Ecological Archives E093-099-A2

Michael A. Sams and Michael J. Keough. 2012. Contrasting effects of variable species recruitment on marine sessile communities. Ecology 93:1153–1163.

Appendix B. Details of study site, experimental set up, statistical methods, and supplementary results, including tables of statistical analysis and figures of common species.

Study sites

Workshops Jetty is situated at the northern end of Port Phillip Bay, Australia, near the mouths of the Yarra and Maribyrnong rivers. It is an estuarine environment characterised by high loads of muddy sediment. The sessile community found at Workshops Jetty is a typical 'fouling' community consisting of many invasive species, including colonial ascidians (Botryllus schlosseri, Botrylloides leachii, and Diplosoma listerianum), solitary ascidians (Styela plicata, Styela clava, Ascidiella aspersa, Ciona intestinalis), bryozoans (Watersipora subtorquata, Bugula neritina, Bugula stolonifera, Bugula flabellata), barnacles (Amphibalanus variegatus and Balanus trigonus) serpulids (Hydroides ezoensis, Ficopomantus enigmaticus) and dominated by the large native solitary ascidian Pyura dalbyi. Native species such as the bryozoan Bugula dentata, the barnacle Elminius modestus, and the serpulid Pomatoceros taeniatus as well as didemnid ascidians (one or two spp of ascidian in the family Didemnidae belonging to Tridemnum and Didemnum that are difficult to distinguish from one another, but easily distinguished from Diplosoma listerianum, also in the same family) , whose status as invasive or native is unknown, are also abundant at this site.

St Leonard's Pier is at the southern end of Port Phillip Bay. It is characterised by coarse sand and lower sediment loads in the water. Unlike Workshops Jetty, the sessile community at St Leonards is characterised by mostly native colonial species including large colonial ascidians (Family Didemnidae), large sponges, bryozoans (Bugula dentata, Celleporaria spp. and the likely non-native  Tricellaria occidentalis), and hydroids.

Santa Cruz harbour is a manmade lagoon designed to harbour small vessels. Experiments at Santa Cruz harbour were run from jetty piles closest to the harbour entrance, which is regularly flushed with water from Monterey Bay.  Major space occupiers of the sessile community in this part of the harbour include the encrusting bryozoan Watersipora subtorquata, anemones of the genus Anthopleura, the barnacle Balanus crenatus, the colonial ascidian Botryllus schlosseri as well as large colonial ascidians and sponges.

Experimental setup and details

Experiments were run at three sites Workshops Jetty, Williamstown (2 experiments), St Leonards (1 experiment) and Santa Cruz (1 experiment) using a similar set up at each sites. At each study site Perspex settlement plates were attached to PVC backing panels with stainless steel bolts through the middle of each plate. Panels were suspended by ropes separated by at least 2.5 metres intervals at random location on horizontal beams, and weighted so that the surface bearing plates faced downward at a depth of approximately 2 metres below the low water mark.  Water depths reached approximately 4-5 m at panel locations Workshops Jetty, Williamstown and Santa Cruz, and 4 m at St Leonards. At Williamstown and St Leonards 100 mm × 125 mm plates were attached to 600 mm × 600 mm PVC backing panels designed to hold 16 panels in an evenly spaced 4 × 4 array. However, the actual number of plates on each panel changed between experiments depending on the number of treatments and replicates of each treatment in each experiment. All panels had multiple replicate plates of every treatment, which were placed randomly in any of the 16 available attachment points. At Santa Cruz 100 × 100 mm plates were attached to 600 mm × 600 mm panels with 25 plates per panel arranged in a 5 × 5 evenly spaced array. All panels at Santa Cruz had 5 replicate plates of each of the 5 treatments used. A brief description of each experiment and the allocation of plates to panels are provided below.

Williamstown Experiment 1

Experiment 1 ran from mid Nov 2008 until late May 2009 at Workshops Jetty, Williamstown.  Treatments consisted of the colonial ascidian B. schlosseri at three densities, high (70 individuals), medium (40 individuals) and low (20 individuals), colonial ascidians belonging to the Family Didemnidae (which consisted of one or two very similar species belonging to the genus Didemnum and/or Tridemnumn) at one density, low (20 individuals) and control plates cleared of any initial recruits. Two replicates plates of each treatment were randomly attached to any of the available spots on three PVC backing panels for an overall total of six replicates for each treatment.

Williamstown Experiment 2

Experiment 2 ran from the late December 2008 until early June 2009 at Workshops Jetty. Treatments consisted of the colonial ascidian Botryllus schlosseri at two densities, medium (40 individuals) and low (20 individuals), a mixture of barnacle recruits comprising mostly Amphibalanus variegates, Balanus trigonus at two densities, medium (80 individuals) and low (40 individuals), a mixture of arborescent bryozoans at one density, low (10 individuals) and cleared control plates. Two replicates plates of each treatment were randomly attached to any of the available spots on to four individual PVC backing panels for a total of eight replicates for each treatment

St Leonards

St Leonards ran from early March 2009 until mid October 2009 at St Leonards pier. Treatments consisted of only arborescent bryozoans at two densities, medium (10 individuals) and low (5 individuals), and cleared control plates. Arborescent bryozoan treatments consisted of mostly Tricellaria occidentalis with a few early recruits of Bugula neritina. Two replicate plates of each treatment were randomly attached to any of the available spots on to four individual PVC backing panels for a total of eight replicates for each treatment Santa Cruz

Santa Cruz

Santa Cruz ran from early August 2009 until mid February 2010 at Santa Cruz harbour. Treatments consisted of the colonial ascidian B. schlosseri at two densities, low (20 individuals) and very-low (10 individuals), the encrusting bryozoans W. subtorquata at one density, very-low (5 individuals), and cleared control plates. Five replicates plates of each treatment were attached in randomly allocated position on four individual PVC backing panels for a total of 15 replicates for each treatment.

Statistical methods

To test for changes in overall community structure we used PERMANOVA (add-on for Primer) using the partly-nested models outlined below. To test for the effects of different recruitment patterns on individual taxa we analysed cover data using repeated-measures ANOVA in the software package SYSTAT. The models used in repeated measures ANOVA were analogous to the partly-nested models used in PERMANOVA. A specific repeated measures procedure is not available for PERMANOVA which is why partly-nested models were used. We did not use 'standard' measures of community structure, such as species diversity indices and species richness as we believe two much information is lost in calculating these attributes. It is our opinion that PERMANOVA incorporates more relevant information and captures important changes in the composition of communities that these other measures can easily miss. Arc-sin transformation had little influence on the distribution of data for most of the taxa in this study and we used raw data in all analysis presented.

Due to the vagaries of recruitment it was impossible to set up orthogonal species by density treatments (i.e. not all species and densities were crossed with each other) in all experiments except for the experiments at St Leonards, which had treatments of only one species. Consequently these experiments were analysed individually as incomplete factorial designs that involved two to three separate analysis (in some cases) examining the effects of treatments of either different species recruitment or different densities within species on the structure and development of communities. Where the effects of different species were compared we used treatments that had the same relative density for each species to control for any density effects. Where significant effects of treatments were detected post-hoc unplanned pairwise comparisons were done using either PERMANOVA pairwise comparisons (where PERMANOVA was significant) or using Tukey's HSD at the appropriate factor level (where repeated-measures ANOVA was significant for individual taxa). In all the models outlined below it was possible to include 'panels' as a blocking factor, as there were multiple replicate plates of each treatment per panel. Panels were only a logistical convenience and not of biological interest, therefore to keep our models simpler we did not include it as a factor.

Williamstown Experiment 1

The overall experimental design for Experiment 1 consisted of four factors, Species (fixed at three levels: Botryllus schlosseri, didemnid ascidians or no initial recruits), Density (fixed at three levels for B. schlosseri treatments: high, med, low, but only at one level for didemnid treatments: low) Time (fixed at five levels: 2 weeks, 1 month, 2 months, 3 months, and 6 months) and Plate (random with six levels nested within the appropriate level for the factor of interest; see below)

Experiment 1 was analysed with two separate models. The first of these analyses tested for the effects of different species on community structure by comparing three factors, Species (fixed with three levels: Botryllus schlosseri low, didemnid low and cleared), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6 months) and plate (random with six levels and nested within Species). The second of these analyses tested for the effects of different densities of B. schlosseri by comparing three factors: Density (fixed with four levels: B. schlosseri high density, B. schlosseri medium density, B. schlosseri low density and cleared), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and plate (random with five levels nested within Treatment).

Williamstown Experiment 2

The overall experimental design for Experiment 2 consisted of four factors, Species (fixed with four levels, Botryllus schlosseri, barnacles, arborescent bryozoans and cleared), Density (fixed with two levels for B. schlosseri treatments, medium and low, two levels for barnacles, medium and low and one level for arborescent bryozoans, low), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and plate (random with eight levels and nested within the appropriate level for the factor of interest).

Experiment 2 was analysed with three separate models. The first of these analyses tested for the effects of different species of recruits on community structure by comparing three factors, Species (fixed at four levels: Botryllus schlosseri low, barnacle low, arborescent bryozoan low and cleared), Time (fixed at five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with eight levels nested within Species). The second of these analyses tested for the effects of different densities of B. schlosseri on community structure by comparing three factors, Density (fixed with three levels: B. schlosseri medium density, B. schlosseri low density or cleared), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with eight levels and nested within Density). The third of these analyses tested for the effects of different barnacle densities on community structure by comparing three factors, Density (fixed with three levels: barnacle barnacle medium density, barnacle low density and cleared), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with eight levels and nested within Density).

St Leonards

The experimental design for Experiment 3 consisted of three factors, Treatments (fixed with three levels: arborescent bryozoans medium, arborescent bryozoans low and cleared, Time (fixed with four levels: 1 month, 2 months, 3 months, and 6 months) and Plate (fixed with eight levels and nested within Treatment). Experiment 3 was analysed as a single complete model outlined above with Treatment and Time as the main factors and Plate nested within Treatment.

Santa Cruz

The overall experimental design for Experiment 4 consisted of four factors, Species (fixed with four levels: Watersipora subtorquata, Botryllus schlosseri and no initial recruits), Density (fixed with two levels for B. schlosseri: low and very-low, and one level for W. subtorquata, very-low), Time (fixed with five levels: 2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with 15 levels and nested within the appropriate treatment level for the factor of interest)

Experiment 5 was analysed with two separate models. The first of these analyses tested for the effects of different species on community composition by comparing three factors, Species (fixed with three levels, Botryllus schlosseri very-low density, Watersipora very-low density and cleared), Time (fixed with five levels:  2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with 15 levels and nested within Species). The second analyses tested for the effects of different densities of B. schlosseri on community composition by comparing three factors, Density (fixed with three levels: B. schlosseri low, B. schlosseri very-low and cleared), Time (fixed with five levels 2 weeks, 1 month, 2 months, 3 months, and 6  months) and Plate (random with 15 levels and nested within Density).

Supplementary Results

Table B1. Results from repeated-measures ANOVA on the main space occupying taxa and partly-nested PERMANOVA on all 42 taxa colonising settlement plates throughout Experiment 1, Williamstown. The first analysis (a) compares the % cover of taxa between treatments that received recruitment of either Botryllus schlosseri at low densities, didemnid ascidians at low densities or no recruits during the initial stages of recruitment. The second analysis (b) compares the % cover of taxa between treatments that received B. schlosseri in high, medium or low densities or no recruits during the initial stages of recruitment. % cover was measured at 2 weeks, 1, 2, 3 and 6 months after manipulations of recruitment. Numbers underneath main factor headings are P values, while values under MS (btwn) are the mean square values between subjects (plates) and values under MS(wthn) are the mean square values within subjects (plates) for each analysis.  Where there were significant differences in the cover of taxa or the overall structure of communities between treatments results of pairwise post-hoc comparisons are shown below. Letters denote treatments types in each pairwise comparison and where there was an interaction with time semicolons separate each time plates were sampled in order from left to right (see key).

  (a) (b)
Spp Time S × T MS
(btwn)
MS
(wthn)
BDen Time BD
× T
MS
(btwn)
MS
(wthn)
d.f. 2 4 8 14 56 3 4 12 20 80
Botryllus schlosseri 0.003 0.000 0.008 10.422 13.715 0.036 0.000 0.002 19.485 16.487 
Didemnid ascidians 0.061 0.000 0.010 32.646 6.731 0.680 0.000 0.208 24.349 7.472
Diplosoma listerianum 0.365 0.001 0.957 74.702 70.958 0.385 0.001 0.860 53.450 19.523
Watersipora subtorquata 0.583 0.000 0.863 54.907 16.566 0.436 0.000 0.837 68.458 18.297
Bugula dentata 0.346 0.000 0.413 95.864 48.846 0.739 0.000 0.536 47.082 80.513
Bugula stolonifera 0.430 0.001 0.834 13.692 1.563 0.939 0.000 0.985 17.025 18.621
Cyanea capillata 0.546 0.003 0.393 70.492 26.241 0.378 0.000 0.201 14.338 39.596
Amphipod tubes 0.641 0.000 0.371 73.064 33.389 0.827 0.000 0.353 70.983 31.655
PERMANOVA 0.058 0.000 0.000 2188.9 1209 0.2516 0.000 0.005 1912.3 1059.8
  Pair-wise comparisons
for species effects
Pair-wise comparisons for
Botryllus density effects
Botryllus schlosseri B > D B > C D = C;
B > D B = C D = C;
B = D = C;
B = D = C;
B = D = C
BH > C BM > C BL = C BH = BM = BL;
BH = BM = BL = C;
BH = BM = BL = C;
BH = BM = BL = C;
BH = BM = BL = C
Didemnid ascidians D > B D > C B = C;
D = B = C;
D = B = C;
D > C D = B C = B;
D = B = C
 
PERMANOVA B ≠ D ≠ C;
B = C = D;
B = C = D;
B = C = D;
B = C = D
BH > C BM > C BL = C BH = BM = BL;
BH = BM = BL = C;
BH = BM = BL = C;
BH = BM = BL = C;
BH = BM = BL = C

*B = Botryllus low, D = Didemnid, low C = No recruitment
+BH = Botryllus high, BM = Botryllus medium, BL = Botryllus low, C = No recruitment


Table B2. Results from repeated measures ANOVA on the main space occupying taxa colonising settlement plates and partly-nested PERMANOVA on all 42 taxa colonising plates during Experiment 2, Williamstown. The first analysis (a) compares the % cover of taxa between treatments that initially received recruits of either Botryllus schlosseri at low density, barnacles at low density, arborescent bryozoans at low density or no recruits . The second analysis (b) compares the % cover of main space occupying taxa between treatments that initially received B. schlosseri recruits in medium or low density or no recruits. The third analysis (c) compares the % cover of main space occupying species between treatments that initially received recruits of barnacles in medium and low density or no recruits. % cover was measured at 2 weeks, 1, 2, 3 and 6 months after manipulations of recruitment. Numbers underneath main factor headings are P values, while values under MS (btwn) are the mean square error values between subjects (plates) and values under MS (within) are the mean square error values within subjects (plates) for each analysis.  Where there were significant differences in the cover of taxa or the overall structure of communities between treatments results of pairwise post-hoc comparisons are shown below. Letters denote treatments types in each pairwise comparison and where there was an interaction with time semicolons separate each time plates were sampled in order from left to right (see key).

  (a) (b) (c)
Spp Time S × T MS
(btwn)
MS
(wthn)
BDen Time BD
× T
MS
(btwn)
MS
(wthn)
BDen Time BD
× T
MS
(btwn)
MS
(wthn)
d.f. 3 4 12 20 80 2 4 8 15 60 2 4 8 15 60
Botryllus schlosseri 0.166 0.000 0.001 18.578 13.272 0.170 0.000 0.051 39.400 45.628 0.507 0.000 0.039 9.418 6.191
Didemnid 0.374 0.000 0.815 144.382 39.793 0.702 0.000 0.947 167.550 55.310 0.127 0.000 0.581 91.633 32.894
Diplosoma listerianum 0.788 0.000 0.986 70.353 48.406 0.697 0.000 0.974 79.040 59.477 0.543 0.000 0.344 68.143 49.104
Watersipora subtorquata 0.498 0.000 0.546 332.248 0.863 0.280 0.000 0.370 277.290 80.482 0.031 0.000 0.001 238.874 67.330
Amphibalanus variegatus 0.396 0.000 0.639 3.387 2.058 0.480 0.000 0.464 5.038 2.097 0.812 0.000 0.348 1.661 0.773
Balanus trigonus 0.000 0.006 0.023 1.084 0.721 0.065 0.010 0.595 1.013 0.572 0.003 0.007 0.204 2.774 1.423
Ficopomatus enigmaticus 0.070 0.000 0.961 1.986 2.893 0.044 0.000 0.909 1.955 3.369 0.149 0.040 0.378 2.487 2.970
Medium serpulids 0.223 0.002 0.940 8.954 2.838 0.232 0.094 0.924 11.470 2.817 0.253 0.017 0.974 10.293 3.094
Cyanea capillata 0.591 0.000 0.816 4.408 3.250 0.619 0.000 0.506 5.535 3.237 0.077 0.000 0.353 4.366 4.659
Amphipod tubes 0.253 0.000 0.744 39.460 17.999 0.336 0.000 0.520 53.715 19.712 0.058 0.000 0.005 57.185 15.166
PERMANOVA 0.105 0.0002 0.907 1371.6 650.87 0.178 0.000 0.296 1350.1 733.53 0.000 0.000 0.027 1326.7 624.58
  Pair-wise comparisons
for species effects*
Pair-wise comparisons for
Botryllus density effects+
Pair-wise comparisons for
barnacle density effects#
Botryllus schlosseri B > C B > A B = Ba Ba = C = A;
B = Ba = A = C;
B = Ba = A = C;
B = Ba = A = C;
B = Ba = A = C
  BaM > C BaM = BaL BaL = C;
BaM = BaL = C;
BaM = BL = C;
BaM = BL = C;
BM = BL = C;
BM = BL = C
Balanus trigonus Ba > B Ba > A Ba > C B = A = C;
Ba > B Ba > A Ba > C B = A = C;
B = Ba = A = C;
B = Ba = A = C;
Ba > C Ba = A Ba = B B = A = C
  BM > C BM = BL BL = C;
BM > C BL > C BM = BL;
BM > C BM = BL BL = C;
BM = BL = C;
M = BL = C
Ficopomatus enigmaticus   BL > C BL = BM BM = C  
Watersipora subtorquata     BM = BL = C;
BM = BL = C;
BM < C BM = BL BL = C;
BM < C BM = BL BL = C;
BM = BL = C
Amphipod tubes     BM = BL = C;
BM = BL = C;
BM > C BM = BL BL = C;
BM > C BM > BL BL = C;
BM = BL = C
PERMANOVA     BM ≠ C BM = BL BL = C;
BM = BL = C;
BM ≠ C BM ≠ BL BL = C;
BM ≠ C BM ≠ BL BL = C;
BM = BL = C

*B = Botryllus low, Ba = Barnacle Low, A = Arborescent bryozoan, C = No recruitment;
+BM = Botryllus medium, BL = Botryllus low, C = No recruitment;
#BaM = barnacle medium, BaL = barnacle, low C = No recruitment


Table B3. Results from repeated-measures ANOVA on the main space occupying taxa and partly-nested PERMANOVA on all 41 species/taxa colonising settlement plates throughout Experiment 3, St Leonards. The analysis compares the % cover of taxa between treatments that initially received recruits of arborescent bryozoans at medium and low densities or received no initial recruits. % cover was measured 1, 2, 3 and 6 months after manipulations of recruitment. Numbers underneath main factor headings are P values, while values under MS (btwn) are the mean square error values between subjects (plates) and values under MS (within) are the mean square error values within subjects (plates) for each analysis.

  Treatment Time Treat
× Time
MS
(btwn)
MS
(wthn)
d.f. 2 3 6 21 63
Tricellaria  occidentalis 0.247 0.000 0.874 185.667 734.080
Bugula dentata 0.751 0.000 0.760 79.953 22.350
Biflustra perfragilis 0.519 0.000 0.121 1.885 2.535
Hydroid 0.348 0.000 0.807 206.798 108.601
Didemnids 0.758 0.000 0.950 12.973 9.251
Clear Tube Sponge 1 0.534 0.000 0.641 1.515 0.858
PERMANOVA 0.220 0.000 0.403 2581.8 1269.2

Table B4. Results from repeated-measures ANOVA on the main space occupying taxa and partly-nested PERMANOVA on all 32 taxa colonising settlement plates throughout Experiment 4, Santa Cruz. The first analysis (a) compares the % cover of taxa between treatments that received recruitment of either Botryllus schlosseri at very-low densities, Watersipora subtorquata at very-low densities or no recruits during the initial stages of recruitment. The second analysis (b) compares the % cover of taxa between treatments that received Botryllus schlosseri in low or very-low densities or no recruits during the initial stages of recruitment. % cover was measured at 2 weeks, 1, 2, 3 and 6 months after manipulations of recruitment. Numbers underneath main factor headings are P values, while values under MS (btwn) are the mean square error values between subjects (plates) and values under MS(within) are the mean square error values within subjects (plates) for each analysis.  Where there were significant differences in the cover of taxa or the overall structure of communities between treatments results of pairwise post-hoc comparisons are shown below. Letters denote treatments types in each pairwise comparison and where there was an interaction with time semicolons separate each time periods plates were sampled in order from left to right (see key).

  (a) (b)
Spp Time S × T MS
(btwn)
MS
(wthn)
BDen Time BD
× T
MS
(btwn)
MS
(wthn)
d.f. 2 4 8 27 108 2 4 8 27 108
Ascidia ceratodes 0.503 0.000 0.712 15.944 5.444 0.325 0.000 0.225 23.192 8.965
Botryllus schlosseri 0.000 0.000 0.000 62.826 18.018 0.000 0.000 0.000 127.555 48.366
Balanus crenatus 0.033 0.000 0.000 10.445 5.418 0.966 0.000 0.039 29.554 11.105
Tricellaria porteri 0.297 0.000 0.555 5.615 2.063 0.084 0.000 0.002 10.938 3.604
Watersipora subtorquata 0.000 0.000 0.000 102.386 42.330 0.259 0.000 0.011 67.725 26.756
Hydroids 0.085 0.000 0.673 26.127 20.312 0.793 0.000 0.246 0.230 0.337
Pale sponge 0.026 0.000 0.012 0.817 0.337 0.789 0.000 0.166 1.039 0.927
PERMANOVA 0.000 0.000 0.000 1980.04 1100.2 0.001 0.001 0.001 2186.8 1090.4
  Pair-wise comparisons
for species effects*
Pair-wise comparisons for
Botryllus density effects+
Botryllus schlosseri B > W B > C W = C;
B > W B > C W = C;
B > W B > C W = C;
B > W B > C W = C;
B > W B > C W = C
BM = BL > C;
BM = BL > C;
BM = BL > C;
BM = BL > C;
BM = BL = C
Balanus crenatus B = W = C;
B = W = C;
B = W = C;
B = W = C;
B > W B = C W = C
BM = BL = C;
BM = BL = C;
BM = BL = C;
BM = BL = C;
BM = BL > C
Tricellaria porteri   BM = BL = C;
BM > C = BL BM = BL;
BM = BL = C;
BM = BL = C;
BM = BL = C
Watersipora subtorquata W > B W > C B = C;
W > B W > C B = C;
W > B W > C B = C;
W > B W > C B = C;
W > B W > C B = C
BM = BL = C;
BM ≠ BL = C BM = C;
BM = BL = C;
BM = BL = C;
BM = BL = C
Pale sponge B = W = C;
B = W = C;
B = W = C;
B > W = C;
B = W = C
 
PERMANOVA B ≠ W B ≠ C W = C;
B ≠ W ≠ C;
B ≠ W ≠ C;
B ≠ W ≠ C;
W ≠ B W ≠ C B = C
BM = BL ≠ C BM ≠ C;
BM = BL ≠ C BM ≠ C;
BM = BL ≠ C BM ≠ C;
BM= BL ≠ C BM ≠ C;
BM = BL ≠ C BM = C

*W = Watersipora low, B = Botryllus low, C = No recruitment
+ BM = Botryllus medium BL = Botryllus low C = No recruitment


Fig. B1. % cover of major space occupying taxa found on settlement plates exposed to initial recruitment of either B. schlosseri in high, medium  or low densities, didemnid ascidians at low densities or no initial recruitment at Williamstown Experiment 1 at samples taken 2 weeks, 1, 2, 3 and 6 months after manipulations. For details of analysis see Table B1.



Fig. B2. % cover of other common taxa found on settlement plates exposed to initial recruitment of either B. schlosseri in medium or low density, a mixture of barnacles in medium or low density, a mixture of arborescent bryozoans at low density or no initial recruitment at Williamstown Experiment 2 at samples taken 2 weeks, 1, 2, 3 and 6 months after manipulations. Asterisks denote times when there were differences in cover between treatments according to repeated-measures ANOVA. For details see pairwise comparison on Table B2.



Fig. B3. % cover of the 4 most abundant taxa found on settlement plates exposed to initial recruitment of arborescent bryozoans in medium or low density, or no initial recruitment at Experiment 3 (St Leonards) from samples taken 1, 2, 3 and 6 months after manipulations. For details of analysis see Table B3.



Fig. B4. % cover of free space measured for each experiment in this study (see figures of individual taxa for each respective experiment for experimental details).


[Back to E093-099]