J. Mar. Biol. Ass. U.K. (2004), 84, 1117^1130 Printed in the United Kingdom
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures N.A. Knott*P, A.J. Underwood, M.G. Chapman and T.M. GlasbyO Centre for Research on Ecological Impacts of Coastal Cities, Marine Ecology Laboratories A11, University of Sydney, NSW 2006, Australia. *Present address: Department of Zoology, University of Melbourne, Victoria, 3010, Australia. OPresent address: NSW Department of Primary Industries, Port Stephens Fisheries Centre, Private Bag 1, Nelson Bay, NSW 2315, Australia. PCorresponding author, e-mail:
[email protected]
Subtidal assemblages of epibiota on vertical and on horizontal surfaces of two natural reefs and two concrete breakwalls were sampled photographically during autumn and winter of 1998. Di¡erences in the assemblages on the two types of substrata (natural reefs and concrete breakwalls) were detected between assemblages on horizontal surfaces, but not on vertical surfaces. The covers of several individual taxa (e.g. Herdmania momus, serpulid polychaetes, coralline encrusting algae) and number of sponge taxa showed clear di¡erences between the two types of substrata. There were great di¡erences between the assemblages on vertical and horizontal surfaces on each natural reef and arti¢cial structure. Invertebrates consistently covered a larger area on vertical than on horizontal surfaces with sponges (as a group) and the ascidian Herdmania momus, the dominant invertebrates on these reefs, clearly showing this pattern. Nevertheless, this pattern was complex for sponges because several species covered a larger area on horizontal than on vertical surfaces and there was no di¡erence in the number of taxa of sponges between the two orientations on natural reefs. Algae, contrary to the results of previous studies, did not show any consistent di¡erences in their covers on vertical or on horizontal surfaces. The results of this study indicated that orientation may be of greater in£uence on the biological diversity of epibiota on subtidal reefs than whether reefs are natural or arti¢cial.
INTRODUCTION Urbanization and industrialization of coastal areas has meant that arti¢cial structures have become common and extensive features of many marine environments (Walker, 1988; Glasby & Connell, 1999), sometimes replacing natural rocky reefs. The surfaces of these arti¢cial structures are rapidly colonized by invertebrates and algae (epibiota), and thereby become new habitats for these organisms. Evaluating whether di¡erences exist between assemblages of epibiota on arti¢cial structures and those on natural reefs is important in order to evaluate the potential ecological e¡ects of introducing arti¢cial structures into the marine environment (Glasby & Connell, 1999). Few studies have, however, investigated whether di¡erences exist between subtidal assemblages on arti¢cial structures and on natural reefs (Connell & Glasby, 1999; Gabriele et al., 1999; Glasby, 1999a). In sheltered estuaries, subtidal assemblages on pier pilings and pontoons di¡er from those on natural reefs (Connell & Glasby, 1999; Glasby, 1999a). In the Adriatic Sea, Gabriele et al. (1999) also found that subtidal assemblages of epibiota on wave-exposed natural reefs di¡ered from those on arti¢cial structures. A potential explanation for the di¡erences observed by Gabriele et al. (1999) was, however, orientation because the surfaces on natural reefs tended to be horizontal while those on arti¢cial structures tended to be vertical. The orientation of surfaces on subtidal reefs appears to be a very important feature of these habitats (Hartnoll, Journal of the Marine Biological Association of the United Kingdom (2004)
1983; Sebens, 1985, 1991). A general pattern appears to be that invertebrates cover a larger area on vertical than on horizontal surfaces, while algae show the opposite pattern (Witman & Sebens, 1991; Baynes, 1999). Numerous observations have been made about such di¡erences in subtidal assemblages of epibiota on vertical and on horizontal surfaces (Logan et al., 1984; Wendt et al., 1989; De Kluijver, 1991; Ginn et al., 2000), but few studies have evaluated these patterns clearly and quantitatively. Furthermore, no studies have been done on subtidal reefs in the southern hemisphere to compare assemblages on vertical and on horizontal surfaces. Only with quanti¢ed patterns can logical hypotheses about possible explanations be formulated and tested (Andrew & Mapstone, 1987; Underwood et al., 2000). Sponges and ascidians are often the dominant invertebrates in terms of cover on subtidal reefs in temperate regions (Roberts & Davis, 1996; Underwood & Chapman, 1996; Roberts et al., 1998; Bell & Barnes, 2000). Despite this, the ecology of sponges and ascidians on waveexposed reefs has been studied relatively little (Keough, 1999; Roberts, 1999; Ginn et al., 2000). This is often because the taxonomy of many species is not known and few people know how to identify them reliably (Roberts & Davis, 1996). This is especially true for Australian species (Roberts & Davis, 1996; Keough, 1999; Roberts, 1999). Sponges and ascidians were the main focus of this study, although other epibiotic species were also sampled. Concrete breakwalls are conspicuous arti¢cial structures on wave-exposed coasts. They are often extremely
1118
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
large, stretching for kilometres, and they are structurally similar to natural reefs, having both vertical and horizontal surfaces. These structures were sampled along with natural reefs to test two sets of predictions. First, that assemblages of epibiota on vertical and on horizontal surfaces on concrete breakwalls would be di¡erent from those on natural reefs. Second, it was predicted that assemblages on vertical surfaces would be di¡erent from those on horizontal surfaces, speci¢cally because invertebrates would cover a larger area on vertical than on horizontal surfaces and because algae would show the opposite pattern.
MATERIALS AND METHODS Locations and sampling methods
Wave-exposed subtidal reefs were sampled photographically at four locations in the Sydney^ Illawarra region (Figure 1): two natural rocky reefs, Bare Island (autumn, 1998) and Dobroyd Head (winter, 1998) and two concrete breakwalls, Port Kembla (autumn, 1998) and Port Botany (winter, 1998). Each surface on each natural reef and on each breakwall was sampled with ¢ve haphazardlyplaced photo-quadrats. Each photo-quadrat was 3857 cm and was taken with a Nikonos V underwater camera with a 28 mm lens and strobe attached to a frame. The taxa on each surface were recorded in the ¢eld to aid identi¢cation when sampling photographs in the laboratory. Breakwalls were constructed of large concrete blocks (approximately 1.81.81.8 m at Port Botany and 1.51.51.8 m at Port Kembla). The tops (horizontal surfaces) and the sides (vertical surfaces) of these blocks were sampled. Natural rocky reefs did not have discrete surfaces like those on the blocks of the breakwalls. Instead, areas of approximately 1.81.8 m on the reefs were selected haphazardly and sampled. On each natural reef and breakwall, ten vertical and ten horizontal surfaces were sampled. Horizontal surfaces and vertical surfaces were interspersed by alternating between sampling a vertical surface and then a horizontal surface along each reef. There were at least ¢ve metres between each of the surfaces. The depths of the surfaces were also interspersed and were between four and eight metres below mean low water springs. Vertical surfaces were also selected haphazardly in relation to the direction they faced (i.e. perpendicular, parallel or away from the direction of the wave-action). Percentage covers of epibiota were estimated by projecting the photograph of each quadrat onto a screen at a 1 to 1 ratio and identifying the taxon under each of 100 regularly spaced points superimposed over the entire quadrat. Taxa which were in the quadrat but not under a point were recorded as having a nominal cover of 0.5%. Only taxa that were directly in contact with the substratum (i.e. primary cover) were sampled. Data from the ¢ve photo-quadrats per surface were averaged to increase the precision of the estimates of the cover of epibiota and because often one continuous sheet of one species of sponge would cover more than one photo-quadrat. Invertebrate taxa were identi¢ed to the lowest taxonomic level practical. The taxonomy of sponges is di⁄cult Journal of the Marine Biological Association of the United Kingdom (2004)
and there are few guides on species on the Australian coast. Some of the sponge taxa are, therefore, identi¢ed as morpho-species (Oliver & Beattie, 1996). Invertebrates that could not be identi¢ed to species from photographs were sampled at higher taxonomic levels e.g. bryozoans. Algae were sampled as functional groups (i.e. ¢lamentous and foliose algae) or taxonomic groups (i.e. Peyssonnelia sp.). All species, morpho-species, functional and taxonomic groups were used in the multivariate analyses of assemblages of epibiota. Univariate analyses were done on taxa that occurred on more than ¢ve per cent of surfaces. Numbers of taxa were only compared for sponges and ascidians because these taxa were the focus of the study and because it was di⁄cult to identify many of the other taxa to species or morpho-species from photographs. Silt matrix, sediment and bare space were also sampled. The silt matrix was a combination of ¢lamentous algae and sediment (Roberts et al., 1994). Both silt and sand without obvious ¢lamentous algae were recorded as sediment. Analyses of data
Non-parametric multivariate analysis of variance (NPMANOVA; Anderson, 2001) was used to test hypotheses about di¡erences between assemblages of epibiota. The analyses were split into two sets because it is currently di⁄cult to analyse more than two factors using NPMANOVA. First, the hypothesis that assemblages of epibiota on arti¢cial reefs would di¡er from those on natural reefs was tested using data from assemblages on vertical surfaces and then separately on data from assemblages on horizontal surfaces. Second, the hypothesis that assemblages of epibiota on vertical surfaces would di¡er from those on horizontal surfaces was tested using data from assemblages on natural reefs and then separately on data from assemblages on concrete breakwalls. Multivariate analyses were used to test hypotheses about di¡erences between assemblages using data on percentage covers of taxa or the presence or absence of taxa. Percentage covers were used to test the hypothesis that di¡erences existed between assemblages due to the composition of assemblages, frequency of occurrence and/or the covers of taxa. Measures of presence or absence were used to test for di¡erences between assemblages due to composition and/or frequency of occurrence of taxa (Clarke & Warwick, 1994). Common taxa tend to be most important in analyses with measures of percentage covers of taxa and the relative importance of patchily distributed and less abundant taxa increases with measures of presence and absence (Clarke & Warwick, 1994). All multivariate analyses used Bray ^ Curtis dissimilarities. Two-dimensional non-metric multi-dimensional scaling (nMDS) was used to visualize these multivariate patterns of assemblages (Clarke & Warwick, 1994). The SIMPER procedure (PRIMER; Clarke & Warwick, 1994) was used to identify taxa that contributed to di¡erences between assemblages. Univariate analyses of variance (ANOVA) were used to test hypotheses about di¡erences in cover and number of taxa of epibiota on vertical and on horizontal surfaces on natural reefs and on concrete breakwalls. All data were transformed to x0.25 to make the variances of most taxa
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
N.A. Knott et al. 1119
Figure 1. Locations of subtidal reefs sampled in this study. Bare Island and Dobroyd Head are natural reefs and Port Botany and Port Kembla are concrete breakwalls. Journal of the Marine Biological Association of the United Kingdom (2004)
1120
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
Figure 2. nMDS ordinations comparing assemblages of epibiota on natural reefs (¢lled symbols) and on concrete breakwalls (empty symbols). Assemblages on: (A) vertical surfaces; and (B) on horizontal surfaces (N¼10). Natural reefs Bare Island (~) and Dobroyd Head (&); and concrete breakwalls Port Botany (*) and Port Kembla (^).
Table 1. Comparisons of assemblages of epibiota on natural rocky reefs and on concrete breakwalls using non-parametric multivariate analysis of variance (NPMANOVA) with measures of percentage cover or presence and absence of taxa. Substratum (Su) was a ¢xed orthogonal factor with two levels; natural reefs and concrete breakwalls. Reef was a random nested factor with two levels (N¼10). Percentage cover Source Vertical surfaces (73 taxa) Substratum Reef (Su) Residual Horizontal surfaces (70 taxa) Substratum Reef (Su) Residual
Presence/absence
df
MS
F
P
MS
F
P
1 2 36
7170 6023 1467
1.2 4.1
40.25 ***
11522 8586 1007
1.3 8.5
40.25 ***
1 2 36
6493 8074 1337
0.8 6.0
40.25 ***
23146 5458 1014
4.2 5.4
** ***
**, P50.01; ***, P50.001. Journal of the Marine Biological Association of the United Kingdom (2004)
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
N.A. Knott et al. 1121
Figure 3. Common epibiota that showed di¡erences in their percentage covers or number of taxa on natural reefs and on concrete breakwalls. Bare Island (BI) and Dobroyd Head (DH) are natural reefs; Port Botany (PB) and Port Kembla (PK) are concrete breakwalls. &, Vertical surfaces; , horizontal surfaces (N¼10).
more homogenous and the rest were transformed so that the results of all analyses could be interpreted at the same scale. Variances were tested for homogeneity using Cochran’s C-test. The covers of many taxa were heterogeneous, but because analysis of variance is robust to heterogeneity with large numbers of samples and balanced experimental designs (Underwood, 1997), these analyses were still done. Means were compared using Student ^ Newman ^ Keuls (SNK) tests for factors that showed signi¢cant variation.
RESULTS Filamentous algae, alone or with sediment (silt matrix), and encrusting coralline algae covered most of the surfaces on natural reefs and on concrete breakwalls. Invertebrates also covered large areas, with sponges and ascidians being the most dominant invertebrates in terms of cover. Barnacles, bryozoans, bivalves, soft corals, vermetid gastropods and serpulid polychaetes had much smaller covers and made up the rest of the invertebrate assemblages. Generally, covers of invertebrates were patchy among replicate surfaces (metres apart) and were often variable between reefs (tens of kilometres apart). Journal of the Marine Biological Association of the United Kingdom (2004)
Natural reefs versus arti¢cial reefs Multivariate analyses
There were often great di¡erences between assemblages on reefs of the same type (Table 1). This variation between assemblages on reefs made it di⁄cult to detect di¡erences between assemblages on natural reefs and on concrete breakwalls. With measures of percentage covers, assemblages on vertical surfaces were similar on natural reefs and on concrete breakwalls (Figure 2A, Table 1), as were those on horizontal surfaces (Figure 2B, Table 1). In terms of the composition of taxa (presence or absence data), the assemblages on vertical and on horizontal surfaces on concrete breakwalls appeared to di¡er from those on natural reefs (Figure 2A,B). On vertical surfaces, however, the assemblages on the two concrete breakwalls were very di¡erent (Figure 2A). Thus, no di¡erence between the type of reef was detected (Table 1). On horizontal surfaces, the assemblages on each breakwall were relatively similar (Figure 2B), as were the assemblages on each natural reef, while the assemblages on each type of substratum di¡ered clearly from each other (Figure 2B, Table 1). The SIMPER analyses indicated that the sponges Haliclona sp. (2.8% dissimilarity), Sponge sp. 2 (2.8%), Chondropsis sp. (2.7%), the ascidian Herdmania
1122
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
Table 2. Analyses of variance of common epibiota which did show patterns of di¡erence between their cover or number of taxa on natural rocky reefs and on concrete breakwalls. Substratum (Su) is a ¢xed orthogonal factor; natural reefs and concrete breakwalls. Reef (Re) is a random factor and is nested in substratum. Orientation (Or) is a ¢xed orthogonal factor with two levels; vertical and horizontal (N¼10). Data for all taxa were transformed to x0.25. ‘ ’, indicates that the test could not be interpreted because of a signi¢cant interaction between orientation and reef or substratum. NS, P40.05; *, P50.05; **, P50.01; ***, P50.001. Results of Student^Newman^Keuls (SNK) tests are presented below analyses. Source
df
MS
F
P
(A) Ascidians Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
1 2 1 1 2 72
5.07 0.16 8.27 1.28 0.38a 0.36a
32.48 * 0.43a 40.25 21.55a *** 3.34a NS 1.06 40.25
(F) Cymbastela concentrica1 Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
1 2 1 1 2 72
1 2 1 1 2 72
F
MS
P
(B) Herdmania momus 14.38 0.15 7.25 1.17 0.42a 0.39a
F
P
(C) Serpulids
96.87 ** 1.228 41094b 0.38a 40.25 0.000b 0.00a 17.36a *** 0.055 85.59a 2.79a NS 0.055 85.59a 1.06 40.25 0.001a,b 0.01 0.103a,b
(G) Ircinia sp.1
*** 40.25 40.25 40.25 40.25
(H) Sponge sp. 11
1.81 23.06 2.72 111.23 0.08 0.83a 40.25 0.02 0.21 1.81 23.06a 0.45 0.68 1.81 23.06a *** 0.45 0.68 0.08a 0.83 40.25 0.67 5.66 **3 0.09a 0.12 V: N¼C, H: N4C V: N4C, H: N¼C N: V5H, C:V¼H BI: V4H, DH: V¼H PB: V¼H, PK: V¼H (K) Oysters1
Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
MS
3.08 0.10 1.90 1.90 0.13 0.30
31.65 * 1.19 40.25 14.82 NS 14.82 NS 1.57 NS
MS
F
P
(D) Polymastia sp. 1.160 863b 0.001b 0 0.041 1173 0.041 1173a 0.000a,b 0 0.084a,b
MS
F
P
(E) Chondropsis sp.1
*** 4.438 3781 40.25 0.001 0.01a 40.25 40.25 2.165 122.3a 40.25 2.165 122.3a *** 40.25 0.018a 0.10 40.25 0.177a V: N¼C, H: N4C2 N:V5H, C:V¼H
(I) Dendroceratid sp. 11 (J) Capnella gaboensis1 3.77 20.36 11.59 10.20 0.19 1.06 1.14 3.85 1.35 0.93 2.47 1.61 1.35 0.93 2.47 1.61 1.46 8.39 ***3 1.54 5.20 **3 0.17 0.30 V: N¼C, H: N¼C V: N¼C, H: N¼C BI: V¼H, DH: V5H BI:V¼H,DH:V5H PB: V¼H, PK: V¼H PB:V¼H,PK:V¼H
(L) Sponge sp. 24? (M) No. of sponge taxa4 (iv) Encrusting corallines4
4.43 2.66 2.12 13.67 1.67 14.70 0.16 1.94 1.34 1.76 4.00 19.59 0.47 0.61 2.05 10.06 0.76 6.71 **3 0.20 2.55 0.11 0.08 V: N¼C, H: N¼C BI: V¼H, DH: V4H PB: V¼H, PK: V¼H
NS NS * NS NS
3.31 32.38 2.91 0.10 1.65 0.29a 0.70 2.97 0.46 1.22 5.17 4.42 0.24 3.81 *3 0.71 0.06 0.33a V: N¼C, H: N4C BI: V¼H, DH: V¼H PB: V4H, PK: V4H
6.98a ** 1.20 40.25 0.97 40.25 8.00 NS 2.01a NS
a and b , Post-hoc pooling was done to increase the power of the main tests if P40.25. Superscripts identify the MS that were pooled and F ratios that changed due to this pooling (Underwood, 1997). 1, Variances were heterogeneous (Cochran’s C-test, P50.01). 2, Abbreviations for the SNK tests: V, vertical; H, horizontal; N, natural reef; C, concrete breakwall; BI, Bare Island; DH, Dobroyd Head; PB, Port Botany; PK, Port Kembla. 3, Because of this interaction separate analyses of variance were done on the cover or number of taxa on vertical or horizontal surfaces to test whether there were di¡erences between natural reefs or concrete breakwalls. 4,Variances were heterogeneous (Cochran’s C-test, P 50.05).
momus (2.6%) and the soft coral Capnella gaboensis (2.5%) were the ¢ve most important taxa responsible for the di¡erence between assemblages on horizontal surfaces on natural reefs and on concrete breakwalls. Nevertheless, many other taxa also contributed to this di¡erence in small amounts. Univariate analyses
Several invertebrate taxa showed clear di¡erences between natural reefs and concrete breakwalls. The total cover of ascidians was greater on breakwalls than on natural reefs (Figure 3A, Table 2A). This pattern was almost exclusively due to the large solitary ascidian Journal of the Marine Biological Association of the United Kingdom (2004)
Herdmania momus (Figure 3B, Table 2B). Serpulid polychaetes also showed the same pattern, although their cover on these reefs was relatively small (Figure 3C, Table 2C). Many sponges were observed only on natural reefs e.g. Polymastia sp., Chondropsis sp., Cymbastela concentrica, Ircinia sp., Sponge sp. 1 and Dendroceratid sp. (Figure 3D ^ I, Table 2D ^ I). The soft coral Capnella gaboensis also showed this pattern (Figure 3J, Table 2J), while oysters also tended to be found mainly on natural reefs rather than on concrete breakwalls (Figure 3K, Table 2K). Sponge sp. 2 was the only sponge that covered a larger area on breakwalls than on natural reefs (Figure 3L),
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
N.A. Knott et al. 1123
Figure 4. Common epibiota that did not show di¡erences in their percentage covers or number of taxa on natural reefs and on concrete breakwalls. Bare Island (BI) and Dobroyd Head (DH) are natural reefs; Port Botany (PB) and Port Kembla (PK) are concrete breakwalls. &, Vertical surfaces; , horizontal surfaces (N¼10).
although this was not statistically signi¢cant (Table 2L). A striking result of this study was that there were fewer taxa of sponge on horizontal surfaces on concrete breakwalls than on natural reefs (Figure 3M, Table 2M). On vertical surfaces, however, there was no di¡erence between the two types of reefs (Figure 3M, Table 2M). Encrusting coralline algae covered a larger area on natural reefs than on concrete breakwalls (Figure 3N, Table 2N). The total covers of invertebrates, sponges, barnacles and bryozoans on natural reefs were not di¡erent signi¢cantly from those on concrete breakwalls (Figure 4A ^ D, Table 3A ^ D). Similarly, most individual invertebrate taxa showed no di¡erences between the types of substratum (Figure 4E ^ L). Almost all of these groups and individual taxa varied greatly between reefs of each type (Table 3A ^ K). The cover of the bivalve Cleidothaerus albidus, however, varied little between reefs and also between types of reefs (Table 3L, Figure 4L). The number of ascidian taxa was also similar on natural reefs and on breakwalls for both orientations (Figure 4M, Table 3M). The cover of algae on natural reefs was similar to that on concrete breakwalls, except for encrusting coralline algae (Figure 4N ^ P, Table 3N ^ P). Natural reefs and concrete breakwalls had similar covers of silt matrix and sediment (Figure 4Q,R, Table 3Q,R). The amount of bare space was also similar on the two types of substratum (Figure 4S, Table 3S). Vertical versus horizontal surfaces Multivariate analyses
At most reefs, assemblages of epibiota on vertical surfaces were di¡erent from those on horizontal surfaces (Figure 5, Table 4). There was, however, an interaction Journal of the Marine Biological Association of the United Kingdom (2004)
between orientation and reef on natural reefs and on concrete breakwalls. This occurred because the nature of the di¡erence between assemblages on vertical and on horizontal surfaces was di¡erent at each reef. This pattern can be seen most clearly in the nMDS plot of assemblages on concrete breakwalls with measures of presence or absence of taxa (Figure 5B). In this plot, assemblages on vertical surfaces are separated from those on horizontal surfaces, but assemblages on each reef also separate from each other. Most assemblages showed the same pattern with percentage cover and presence or absence data, except for Bare Island (Table 4, Figure 5A). This indicated that, on most of these reefs, assemblages on vertical and on horizontal surfaces di¡ered mostly because of their composition and frequency of occurrence. Using percentage covers, there were often great di¡erences between assemblages on surfaces with the same orientation within a reef. This was indicated by the large average Bray ^ Curtis dissimilarities (Table 4). With measures of presence or absence, however, assemblages were relatively more similar (Table 4). Filamentous algae, alone or with silt, and encrusting coralline algae were the main taxa that contributed to the dissimilarity of assemblages on vertical and on horizontal surfaces using measures of percentage cover of taxa (each 420%: SIMPER analysis). Few invertebrates contributed more than 5% to the di¡erences. The exceptions were the soft coral Capnella gaboensis (12%) on natural reefs and the large solitary ascidian Herdmania momus (7%) on concrete breakwalls. Although, individually, few invertebrate taxa contributed greatly to di¡erences between assemblages on vertical and on horizontal surfaces, many contributed
1124
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
Table 3. Analyses of variance of common epibiota which did not show any patterns of di¡erence between their cover or nuumber of taxa on natural rocky reefs and on concrete breakwalls. Abbreviations are the same as in Table 2. Source
df
MS
F
P
(A) Invertebrates Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
1 2 1 1 2 72
1 2 1 1 2 72
0.53 1.19 1.49 0.04 0.43 0.15
1 2 1 1 2 72
1 2 1 1 2 72
(B) Sponges
(G) Haliclona sp.
(L) Cleidothaerus albidus
MS
F
P
(C) Barnacles
MS
P
(D) Bryozoans
1.10 0.53 0.60 2.08 9.76 2.60 3.09 3.81 1.56 0.19 0.24 1.15 0.81 3.81 *3 1.66 0.21 0.40 V: N¼C, H: N¼C BI: V¼H, DH: V4H PB: V4H, PK: V4H (H) Clathria pyramida
F
0.23 6.52 0.94 0.69 4.18
(Q) Silt matrix4
Journal of the Marine Biological Association of the United Kingdom (2004)
F
(M) Number of ascidian taxa
(R) Sediment4 0.30 0.68 49.16 0.14 9.27
P
(E) Pronax sp.
0.99 0.11 40.25 8.73 16.96a *** 2.39 7.05a * 0.65 1.91a 40.25 0.34a 0.66 40.25 0.52a V: N¼C, H: N¼C BI: V4H, DH: V¼H PB: V¼H, PK: V4H
(I) Halisarca sp.1
(J) Botrylloides leachi4
0.60 0.23 0.05 0.15 40.25 0.134 0.70 2.60 6.52 0.31 1.11a 40.25 0.194 1.37 1.56 0.94 2.53 10.55a ** 1.248 2.17 1.15 0.69 0.03 0.14a 40.25 0.001 0.00 1.66 4.18 *3 0.24a 0.86 40.25 0.576 4.07 0.40 0.28a 0.142 V: N¼C, H: N¼C V: N¼C, H: N¼C BI: V¼H, DH: V4H BI: V¼H, DH: V4H PB: V¼H, PK: V¼H PB: V¼H, PK: V4H
2.72 3.11 NS 0.19 2.70 0.04 0.88 2.91a NS 0.30 1.76 0.12 0.01 0.04a 40.25 0.04 3.41 79.04 0.66 3.61a NS 0.06 0.76 0.23 0.18a 0.61 40.25 1.13 5.21 **3 1.61 0.30a 0.17 0.17 V: N¼C, H: N¼C BI: V¼H, DH: V5H PB: V¼H, PK: V¼H
small amounts and, as a group, they contributed on average over 20% to the dissimilarity between these assemblages. Moreover, invertebrates contributed greatly to di¡erences between the composition of assemblages on vertical and on horizontal surfaces (i.e. using only presence or absence of taxa). Silt matrix, encrusting corallines and ¢lamentous algae contributed little, however, to compositional di¡erences between surfaces.
MS
*3
(N) Filamentous algae
0.12 0.21 1.43 3.86 40.25 0.18 0.21 40.25 0.04 0.01 0.54 14.22 0.37 2.17 40.25 0.87 4.10a * 5.08 15.18 1.46 4.69 4.03 12.04 NS 3.87 197.32a ** 3.11 1.53 0.12 0.37 0.25 0.75 40.25 0.01 0.64a 40.25 0.01 0.00 0.31 8.20 ***3 0.33 1.97 NS 0.02a 0.09 40.25 2.03 6.07 **3 0.04 0.17 0.21a 0.34 V: N¼C, H: N¼C V: N¼C, H: N¼C BI: V¼H, DH: V4H BI: V4H, DH: V4H PB: V4H, PK: V¼H PB: V4H, PK: V¼H (P) Peyssonelia sp.
Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
P
0.44 40.25 4.06 7.04 NS 8.10 *** 0.58 3.54 * 3.47 NS 0.45 2.32a 40.25 0.09 40.25 0.23 1.18a 40.25 2.92 NS 0.19a 1.19a 40.25 0.16a
(K) Serpulorbis sipho4 Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
F
1.00 0.78 40.25 3.83 1.37 40.25 1.29 10.98 *** 2.80 12.76a *** 5.92 30.02 * 4.19 54.93a *** 1.85 9.40 NS 0.91 11.92a NS 0.20 1.68 NS 0.08a 0.35 40.25 0.12 0.22a
(F) Darwinella sp. Substratum Reef (Su) Orientation Or Su Or Re (Su) Residual
MS
*3
(O) Foliose algae 1.93 0.71 7.23 1.60 2.12 0.41
2.70 1.76 3.41 0.76 5.21
**3
(S) Bare space
0.28 0.11 NS 2.68 10.58a *** 3.42 22.67a *** 0.25 1.62a 40.25 0.15a 0.60 40.25 0.25a V: N¼C, H: N¼C BI: V5H, DH: V5H PB: V5H, PK: V5H **3
Univariate analyses
Table 5 provides a general summary of the patterns of covers and numbers of taxa of the common epibiota on vertical and on horizontal surfaces. Invertebrates consistently covered a larger area on vertical than on horizontal surfaces (Figure 4A, Table 3A). Sponges and ascidians were largely responsible for this pattern (Figures 4B & 3A, Tables 3B & 2A).
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
N.A. Knott et al. 1125
Figure 5. nMDS ordinations comparing assemblages of epibiota on vertical surfaces (¢lled symbols) and on horizontal surfaces (empty symbols). Assemblages on: (A) natural reefs Bare Island (~, ~) and Dobroyd Head (&, &) (N¼10); and on (B) concrete breakwalls Port Botany (*, *) and Port Kembla (^, ^) (N¼10).
The greater cover of sponges on vertical than on horizontal surfaces was due to several taxa. Pronax sp., Halisarca sp. and Sponge sp. 2 clearly showed this pattern (Figures 4E,I & 3L, Tables 3E,I & 2L). Darwinella sp. and Haliclona sp. showed the same trend at most sites, but this was not statistically signi¢cant (Figure 4F,G, Table 3F,G). The pattern of cover of ascidians was due primarily to Herdmania momus (Figure 3B, Table 2B), although the colonial ascidian Botrylloides leachi also showed the same pattern on most reefs (Figure 4J, Table 3J). Rarer species of ascidian, however, appeared to cover similar amounts of Journal of the Marine Biological Association of the United Kingdom (2004)
area on each orientation (e.g. colonial ascidians Didemnum moseleyi, Eudistoma sp. and Polyandrocarpa lapidosa and solitary ascidians Cnemidocarpa pedata and Phallusia obesa). Many other invertebrates also showed the same general pattern as sponges and ascidians. Barnacles, bryozoans, oysters and the vermetid gastropod Serpulorbis sipho covered a larger area on vertical than on horizontal surfaces at most reefs (Figures 4C,D,K & 3K, Tables 3C,D,K & 2K). Surprisingly, the covers of several invertebrate taxa did not di¡er between the two orientations. For
1126
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
Table 4. Comparisons of assemblages of epibiota on vertical and on horizontal surfaces on natural rocky reefs and on concrete breakwalls using non-parametric multivariate analysis of variance (NPMANOVA) with measures of percentage cover or presence and absence of taxa. Orientation (Or) was a ¢xed orthogonal factor with two levels; vertical and horizontal. Reef (Re) was a random orthogonal factor with two levels (N¼10). A posteriori pair-wise comparisons and average Bray^Curtis dissimilarities are also shown beneath each analysis. NS, P40.05, **, P50.01, ***, P50.001. Natural reefs (taxa 68) Percentage cover
Breakwalls (53 taxa)
Presence/absence
Percentage cover
Presence/absence
Source
df
MS
F
P
MS
F
P
MS
F
P
MS
F
P
Orientation Reef Or Re Residual
1 1 1 36
9299 2922 7781 1579
1.2 1.9 4.9
***
9658 6962 3909 1235
2.5 5.6 3.2
**
5818 12034 5457 1225
1.1 9.8 4.5
***
10958 14199 3016 787
3.6 18.0 3.8
***
BI 54 61 58NS
DH 49 50 72***
Vertical Horizontal V vs H
BI 52 53 60***
DH 42 47 58**
PB 58 42 58**
PK 47 39 53***
PB 36 43 49***
PK 44 28 55***
BI, Bare Island; DH, Dobroyd Head; PK, Port Kembla; PB, Port Botany.
Table 5. Overview of patterns of cover of common epibiota on vertical and on horizontal surfaces. Vertical4horizontal Consistent patternsa Invertebrates Sponges Ascidians Pronax sp. Halisarca sp. Sponge sp. 1 Herdmania momus Serpulorbis sipho Variable patternsc Botrylloides leachiiexcept BI, PB: V¼H Barnaclesexcept BI: V¼H Bryozoansexcept DH: V¼H
Vertical¼horizontal
Vertical5horizontal
Polymastia sp.b Serpulids Encrusting corallines Peyssonnelia sp.
Cymbastela concentricab Chondropsis sp.b
Clathria pyramidaexcept DH: V4H Ircinia sp.b, except BI: V4H Filamentous algaeexcept DH, PB: V4H Foliose algaeexcept DH, PK: V5H
Dendroceratid sp.except BI: V¼H Cleidothaerus albidus Capnella gaboensisb, except BI: V¼H
Trendsd Darwinella sp. Haliclona sp.except BI: V¼H Oysters
Sponge sp. 2
a, Main e¡ects or consistent pattern at each reef of occurrence. b, Taxa only found on natural reefs. c, Abbreviations: BI, Bare Island; DH, Dobroyd Head; PB, Port Botany; PK, Port Kembla; V, vertical; H, horizontal. d, Non-signi¢cant.
example, serpulids and the sponge Polymastia sp. covered similar amounts of area on vertical and on horizontal surfaces (Figure 3C,D, Table 2C,D). The sponges Ircinia sp. and Clathria pyramida showed similar patterns on some reefs, although on others they covered a larger area on vertical surfaces (Figures 3G & 4H, Tables 2G & 3H). Thus, the pattern of cover on vertical and on horizontal surfaces varies among reefs for some invertebrates. In contrast, several taxa covered a larger area on horizontal surfaces. The sponges Cymbastela concentrica and Journal of the Marine Biological Association of the United Kingdom (2004)
Chondropsis sp. clearly showed this alternative pattern (Figure 3E,F, Table 2E,F). Furthermore, Sponge sp. 1 and the bivalve Cleidothaerus albidus consistently covered a larger area on horizontal than on vertical surfaces, although these di¡erences were not statistically signi¢cant (Figures 3H & 4L, Tables 2H & 3L). The sponge Dendroceratid sp. and the soft coral Capnella gaboensis were observed only on natural reefs and each covered a larger area on horizontal than on vertical surfaces at Dobroyd Head, but similar amounts of area at Bare Island (Figure 3I, J, Table 2I, J).
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures The numbers of taxa of sponges and of ascidians also di¡ered between vertical and horizontal surfaces. On concrete breakwalls there was a greater mean number of sponge taxa on vertical surfaces (Figure 3M, Table 2M). On natural reefs, however, the number of sponge taxa on each orientation was similar (Figure 3M, Table 2M), although at Dobroyd Head, there was the opposite trend, with there being more taxa on horizontal surfaces (Figure 3M). The number of ascidian taxa was, however, consistently greater on vertical surfaces at all reefs (Figure 4M, Table 3M). Algae, especially encrusting coralline algae and ¢lamentous algae, covered large areas on all reefs (Figures 3O & 4N). Generally, algae showed no consistent di¡erence in their cover on vertical or horizontal surfaces. Encrusting corallines and Peyssonnelia sp. covered similar amounts of area on vertical and on horizontal surfaces (Figures 3N & 4P, Tables 2N & 3P). Filamentous and foliose algae also tended to show no di¡erences in their covers on the two orientations at most reefs (Figure 4N,O, Table 3N,O). The silt matrix covered very large areas on both vertical and horizontal surfaces at all reefs (Figure 4Q, Table 3Q). It appeared that the ¢lamentous algae trapped silt both on horizontal and on vertical surfaces and the covers were usually similar on both surfaces. Sediment covered a larger area on horizontal surfaces (Figure 4R, Table 3R). There were only very small patches of bare space on the reefs, but these were larger on vertical than on horizontal surfaces (Figure 4S, Table 3S).
DISCUSSION Natural reefs versus concrete breakwalls
The large and similar covers of ¢lamentous algae (alone and with silt) on natural reefs and on concrete breakwalls made assemblages on the two types of reefs appear similar. Moreover, the great variability in the covers of taxa on reefs of the same type made it di⁄cult to detect potential di¡erences between assemblages on natural reefs and those on concrete breakwalls. Despite the assemblages appearing similar on natural reefs and on concrete breakwalls, many taxa showed clear di¡erences in their covers. For example, Herdmania momus and serpulids covered a larger area on concrete breakwalls than on natural reefs, while encrusting coralline algae showed the opposite pattern. These di¡erences were, however, not enough to make the whole assemblages di¡er when using measures of percentage cover. In terms of composition and frequency of occurrence (measures of presence or absence), assemblages on arti¢cial reefs either appeared di¡erent (e.g. vertical surfaces) or were signi¢cantly di¡erent (e.g. horizontal surfaces) from those on natural reefs. On vertical surfaces, this apparent di¡erence was probably due to the spatial relationships among taxa that showed di¡erences in their covers on the two types of reefs (e.g. their combined presence or absence on replicate surfaces). On horizontal surfaces, the di¡erences between assemblages on arti¢cial and on natural reefs was due mainly to many taxa being found on natural reefs but not on concrete breakwalls while, conversely, all of the species on concrete breakwalls Journal of the Marine Biological Association of the United Kingdom (2004)
N.A. Knott et al. 1127
were found on natural reefs. This was clearly illustrated in the greater number of sponge taxa on horizontal surfaces on natural reefs than on horizontal surfaces on concrete breakwalls. This is an interesting result because previous studies have found that subtidal assemblages on arti¢cial structures were more diverse than on natural reefs (Connell & Glasby, 1999; Glasby, 1999a). Several physical di¡erences between the surfaces on natural rocky reefs and on concrete breakwalls could be responsible for the observed di¡erences in the covers and numbers of taxa of epibiota. There were clear di¡erences in the material of the breakwalls and the natural reefs, i.e. concrete versus sandstone. Moreover, the surfaces of the breakwalls appeared relatively smooth compared with the pitted and grooved surfaces of the natural reefs. These di¡erences in material or the structure of the surfaces may a¡ect the recruitment and growth of invertebrates causing the observed di¡erences in the cover and richness of taxa. Furthermore, di¡erences in the richness of sponges may have also been due to the age of the concrete breakwalls. These breakwalls are 25 and 50 years old (Port Botany and Port Kembla, respectively). Despite being in the water for long periods of time, the assemblages of sponges that have developed on their horizontal surfaces are quite depauperate of species compared with natural rocky reefs. Therefore, there are either strong di¡erences between the habitats provided by natural reefs and by concrete breakwalls, or the recruitment of sponges is so slow that it takes very long periods of time before assemblages on arti¢cial structures become like those on natural reefs on wave-exposed coasts. Considering the numerous studies that have reported poor recruitment of sponges (Keough, 1984; Butler, 1986; Glasby, 1999b) or no recruitment at all (Wilkinson, 1978; Dayton, 1979; Knott, 2002), the latter explanation appears to be likely. Previous quantitative comparisons of subtidal assemblages of epibiota on natural reefs and on arti¢cial structures (Connell & Glasby, 1999; Glasby, 1999a) have found that many taxa had very di¡erent covers on either type of substratum. The present study, however, has found that many taxa were more variable between reefs of the same type than between natural reefs and concrete breakwalls, although several taxa did show clear di¡erences between the two types of substratum. Most arti¢cial structures compared in previous studies have been very di¡erent from natural reefs in several ways. For example, pilings are usually shaded and pontoons £oat; and they are usually isolated from other hard substrata. These factors alone appear to be responsible for much of the di¡erences between these structures and natural rocky reefs e.g. shading (Glasby, 1999b); £oating (Holloway & Connell, 2002); isolation (Keough, 1984). The concrete breakwalls sampled in the present study are probably more similar to natural reefs because they appear to be shaded similarly to natural reefs and they do not move. Moreover, they are more similar in their three-dimensional structure to natural reefs, possessing upward facing horizontal and vertical surfaces. Reefs were also separated by relatively large distances (20^70 km), hence, spatial variation was likely to be substantial. Previous studies were done over smaller spatial scales (tens to hundreds of metres) and, therefore, were likely to have had less spatial variation ^ making them more likely to have detected potential
1128
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
e¡ects. Although the covers of several taxa on natural reefs were di¡erent from those on concrete breakwalls, many other taxa appeared not to be a¡ected by whether the habitat was natural or arti¢cial. For example, the sponges Pronax sp., Darwinella sp., Clathria pyramida, the ascidian Botrylloides leachi, barnacles, bryozoans and ¢lamentous and foliose algae covered similar amounts of area on concrete breakwalls and on natural reefs. Although only two natural reefs and two arti¢cial reefs were sampled in this study, the range of covers of these taxa on the concrete breakwalls appear to be within that found on several nearby natural reefs sampled in other studies (Roberts & Davis, 1996; Underwood & Chapman, 1996; Wright et al., 1997). It would appear, therefore, that the material and the structure of surfaces are not important features of subtidal reefs for many epibiotic taxa. Vertical versus horizontal surfaces
Assemblages of epibiota on vertical surfaces were usually very di¡erent from those on horizontal surfaces. This ¢nding is consistent with the general pattern found on subtidal reefs in the northern hemisphere (Hartnoll, 1983; Sebens, 1985; Wendt et al., 1989; De Kluijver, 1991). There was, however, much variation in these assemblages within orientations and between reefs. Often, the variation in assemblages on the two orientations was only slightly greater than that among assemblages on surfaces with the same orientation. This great variation among assemblages is important and should be recognized, rather than ignored, because it is part of the complexity of these assemblages (Sebens, 1991; Ginn et al., 2000). Nevertheless, such variation makes it di⁄cult to detect patterns in these assemblages and possibly explains why so few studies have attempted to quantify the e¡ects of orientation on established assemblages of epibiota. Surprisingly, assemblages on vertical surfaces were not always found to be di¡erent from those on horizontal surfaces (e.g. at Bare Island). This indicates that the e¡ects of orientation may not be as consistent as previously thought. The general observation that invertebrates cover a larger area on vertical than on horizontal surfaces on subtidal reefs in the northern hemisphere (Logan et al., 1984; Wendt et al., 1989; Witman & Sebens, 1991; Baynes, 1999; Ginn et al., 2000) clearly also existed on subtidal reefs of the Sydney ^ Illawarra coast. Many invertebrates showed this pattern (e.g. Pronax sp., Herdmania momus, barnacles, bryozoans) and these taxa contributed greatly to di¡erences between assemblages on vertical and on horizontal surfaces. Nevertheless, there were several invertebrate taxa that either showed no clear pattern of cover on these surfaces (e.g. Polymastia sp., Clathria pyramida) or, oppositely, covered a larger area on horizontal than on vertical surfaces (e.g. Capnella gaboensis, Cleidothaerus albidus). Wendt et al. (1989) and Witman & Sebens (1991) also found several species that either showed no e¡ect of orientation or covered a larger area on horizontal than on vertical surfaces. The existence of these alternative patterns should be acknowledged in order to describe the patterns of covers of invertebrates on subtidal reefs appropriately. Algae contributed greatly to the similarity of assemblages on vertical and on horizontal surfaces. Filamentous, Journal of the Marine Biological Association of the United Kingdom (2004)
foliose and encrusting coralline algae and Peyssonnelia sp. each covered similar amounts of area on vertical and on horizontal surfaces on most reefs. Because algae were relatively common and covered large areas on natural reefs and on concrete breakwalls, they were very prominent when the assemblage was measured in terms of the percentage covers. Considering di¡erences in assemblages in terms of presence or absence of taxa reduced the importance of algae and the resulting patterns were mainly driven by di¡erences in the invertebrate taxa. This was because many invertebrates occurred more frequently on either vertical or horizontal surfaces. Moreover, several invertebrate taxa were found almost exclusively on either vertical surfaces (e.g. Halisarca sp., Serpulorbis sipho) or on horizontal surfaces (e.g. Cymbastela concentrica, Chondropsis sp.). The total cover of sponges, as predicted, was greater on vertical than on horizontal surfaces, supporting the general pattern of cover of sponges on subtidal reefs (Wendt et al., 1989; Witman & Sebens, 1991; Ginn et al., 2000). Nevertheless, there was not a simple relationship between the cover of sponges and orientation because there was much variability in the patterns of cover of individual species, as has been found in other studies (Logan et al., 1984; Wendt et al., 1989; Witman & Sebens, 1991). Furthermore, patterns of cover of species varied among reefs. For example, Clathria pyramida covered a larger area on vertical than on horizontal surfaces at Dobroyd Head, but similar amounts of area on each orientation at Bare Island and at Port Botany. It is, therefore, necessary to sample more than one reef to assess patterns of cover of sponges adequately. The total cover of ascidians was also larger on vertical than on horizontal surfaces. This pattern has also been reported by Gotelli (1987) for a colonial ascidian. Wendt et al. (1989), however, observed that solitary ascidians covered similar amounts of area on vertical and on horizontal surfaces. Similarly, Logan et al. (1984) found that several species of ascidian were either present on both orientations or were only on horizontal surfaces. In the current study, the solitary ascidian Herdmania momus clearly covered a greater area on vertical surfaces and the colonial ascidian Botrylloides leachi showed the same pattern, although this was not statistically signi¢cant. Nonetheless, rarer colonial and solitary ascidians appeared to cover similar amounts of area on vertical and on horizontal surfaces. Therefore, it would appear that the e¡ect of orientation on the cover of ascidians also cannot be generalized across species. Barnacles, bryozoans and serpulid polychaetes covered a greater area on vertical than on horizontal surfaces and soft coral Capnella gaboensis showed the opposite pattern. These results are consistent with those of Logan et al. (1984) and Wendt et al. (1989). The covers of oysters and Cleidothaerus albidus on vertical surfaces were also di¡erent from those on horizontal surfaces. There appear to be no published studies that have sampled quantitatively the cover of these molluscs on vertical and on horizontal surfaces of subtidal reefs, so the consistency of these patterns cannot be assessed. Several physical and biological factors appear to di¡er between vertical and horizontal surfaces and may play a role in creating the di¡erences in the cover of invertebrates
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures on the two orientations. The total cover of invertebrates on vertical and on horizontal surfaces was negatively related to the cover of sediment (which covered approximately 25% of the area on horizontal surfaces). Sedimentation is known to a¡ect algae (Airoldi, 2003), but there appear to be few manipulative ¢eld studies on the e¡ects of sedimentation on invertebrates. Competition for space between invertebrates and algae has been hypothesized to be an important in£uence on the cover of invertebrates on vertical and horizontal surfaces (Hartnoll, 1983; Sebens, 1985). Baynes (1999), on a tropical rocky reef, found a negative relationship between the cover of tur¢ng algae and invertebrates on the two surface orientations. In the current study, however, no such relationship was observed. Other variables which may di¡er between the two orientations are grazing (Harris & Irons, 1982), light (Young & Chia, 1984) and water-£ow (Witman & Sebens, 1991). These factors and sedimentation o¡er potential explanations for the observed patterns of cover of invertebrates on vertical and on horizontal surfaces. Rarely, however, have these factors been evaluated to explain di¡erences between assemblages on vertical and on horizontal surfaces (except for grazing: Baynes, 1999). Testing hypotheses about the e¡ects of these factors to evaluate them as possible explanations for the patterns observed in this and other studies would be extremely valuable in understanding the basic ecology of epibiota on subtidal reefs. The strong and consistent di¡erences between assemblages of epibiota on vertical and on horizontal surfaces suggest that the orientation of surfaces of arti¢cial structures may be of greater importance than their material or structure in terms of their e¡ects on the biological diversity of epibiota in marine habitats. Therefore, the orientation of surfaces of arti¢cial structures introduced into marine environments should be considered carefully, in relation to the surrounding substrata, in order to minimize the e¡ects such structures will have on the biological diversity of these areas. This study was supported by the Australian Research Council through its Special Research Programme and by a Postgraduate Scholarship from the Centre for Research on Ecological Impacts of Coastal Cities to Nathan Knott. We thank F. Barros, I. Carlson, G. House¢eld, B. Kelaher and M. Sage for help in the ¢eld. Two anonymous referees made helpful comments on an earlier draft of the paper.
REFERENCES Airoldi, L., 2003. The e¡ects of sedimentation on rocky coast assemblages. Oceanography and Marine Biology, 41, 161^236. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology, 26, 32^46. Andrew, N.L. & Mapstone, B.D., 1987. Sampling and the description of spatial pattern in marine ecology. Oceanography and Marine Biology. Annual Review, 25, 39^90. Baynes, T.W., 1999. Factors structuring a subtidal encrusting community in the southern gulf of California. Bulletin of Marine Science, 64, 419^450. Bell, J.J. & Barnes, D.K.A., 2000. The in£uences of bathymetry and £ow regime upon the morphology of sublittoral sponge communities. Journal of the Marine Biological Association of the United Kingdom, 80, 707^718.
Journal of the Marine Biological Association of the United Kingdom (2004)
N.A. Knott et al. 1129
Butler, A.J., 1986. Recruitment of sessile invertebrates at ¢ve sites in the Gulf of St. Vincent, South Australia. Journal of Experimental Marine Biology and Ecology, 97, 13^36. Clarke, K.R. & Warwick, R.M., 1994. Change in marine communities: an approach to statistical analysis and interpretation. Plymouth, UK: Natural Environment Research Council. Connell, S.D. & Glasby, T.M., 1999. Do urban structures in£uence local abundance and diversity of subtidal epibiota? A case study from Sydney Harbour, Australia. Marine Environmental Research, 47, 373^387. Dayton, P., 1979. Observations of growth, dispersal and population dynamics of some sponges in McMurdo Sound, Antarctica. In Sponge biology (ed. C. Levi and N. BouryEsnault), pp. 271^282. Paris: Centre National de la Recherche Scienti¢que. De Kluijver, M.J., 1991. Sublittoral hard substrate communities o¡ Helgoland. Helgola«nder Meeresuntersuchungen, 45, 317^344. Gabriele, M., Bellot, A., Gallotti, D. & Brunetti, R., 1999. Sublittoral hard substrate communities of the northern Adriatic Sea. Cahiers de Biologie Marine, 40, 65^76. Ginn, B.K., Logan, A. & Thomas, M.L.H., 2000. Sponge ecology on sublittoral hard substrates in a high current velocity area. Estuarine, Coastal and Shelf Science, 50, 403^414. Glasby, T.M., 1999a. Di¡erences between subtidal epibiota on pier pilings and rocky reefs at marinas in Sydney, Australia. Estuarine, Coastal and Shelf Science, 48, 281^290. Glasby, T.M., 1999b. E¡ects of shading on subtidal epibiotic assemblages. Journal of Experimental Marine Biology and Ecology, 234, 275^290. Glasby, T.M. & Connell, S.D., 1999. Urban structures as marine habitats. Ambio, 28, 595^598. Gotelli, N.J., 1987. Spatial and temporal patterns of reproduction, larval settlement, and recruitment of the compound ascidian Aplidium stellatum. Marine Biology, 94, 45^51. Harris, L.G. & Irons, K.P., 1982. Substrate angle and predation as determinants in fouling community succession. In Arti¢cial substrates (ed. J. Cairns), pp. 131^174. Michigan: Ann Arbor Science. Hartnoll, R.G., 1983. Substratum. In Sublittoral ecology: the ecology of the shallow sublittoral benthos (ed. R. Earll and D.G. Erwin), pp. 97^124. Oxford: Clarendon Press. Holloway, M. & Connell, S., 2002. Why do £oating structures create novel habitats for subtidal epibiota? Marine Ecology Progress Series, 235, 43^52. Keough, M.J., 1984. E¡ects of patch size on the abundance of sessile marine invertebrates. Ecology, 65, 423^437. Keough, M.J., 1999. Sessile animals. In Under southern seas: the ecology of Australia’s rocky reefs (ed. N.L. Andrew), pp. 136^145. Sydney: University of New South Wales Press. Knott, N.A., 2002. Epibiota on natural reefs and arti¢cial structures. PhD thesis, University of Sydney, Sydney, Australia. Logan, A., Page, F.H. & Thomas, M.L.H., 1984. Depth zonation of epibenthos on sublittoral hard substrates o¡ Deer Island, Bay of Fundy, Canada. Estuarine, Coastal and Shelf Science, 18, 571^592. Oliver, I. & Beattie, A.J., 1996. Invertebrate morphospecies as surrogates for species: a case study. Conservation Biology, 10, 99^109. Roberts, D., 1999. Rocky reefs beyond most divers. In Under southern seas: the ecology of Australia’s rocky reefs (ed. N.L. Andrew), pp. 18. Sydney: University of New South Wales Press. Roberts, D.E. & Davis, A.R., 1996. Patterns in sponge (Porifera) assemblages on temperate coastal reefs o¡ Sydney, Australia. Marine and Freshwater Research, 47, 897^906. Roberts, D.E., Fitzhenry, S.R. & Kennelly, S.J., 1994. Quantifying subtidal macrobenthic assemblages on hard substrata using a jump camera method. Journal of Experimental Marine Biology and Ecology, 177, 157^170.
1130
N.A. Knott et al.
Epibiota on vertical and on horizontal surfaces on natural reefs and on arti¢cial structures
Roberts, D.E., Smith, A., Ajani, P. & Davis, A.R., 1998. Rapid changes in encrusting marine assemblages exposed to anthropogenic point-source pollution: a ‘beyond BACI’ approach. Marine Ecology Progress Series, 163, 213^224. Sebens, K.P., 1985. Community ecology of vertical rock walls in the Gulf of Maine, USA: small-scale processes and alternative community states. In The ecology of rocky coasts (ed. P.G. Moore and R. Seed), pp. 346^371. Sevenoaks, Kent, England: Hodder & Stoughton. Sebens, K.P., 1991. Habitat structure and community dynamics in marine benthic systems. In Habitat structure: the physical arrangement of objects in space (ed. S.S. Bell et al.), pp. 211^234. New York: Chapman & Hall. Underwood, A.J., 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge: Cambridge University Press. Underwood, A.J. & Chapman, M.G., 1996. Subtidal assemblages on rocky reefs at a cli¡-face sewage outfall (North Head, Sydney, Australia): what happened when the outfall was turned o¡ ? Marine Pollution Bulletin, 33, 293^302. Underwood, A.J., Chapman, M.G. & Connell, S.D., 2000. Observations in ecology: you can’t make progress without understanding the patterns. Journal of Experimental Marine Biology and Ecology, 250, 97^115.
Journal of the Marine Biological Association of the United Kingdom (2004)
Walker, H.J. ed., 1988. Arti¢cial structures and shorelines. Los Angeles: Kluwer Academic Publishers. Wendt, P.H., Knott, D.M. & Van Dolah, R.F., 1989. Community structure of the sessile biota on ¢ve arti¢cial reefs of di¡erent ages. Bulletin of Marine Science, 44, 1106^1122. Wilkinson, C.R., 1978. Microbial associations in sponges. I. Ecology, physiology and microbial populations of coral reef sponges. Marine Biology, 49, 161^167. Witman, J.D. & Sebens, K.P., 1991. Distribution and ecology of sponges at a subtidal rock ledge in the Central Gulf of Maine. In New perspectives in sponge biology (ed. K. Ru«tzler), pp. 391^396. Washington: Smithsonian Institution Press. Wright, J.T., Benkendor¡, K. & Davis, A.R., 1997. Habitat associated di¡erences in temperate sponge assemblages: the importance of chemical defence. Journal of Experimental Marine Biology and Ecology, 213, 199^213. Young, C.M. & Chia, F.-S., 1984. Microhabitat-associated variability in survival and growth of subtidal solitary ascidians during the ¢rst 21 days after settlement. Marine Biology, 81, 61^68.
Submitted 3 March 2004. Accepted 26 August 2004.
