephenson mapping software. The positions of the sites for et al. 1974), by an intensive grab-sampling the 2005 sampling are shown in Fig. 2, and the study where fifteen stations were sampled in coordinates are given in Table 1. All sites are quintuplicate in each of four seasons for two enclosed within an area of approx. 3 km2. years (from March 1970 to December 1971). The Topographic grouping of sites can be sum¬ objective of that work was to investigate a marised as: 'Goat Island slope' (Sites 1-5); 'North¬ suspected complex benthic biota and to attempt west gutter' (6-10,13); 'Southern shallows' (11, to resolve the complexity into a number of 12,14,15). spatial and temporal patterns. Depths of sampling sites ranged from 3.9-9.2 Stephenson et al. (1974) sampled five sites in m at the time of sampling (Table 1), reflecting each of three areas to reflect contrasting bottom the depths of 2.4-93 m given by Stephenson et topographies as revealed in published charts al. (1974). The apparent slightly greater depth at available at the time. While they stated in their our shallowest sites may be due to factors such paper that the site 'positions have been estab¬ as sampling on a higher tide, or differences in lished by horizontal sextant angles of conspicu¬ specific site location. Overall there appears to 404 Memoirs of the Queensland Museum — Nature • 2010 • 54(3) Macrobenthic communities south of Peel Island have been no significant movement of banks, or then also assessed against sediment structure to changes in depths, since the sampling at the see if any obvious patterns emerged in community beginning of the 1970s. structure. All samples are deposited in the collections of SAMPLING the Queensland Museum, Brisbane. Quintuplicate samples were collected at each site using a long-arm van Veen grab with a SEDIMENTS surface sampling area of 0.1 m2 (total sampling At each site a separate grab sample was taken, area = 0.5 m2/site). Faunal samples were and subsequently a 200-250 gm subsample was washed on board the vessel through a series of taken in the lab for sediment particle size graded sieves down to a 0.5 mm mesh, and the analysis. Sediments were washed through a series contents of each sieve were washed into a large of graded sieves (2.0,1.0, 0.5,0.250, 0.125, 0.063 plastic bag and preserved with 4% fomalin. In mm). Similar sieves were used by Stephenson et the laboratory the faunal samples were again al (1974) (viz. 1.98,1.02, 0.53, 0.211, 0.15, 0.099 washed, transferred to 70% ethanol, and stained mm) to grade the series of retained fractions as with Rose Bengal. The fauna was removed from gravel, very coarse sand, coarse sand, medium the samples using elutriation, and by hand¬ sand, fine sand and very fine sand respectively, picking using forceps under a dissecting micro¬ and the non-retained fraction as mud. The scope. Despite the 0.5 mm fraction being retained, retained portion of each sieve was air dried, sorting was only undertaken to the 1 mm stage baked in a microwave oven to remove all due to time and labour constraints, and because moisture, and weighed; the remaining sediment this was sufficient to provide a valid compari¬ was washed into a coffee-filter paper, similarly son with the original sampling regime of dried and weighed (minus the weight of the Stephenson et al (1974). The initial sort was to filter paper). Weights were converted to percent¬ major taxa, followed by more precisely splitting ages so that the composition of each sample group by group into recognisable OTUs could be compared. Results are presented in (operational taxonomic units). Identification Table 2. The general results are not dissimilar to was undertaken to the lowest taxonomic rank those of Stephenson et al. (1974). Figure 7 possible depending on available expertise. clearly shows the significant differences in Unfortunately it was logistically impossible to sediment composition between the derived site- check identifications against the original groups. These are discussed further in relation reference collection of Stephenson et al (1974), to faunal trophic patterns later in the paper. so there is not necessarily concordance in nomen¬ Stephenson et al (1974) considered 'mud' to be clature between that and the present study. In the portion not retained by their finest sieve (99 addition, there have been significant changes in pm) — in the present study our finest sieve was nomenclature in many groups over the last 35 63 pm and thus the results are not directly years, and we have not tried to track these when comparable. Their 'mud' component could be comparing the two data sets. It is nevertheless expected to be relatively higher than our result, interesting, as will be further discussed, that as less was retained by their finest sieve. This is there has been a very clear and real shift in highlighted by Site-Group II (Sites 2-5) where characterising species at many sites. they found a >50% mud component (< 99 pm), whereas our result was 34% for <63 pm, but Individual species counts from the replicate samples were lumped for further analyses in 87% at 63 pm. Thus the sites we sampled were similarly very muddy, but not directly compar¬ order to minimise the effects of micro-patchiness able with the terminology of Stephenson et al between samples, and to get the best possible (1974), at least at the very finest particle size. reflection of the community composition at each site. All species were also assessed for their Stephenson et al (1974) found some signifi¬ trophic status, and assigned to one of five cant changes in sediment structure at a number categories, viz., 1, Suspension feeder; 2, Deposit of sites between sampling times from March feeder; 3, Grazer; 4, Predator/Scavenger; or 5, 1970 and December 1971. In particular important Parasite. The trophic structure of each site was changes were: less mud at Stn 5; less fine sand Memoirs of the Queensland Museum — Nature • 2010 • 54(3) 405 Davie, Brown & Mayer Table 2. Sediment composition at each site as retained by graded sieves, and particle size expressed as a percentage of each sample. Stn No./ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Grain Size 2 mm 0.8 2.5 4.6 0.9 0.7 9.3 10.2 20.4 16.7 8.6 4.3 7.1 4.2 9.0 12.7 1 mm 0.7 1.3 2.3 0.5 0.6 9.7 0.7 8.3 7.9 6.7 2.9 2.0 3.2 8.1 6.4 500 pm 0.7 1.0 1.7 0.5 0.6 16.2 15.2 10.0 11.5 5.4 2.8 2.2 2.4 6.8 8.4 250 pm 18.5 3.8 1.4 0.7 0.7 24.6 26.3 14.6 29.3 11.9 34.8 38.7 5.7 36.3 41.9 125 pm 30.6 6.7 5.2 33.8 6.8 23.6 21.0 13.8 24.1 55.3 48.0 42.8 55.0 33.5 21.0 63 pm 18.9 50.2 50.8 47.0 56.2 4.2 6.1 5.5 2.2 7.7 2.9 2.9 20.7 3.0 5.6 <63 pm 29.9 34.5 34.0 16.5 34.4 12.6 20.5 27.4 8.2 4.4 4.3 4.3 8.8 3.4 4.0 and more mud at Stn 12; more very coarse and a seasonal pattern with winds from north to medium sand and less mud at Shis 13 and 14; east predominating between December and and less coarse, medium and fine sand and April, and from south to southwest between more very fine sand and mud at Stn 15 (11-15). May and August (Newell 1971). The longest Most of these changes over time occurred in the wind-fetches in the sampling area are in an arc group of 'Southern Shallows' sites. The most from south to southeast, and wave-action is aberrant sites in terms of sediments were 1, 4 severest when these winds blow against a and 13 (very similar to the present results), and flooding tide from the north. This, and the they were able to further subdivide sites on the shallow depths, makes the most southerly sites basis of percentages of coarse sand, with sites 6, 11,12,14 and 15, the most wave-affected. 7 & 8 having > 60%, and 9,10 & 13 < 60% (again TROPHIC STRUCTURE very close to the site-groupings based on The relationship between feeding type and sediment obtained in the present work). sediment characteristics has been well WATER MOVEMENTS documented and explored for many years (e.g. Tidal currents flow south around Peel Island see Gray, 1974; Rhoads 1974; Lopez & Levinton from the north-east and north-west, leaving a 1987). The classification of benthic invertebrates 'slacker-water' area which includes sites 1-5 into infaunal trophic feeding groups can be (see Patterson & Witt 1992: fig. 11). Conversely, quite complex, and has even been developed the more westerly area that includes sites 6-13 into a numerical index for ecological mapping has strong tidal currents that may reach 2-3 (Word 1980). knots (at the surface at least), during spring We have adopted a basic system consisting of tides. The more easterly sites (1-5) are also more five categories: 1, suspension feeder; 2, deposit influenced by clean oceanic water flowing from feeder; 3, grazer; 4, predator/scavenger; 5, the north-east through the Rainbow Passage parasite. This is similar to that used in some (Stephenson etal. 1974). The other more westerly recent Australian studies that have explored and southerly sites (6-15) are more influenced the trophic relationships of both tropical (Long by tidal flows from the western and central & Poiner 1994), and temperate macrobenthic portions of the Bay and are thus under greater infaunal communities (Poore & Rainer 1974; terrestrial influence, particularly dilution by Wilson et al. 1993). A 'parasite' category was flood waters (consequently higher turbidity and included initially because of the high incidence greater possibilities of pollution), and greater of a bopyrid isopod in the gill chamber of a temperature variation. The main influence of common porcellanid, Pisidia dispar, however it floods in the area is from the Logan-Albert was removed from the graphical presentation Rivers which discharge from the south. Overall, of the final analyses as no other similar para¬ the whole study area is generally relatively sites were identified, and it merely mirrored the protected from the open Bay to the north, and presence of its host so did not contribute to any waves are due to local winds. There is typically better understanding of community structure. 406 Memoirs of the Queensland Museum - Nature • 2010 • 54(3) Macrobenthic communities south of Peel Island The predators and scavengers were grouped multidimensional scaling (MDS) representation as in this context the definitions are ambiguous, for each data set. The degree of association and many such species commonly switch roles between the two between-sites similarity matrices according to available resources. (namely, using the sediment and the species- counts data respectively) was estimated using Suspension feeding organisms (= filter feeders) the Mantel test of the product-moment either actively pump suspended particles and correlation. Canonical correspondence analysis organic matter through a filtration apparatus, (CCA) was used to derive and interpret the or use complex feeding appendages including inter-relationships between these two sets of mucous nets, to separate such matter from the water column with the aid of bottom currents. variables. Bivalve molluscs, tunicates and bryozoans are RESULTS important suspension feeders, as are some crustaceans and polychaetes. The organic matter In total, a remarkable 564 species were is typically living or dead phyto- and zooplankton collected (see Appendix 1) consisting of 11,892 and bacteria, resuspended benthic particles, individuals (average of 793 individuals/m2). Of and dissolved organic matter. these a large number (150) occurred only once, and only 264 species occurred at least 5 times. Deposit feeders may be either mobile or The fauna was relatively evenly represented sedentary, and feed at or near the surface, or across the three major taxa: Mollusca — 181 burrow to some depth. They feed on living or species (32.1%) of which the largest number dead organic content, often including degraded was bivalves (108 spp.); Crustacea — 160 spp. plant material, and typically ingest sediment (28.4%); and the Annelida — 180 spp. (31.9%). with its attached interstitial meiofauna and The Echinodermata were represented by rela¬ microflora. Deposit feeders typically process at tively few species (16; 2.8%), though a couple of least one body weight in sediment daily (Lopez species played a major characterising role in & Levinton 1987), and considerable amounts of some site communities. sediment are processed in this way. Many species switch between deposit- and suspension-feeding Data Reduction. The large number of species modes (Lopez & Levinton 1987). Switching is recorded meant that a meaningful analysis of often influenced by local environmental vari¬ species correlations required some significant ables (current flow, concentration of suspended data reduction. Initial analyses were restricted particles). Interactions between deposit- and to the 264 species which occurred 5 times or suspension-feeding animals influence nutrient more. The biplot from the full CCA still showed cycling and community structure (Wilson et al. too many species vectors to be interpretable, so 1993). this was re-run with only the 28 most abundant species. These 28 totalled 65% of all captures, Trophic status of each species was obtained and all other species represented less than 0.5% from Poore & Rainer (1974), Fauchald & Jumars each. It was felt that this was reasonable to reflect (1979), Brusca & Brusca (1990) and Todd (2001). the key species defining the communities STATISTICAL METHODS Between-sites similarity matrices, using the SITES CLASSIFICATION Bray-Curtis index on the untransformed data, The site groupings recognised here are the were formed separately for the sediment, result of concordance between three separate species-counts, and trophic structure data sets. site classifications based on: sediment structure Bray-Curtis was used because it does not derive vs sites; species presence and abundance vs similarity from conjoint absences (Clark & sites; and trophic classes vs sites. The combined Warwick 1994), and has been shown to be a sites x sediments x species data were then robust index across both raw and standardised shown in relationship to each other using a data (Faith et al. 1987). These similarities formed canonical correspondence analysis (CCA) biplot. the basis of group-average hierarchical clustering Dendrograms of sediments and species site to produce dendrograms. Dissimilarities, calcu¬ groupings (not presented here) largely revealed lated as one minus similarity, were used for a the same groupings as the MDS analyses. Memoirs of the Queensland Museum — Nature • 2010 • 54(3) 407 Davie, Brown & Mayer FIG. 3. MDS of sites x sediment data indicating 6 site- groupings (I—VI) based on relative proportions of sediment grades. as Site-Group IVa and IVb respectively. although a few sites that appear closer to each other in the MDS plots, moved into neighbouring clades in the dendrograms. Overall we considered the dendrograms did not give as conceptually a satisfying result probably because of the nature of the clustering algorithm, and that the MDS analyses gave a more visually understandable result in a two dimensional framework. Under multidimensional scaling, two dimen¬ sions adequately represented these data, with stresses of much less than 0.2. These stress values were 0.042 for the sediments data, 0.021 for the trophic-level counts and 0.095 for the species-counts. These two-dimensional multi¬ dimensional patterns of sites are shown in Figs 2, 3 and 5 respectively. The Mantel test showed a significant (P<0.01) association between the similarity matrices, with a correlation of 0.51. The resultant biplot is shown in Fig. 6. Sediments x Sites classification. The MDS of sites by sediment data (Fig. 3) suggests 6 site- discussion of the trophic analysis. Noteworthy groupings (I—VI) based on relative proportions is that Sites 6, 7 & 9 clustered most closely to¬ of sediment grades. As found by Stephenson et gether according to sediment composition with ni (1974), sites 1 (Site-Group I) and 4 (Site-Group Site 8 being the outlier within the group. However, III) differ significantly from all other individual according to spp x sites and trophic structure x sites. The specific differences in sediment sites groupings, sites 6 & 7 are distinctly structure are explored in more detail under the separated from 8 & 9, although relatively closely 408 Memoirs of the Queensland Museum - Nature • 2010 • 54(3) Macrobenthic communities south of Peel Island FIG. 6. CCA Biplot showing 3-parameter representation of species x sites x sediment characteristics. allied. Because sites 6-9 are situated close (Fig. 5) essentially gives the same groupings as together geographically, and thus form a logical the species classification, and in particular sites site-grouping, we have decided to treat these 6,7,8,9,10 & 13 are clustered in close proximity four sites as a single site-group, but use a (this is further discussed under a separate subgrouping notation to indicate that there are heading later). differences in species composition. CCA Biplot. This plot (Fig. 6) provides an Species x Sites classification. The spp. x sites informative 3-parameter visual representation classification (Fig. 4) shows seven discrete of the species x sites x sediment characteristics. site-groups, though as already mentioned, sites In general, the closer the arrow to the centre the 6 & 7 and 8 & 9 are treated as Site-Group IVa more evenly distibuted the values, such that and IVb respectively. Comparison with the map inner cluster of species are the most wides¬ of sites (Fig. 2) shows that the the site-groups all pread across all sites. Some strong trends in the include sites that are clustered close to each data are apparent. Site 1 (Site-Group I) is strongly other topographically. characterised by an increased proportion of 125 Trophic Classes vs Sites classification. The jim sediments and the marked presence of 'sp. analysis of trophic class x sites classification 2' (Mesochaetopterus minutus). Sites 2, 3 and 5 Memoirs of the Queensland Museum — Nature • 2010 • 54(3) 409 Davie, Brown & Mayer (Site-Group II) have the greatest proportion of agreed with that of Stephenson et al. (1974) 63 |Lim sediments and are most strongly charact¬ who, although they included it in their Site- erised by species 22 and 14 (Whiteleggia stephensoni Group I (with sites 2-5), remarked that it was and maldanid sp. 3 respectively). Species 4 aberrant. While it was a relatively muddy/fine (Maldane sp.) and 9 (Golfingia trichocephala) are sand site (49% 63 pm), it was not as obviously characteristic of both Site-Groups I and II. Site 4 muddy as the adjacent sites otherwise included (Site-Group III) is characterised by primarily 63 in Site-Group II (see Fig. 7). In total 95 spp. were Mm and 125 Mm sediments but does not have an present consisting of 2179 individuals (4358 nr2), obvious species characterisation. Sp. 16 (spionid making it the most densely populated of all the sp. 4) lies close to Site 4 on the plot, but this is an sites. There were 16 characterising species (>10 artefact of its dual presence at sites 2, 3, 5 individuals), but it was remarkable for very (Site-Group II) and at the widely separated sites high densities of four species: Mesochaetopterus 6 and 7 (Site-Group IVa). Sites 6-9 (Site-Group minutus, Golfingia trichocephala, Maldane sp. and IVa, b) are clearly characterised by much sandier Ophiura kinbergi. In particular it was markedly sediment grades, and characterised by spp. 23 different from all other sites by having a very (Pharyngeovalvata sp.), 8 (Terebellidcs narribri), 1 large number of the tubicolous, suspension ('Trichomya hirsuta) and 17 (Paraoroidcs sp. 1). feeding polychaete Mesochaetopterus minutus These associations are further discussed in the (1986 nr2), which occurred at no other site. The following Site-Group accounts. other most abundant species, the sipunculid SITE-GROUP CHARACTERISING SPECIES Golfingia trichocephala, characterised this site and The species x sites classification largely agreed Site-Group II, but nowhere else. The deposit with the sediments x sites classification in feeding polychaete Maldane sp. also occurred at supporting the recognition of six major site- Site-Groups II, III, and V, but was three times groups, however there was some disparity more abundant at Site 1 than at the sites between the two pairs of adjacent sites within comprising Site-Group II, and at least six times Site-Group 4, and this was resolved by erecting or more abundant than at sites III and V. The subgroups IVa and IVb for the purposes of predatory Ophiura kinbergi was the other major understanding the dominant species that charact¬ component of the dominant fauna, and is pre¬ erise these communities. sumably responding to the number of prey In order to limit discussion of characterising species present. species for each site-group, we have arbitrarily Twelve species were present in relatively assigned a cut-off of 10 or more individuals lower but consistent numbers. Of these, being present at least one site within the site- Protankyra sp., capitellid sp. 5 and polynoid sp. group. This count represents the sum of quin- 4 are unique to this site; Amphiuridae sp. 1, tuplicate 0.1 m2 grab samples, so in effect, we Ophiuridae sp. 2, Sternaspis scutata, Nematonereis only further consider species composition at unicornis and maldanid sp. 3 are characteristic any given site if they occurred in a density of only this site and the adjacent Site-Group II, greater than 2 individuals per grab sample. VVe while Ophiothrix sp 1 is shared only with feel intuitively that at densities lower than this Site-Group V. Overall it is clear this site shares any given species will be a minor component of it closest affinities with Site-Group II, but the community at that site-group. having four uniquely characteristic species, Where a species is discussed as uniquely including the very abundant Mesochaetopterus characterising a site-group, this is based on the minutus, sets it clearly apart faunistically. This reduced data set. It is possible that such a species species may have a patchy presence, or may be may occasionally occur in other site-groups, seasonal in occurrence, and without it, this site but in sufficiently low numbers as to not make a would be far more faunistically similar to Site- significant contribution to site-group classification. Group II. However there were also significant Site-Group I. This site 'group' is composed sediment differences, and this may, in the end, only of Site 1. All analyses (see Figs 2, 3, 5 & 6) be the determining factor in the observed indicated that Site 1 was unique. This result faunistic differences. 410 Memoirs of the Queensland Museum - Nature • 2010 • 54(3)