114 G
.A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
determined by non-metric multidimensional scaling nMDS on unstandardised fourth root transformed data, using the Bray–Curtis similarity measure. The statistical
significance of differences among locations was assessed using analysis of similarities ANOSIM.
3. Results
3.1. Structure of the habitat The area around the boardwalk and the path supported 65–80 fewer mangrove
pneumatophores than found in the control locations Table 1; Fig. 2A. There were no significant differences in the number of pneumatophores between the forest and edge
habitats, nor were any of the interactions involving habitat significant Table 1. There was very little variation in the number of pneumatophores between the two control
locations Table 1; Fig. 2A, but there was significant variation at the spatial scales of sites within the boardwalk path location and the controls.
Although there were fewer saplings found in the proximity of the boardwalk and path than in the controls, these differences were not detected as being significant Tables 2
and 3. There was substantial variation in the number of saplings at the spatial scale of sites within locations, and also among the control locations Fig. 2B. The biomass of
leaf litter on the substratum near the boardwalk and the path was very similar to the controls Fig. 2Ci, ii. Again, there was substantial variation in the biomass of leaf litter
at the spatial scales of sites and among control locations Tables 2 and 3.
Table 1
2
Example of asymmetrical analysis of variance of the number of pneumatophores in 0.1-m quadrats from two control locations and the boardwalk location. The complete analysis is shown here to illustrate the full design
for the sampling programme. Data were untransformed; n 5 4 quadrats in each site. There were three independent sites sampled in each combination of habitat forest versus edge and location boardwalk versus
controls
Source of variation df
SS MS
F-ratio Sign.
F df Habitat
1 760.50
760.50 0.50
0.55 1,2
Locations 2
31 077.03 15 538.51
b
B walk vs. among controls 1
30 829.51 30 829.51
28.36 0.0005
1,9 Among controls
1 247.52
247.52 0.21
0.66 1,12
Habitat 3 locations 2
3055.08 1527.54
a
H 3 b walk vs. am. cont. 1
3052.56 3052.56
2.88 0.12
1,9 H 3 among controls
1 2.52
2.52 0.00
0.96 1,8
Sites H 3 locations 12
11 937.33 994.78
0.11 1.00
12,54 Sites H 3 b walk
4 2400.17
600.04 3.67
0.01 4,54
Sites H 3 among controls 8
9537.17 1192.15
7.30 0.00
8,54 Residual
54 8820.50
163.34
a
Tested against pooled term: [habitat 3 among controls] 1 [sites habitat 3 among controls] df 5 9, SS 5 9539.69, MS 5 1059.97.
b
Tested against pooled term: [among controls] 1 [sites habitat 3 among controls] df 5 9, SS 5 9784.69, MS 5 1087.19.
G .A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
115
Fig. 2. Mean 1 S.E. number of pneumatophores, saplings, and the biomass of leaf-litter at the boardwalk and path locations and the two control locations. Asterisks indicate significant differences between disturbed
location and the average of the controls P , 0.05, ns indicates no significant difference P . 0.05. N 5 24 quadrats, pooled across three sites and two habitats for comparisons with the boardwalk location; N 5 12
quadrats, pooled across three sites for comparisons with the path location.
3.2. Mollusc assemblage Significantly fewer species of molluscs and individuals total were found in the
sediments near the boardwalk and path than in the control locations Tables 2 and 3; Fig. 3A and B. Of the species of molluscs which were sufficiently abundant to analyse
separately, Salinator solida von Martens, 1878 was significantly less abundant near the boardwalk but not the path, compared with the controls Tables 2 and 3; Fig. 3Ci, while
the reverse was true for Melosidula zonata H. and A. Adams, 1855 Tables 2 and 3;
116 G
.A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129 Table 2
Summaries of asymmetrical analyses of variance on the number of saplings and biomass of leaf litter, and the number of species of molluscs, total number of individuals, number of Modiolus peronianus, Melosidula
2
zonata and Salinator solida in 0.1-m quadrats from two control locations and the boardwalk location. Data for total number of individuals were transformed to log x. Other details as in Table 1
e
Source of variation Saplings
Biomass Species
Individuals Modiolus
Melosidula Salinator
Habitat P . 0.355
P . 0.515 P . 0.990
P . 0.468 P . 0.376
P . 0.390 P . 0.817
Locations
a
B walk vs. among controls P . 0.373
P . 0.770 P , 0.004
P , 0.001 P . 0.452
P . 0.172 P , 0.053
]] ]]
]] Among controls
P , 0.069 P . 0.136
P . 0.310 P . 0.507
P . 0.150 P . 0.812
P . 0.403 ]]
Habitat 3 locations H 3 b walk vs. am. cont.
P . 0.124 P . 0.647
P . 0.592 P . 0.942
P . 0.208 P . 0.401
P . 0.466 H 3 among controls
P . 0.727 P , 0.056
P . 0.366 P , 0.036
P . 0.529 P . 0.771
P . 0.325 ]]
]] Sites H 3 loc
Sites H 3 b walk P . 0.773
P , 0.001 P , 0.003
P , 0.001 P . 0.875
P , 0.061 P . 0.724
]] ]]
]] ]]
Sites H 3 am. controls P , 0.001
P . 0.117 P , 0.003
P , 0.044 P , 0.004
P . 0.1245 P , 0.014
]] ]]
]] ]]
]] Residual
a
Tested against pooled term: [among controls] 1 [sites habitat 3 among controls]
Fig. 3Cii. There were no significant differences in the abundance of the mytilid bivalve, Modiolus peronianus Laseron, 1956, between the boardwalk and the controls Tables 2
and 3. The composition of the molluscan assemblage in the forest habitat near the boardwalk
and the path was significantly different from the control locations, with the latter not being different from each other ANOSIM: boardwalk vs. controls P , 0.001; path vs.
controls P , 0.0001; Fig. 4A and C. The composition of the molluscan assemblage in the edge habitat near the boardwalk was different from each of the controls, but the two
control locations also differed from each other ANOSIM: P , 0.0001; Fig. 4B, suggesting these differences are not correlated with the presence of the boardwalk
through the edge habitat.
In all cases, the abundance of the pneumatophores alone provided the optimal
Table 3 Summaries of asymmetrical analyses of variance on the number of saplings and biomass of leaf litter, and the
number of species of molluscs, total number of individuals, number of Modiolus peronianus and Melosidula
2
zonata in 0.1-m quadrats from two control locations and the path location. Data were untransformed; n 5 4 quadrats in each site. There were three independent sites sampled in each location path versus controls. Only
the forest habitat was sampled see text for further details Source of variation
Pneumat. Saplings
Biomass Species
Individuals Modiolus
Melosidula Locations
Path vs. among controls P , 0.008
P . 0.187 P . 0.534
P , 0.005 P , 0.077
P . 0.129 P , 0.001
]] ]]
]] ]]
Among controls P . 0.837
P . 0.262 P , 0.059
P . 0.906 P . 0.313
P . 0.272 P . 0.255
]] Sites H 3 loc.
Sites H 3 path P . 0.161
P . 0.122 P . 0.182
P , 0.005 P . 0.839
P . 0.978 P . 0.901
]] Sites H 3 am. controls
P , 0.001 P , 0.019
P , 0.085 P , 0.061
P , 0.034 P , 0.005
P . 0.537 ]]
]] ]]
]] ]]
Residual
G .A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
117
Fig. 3. Mean 1 S.E. number of species of mollusc, individual molluscs, Salinator solida and Melosidula zonata at the boardwalk and or path locations and the two control locations. Other details as in Fig. 2.
correlation between
the environmental
data i.e.
numbers of
saplings and
pneumatophores and biomass of litter and the rank similarities underlying the nMDS based on the molluscan assemblage BIOENV:
r 5 0.49 — forest habitat, boardwalk
w
vs. controls; r 5 0.26 — edge habitat, boardwalk vs. controls; r 5 0.38 — forest
w w
habitat, path vs. controls. These results suggest the model that the significant differences in the number of pneumatophores between the boardwalk path location and the controls
was responsible for the observed differences in the abundance and structure of the molluscan assemblage between the treatments. This was examined in the manipulative
experiment described below Section 3.4.
118 G
.A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
Fig. 4. nMDS ordinations on fourth root transformed species abundance data from quadrats sampled at either the boardwalk or path locations and the two control locations. There were n 5 4 replicates from each of three
sites in the forest and edge habitats boardwalk comparisons or forest habitat path comparisons.
3.3. Crab assemblage There were significantly more crabs caught per trap per 24 h trapping period near
the boardwalk than in the control locations Table 4, although the number of crabs caught near the boardwalk was generally small Fig. 5A. The number of crabs varied
significantly between the forest and edge habitats and between the two times of sampling, but there were no significant interactions between habitat nor time and the
comparison of interest, the boardwalk versus the average of the controls Table 4.
Parasesarma erythrodactyla Hess, 1865 and Helograpsus haswellianus were the most abundant of the species caught in the traps, comprising 24 and 52 of the total
number of individuals, respectively. No significant treatment effects i.e. boardwalk vs.
G .A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
119 Table 4
Example of asymmetrical analysis of variance of the number of crabs in pitfall traps per 24-h period from two control locations and the boardwalk location on two separate occasions. The complete analysis is shown here
to illustrate the full design for the sampling programme. Data were untransformed; n 5 5 traps in each site. There were three independent sites sampled in each combination of habitat forest versus edge and location
boardwalk versus controls
Source of variation df
SS MS
F-ratio Sign.
F df Time
1 53.35
53.36 22.18
0.042 1,2
Habitat 1
37.36 37.36
3.99 0.057
1,24 Locations
2 60.08
30.04
b
B walk vs. among controls 1
54.445 54.45
5.17 0.036
1,1 ]
Among controls 1
5.633 5.63
0.52 0.481
1,16 Time 3 habitat
1 5.69
5.69 1.61
0.333 1,2
Time 3 locations 2
4.81 2.41
a
Time 3 b walk vs. am. cont. 1
0.01 0.01
0.00 0.974
1,17 Time 3 among cont.
1 4.80
4.80 0.44
0.515 1,16
Habitat 3 locations 2
59.81 29.91
H 3 b walk vs. am. cont. 1
27.78 27.78
0.87 0.523
1,1 H 3 among controls
1 32.03
32.03 2.95
0.105 1,16
Time 3 habitat 3 locations 2
7.08 3.54
a
Tm 3 H 3 b walk vs. am. cont. 1
5.88 5.88
0.57 0.460
1,17 Tm 3 H 3 among controls
1 1.20
1.20 0.11
0.744 1,16
Sites Tm.H.L 24
224.67 9.36
2.79 0.0001
24,144 Sites Tm.H.Lb walk
8 51.20
6.40 1.91
0.063 8,144
Sites Tm.H.Lam. cont. 16
173.47 10.84
3.23 0.0001
16,144 Residual
144 482.80
3.35
a
Tested against pooled term: [Tm 3 H 3 among controls] 1 [sites Tm 3 H 3 among controls] df 5 17, SS 5 174.667, MS 5 10.275.
b
Tested against pooled term: [among controls] 1 [sites Tm 3 habitat 3 among controls] df 5 17, SS 5 179.100, MS 5 10.535.
among controls, or associated interactions were detected for P . erythrodactyla Tables 5
and 6. There was a significant time 3 habitat 3 treatment interaction detected for H .
haswellianus Tm 3 H 3 b walk vs. am. controls; Tables 5 and 6, resulting from significantly more crabs being caught near the boardwalk than the controls in the edge
habitat but not in the forest habitat at time 1, and no differences between treatments at time 2 Fig. 5B.
There were significantly more species of crabs caught near the path than in the controls at time 1, but the reverse was true for time 2 time 3 path vs. am. controls;
Table 6; Fig. 6A. There were substantially more individuals caught near the path than in the controls at time 1 Fig. 6B but there was no difference between the treatments at
time 2 Fig. 6B, explaining the significant interaction time 3 path vs. among controls; Table 6. An identical result was obtained for Helograpsus haswellianus Whitelegg,
1889 Table 6, which comprised 86 of the individuals caught near the path.
3.4. Experimental manipulation of habitat As expected, the number of pneumatophores was significantly less in the
pneumatophore removal plots, compared with all other treatments, but reduction of the
120 G
.A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
Fig. 5. A Mean 1 S.E. number of crabs at the boardwalk and control locations. N 5 60 traps, pooled across two times, three sites and two habitats. B Mean 1 S.E. number of Helograpsus haswellianus in the
forest and edge habitats at the boardwalk and control locations on two different occasions N 5 15 traps, pooled across three sites. Other details as in Fig. 2.
abundance of pneumatophores and the attached epiphytic algae had no effect on the biomass of leaf litter within the plots. Visual inspection confirmed that the algae had not
recolonised the cleared pneumatophores by the end of the experiment. There were | 83 fewer molluscs in the pneumatophore and algal removal plots than
G .A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
121 Table 5
Summaries of asymmetrical analyses of variance of the number of species of crabs, total number of individuals and numbers of Parasesarma erythrodactyla and Helograpsus haswellianus in pitfall traps per 24 h period
from two control locations and the boardwalk location on two separate occasions. Other details as in Table 4
Source of variation Species
Parasesarma Helograpsus
Time P , 0.061
P . 0.258 P . 0.377
]]] Habitat
P , 0.077 P , 0.068
P . 0.166 ]]]
]]] Locations
B walk vs. among controls P . 0.425
P . 0.154 P . 0.212
Among controls P , 0.035
P . 0.832 P . 0.193
]]] Time 3 habitat
P , 0.026 P . 0.755
P . 0.377 ]]]
Time 3 locations Time 3 b walk vs. am. cont.
P . 0.331 P . 0.675
P , 0.001 ]]]
Time 3 among cont. P . 0.652
P . 0.330 P . 0.795
Habitat 3 locations H 3 b walk vs. am. cont.
P . 0.454 P . 0.546
P . 0.146 H 3 among controls
P , 0.084 P . 0.171
P . 0.193 ]]]
Time 3 habitat 3 locations Tm 3 H 3 b walk vs. am. cont.
P . 0.990 P . 0.587
P , 0.001 ]]]
Tm 3 H 3 among controls P . 0.763
P . 0.832 P . 0.795
Sites Tm.H.L Sites Tm.H.Lb walk
P . 0.165 P , 0.001
P , 0.001 ]]]
]]] Sites Tm.H.Lam. cont.
P , 0.001 P , 0.022
P . 0.981 ]]]
]]] Residual
in the control plots Fig. 7, indicating a marked response by the molluscs to the alterations in the physical structure of the mangrove forest. There was a slight, but not
significant, decline in the number of individuals in the ‘pneumatophore disturbance’ and ‘algal disturbance control’ plots compared with the ‘undisturbed control’ plots Table 7;
SNK tests, Fig. 7. There was no significant difference in the abundance of molluscs in the pneumatophore and algal removal plots suggesting any effect on the molluscs was
Table 6 Summaries of asymmetrical analyses of variance of the number of species of crabs, total number of individuals
and numbers of Parasesarma erythrodactyla and Helograpsus haswellianus in pitfall traps per 24-h period from two control locations and the path location on two separate occasions. Other details as in Table 4
Source of variation Species
Individuals Parasesarma
Helograpsus Time
P . 0.322 P . 0.382
P . 0.239 P . 0.423
Locations Path vs. among controls
P . 0.238 P , 0.001
P . 0.179 P , 0.001
]]] ]]]
Among controls P . 0.990
P . 0.990 P . 0.990
P . 0.990 Time 3 locations
Time 3 path vs. am. cont. P , 0.012
P , 0.001 P , 0.073
P , 0.001 ]]]
]]] ]]]
]]] Time 3 among cont.
P . 0.919 P . 0.916
P . 0.552 P . 0.968
Sites Tm.L Sites Tm.Lpath
P , 0.038 P , 0.069
P . 0.263 P . 0.167
]]] ]]]
Sites Tm.Lam. cont. P , 0.003
P . 0.982 P , 0.004
P . 0.990 ]]]
]]] Residual
122 G
.A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
Fig. 6. Mean 1 S.E. number of species of crab and number of individual crabs at the path and control locations on two different occasions. N 5 15 traps, pooled across three sites. Other details as in Fig. 2.
primarily related to removal of the epiphytic algae from the surface of the pneumatophores.
These results were generally repeated when considering most of the individual taxa sufficiently abundant for univariate analysis. The numbers of Salinator solida, Assiminea
tasmanica Tenison-Woods, 1876 and Modiolus peronianus were all significantly reduced in the pneumatophore and algal removal plots compared with the undisturbed
control plots, but there was no significant difference in the abundance of these species between the two types of removals Fig. 7; Table 7. Again, there was some indication
of a small, experimental artifact arising from manipulations to the experimental plots. In the case of the mussel, Modiolus peronianus, there was a significant artifact associated
with the algal disturbance control Fig. 7. There was a significant treatment 3 site interaction detected for the snail, Melosidula zonata, arising because of the pronounced
and significant experimental artifacts detected at one of the two sites but not the other Fig. 7; Table 7. The second of the two sites showed an identical pattern of decline in
the removal treatments as described for the other species.
The effects of the manipulations were less clear when examining the structure of the entire molluscan assemblage using nMDS and ANOSIM analyses. Although some of the
treatments were significantly different from each other P , 0.03 — ANOSIM; nMDS;
G .A. Skilleter, S. Warren J. Exp. Mar. Biol. Ecol. 244 2000 107 –129
123
Fig. 7. Mean 1 S.E. number of species of mollusc, individual molluscs, Salinator solida, Assiminea tasmanica, Modiolus peronianus and Melosidula zonata in each of five different experimental treatments
unmanipulated controls, pneumatophore disturbance controls, algal disturbance controls, pneumatophore removals and algal removals. N 5 8 plots, pooled across two sites, except for Melosidula zonata where N 5 4
plots at each site. Letters above the histograms indicate groups of means identified as significantly different P , 0.05 by SNK tests; treatments with the same letter were not significantly different P . 0.05.
Fig. 8, the overall pattern suggested that any effects were primarily due to artifacts arising from the experimental manipulations of the pneumatophores and epiphytic algae.
4. Discussion
