Table 1. ANOVAs on the effects of Habitat (Intertidal vs Subtidal), Region (EA vs SA), Location, Grazers (Present, Removed and Procedural control) and Nutrients (Enhanced vs Ambient) on the amount of unoccupied space (%), percentage cover of algal turfs and foliose macroalgae, 17 weeks after the experiment was started.
Unoccupied space | Algal turfs | Foliose macroalgae | |||||
Source of variation | df | MS | F | MS | F | MS | F |
Habitat = H | 1 | 4657.41 | 149.98*** | 78940.45 | 52.90 | 821.40 | 1.27 |
Region = R | 1 | 36.62 | 0.20 | 28181.11 | 3.61 | 2522.02 | 3.91 |
Location (R) = L(R) | 2 | 186.56a | 8.94*** | 7806.82 | 19.35*** | 645.41 | 4.36* |
Grazers = G | 2 | 1635.98 | 35.71** | 15789.64 | 18.04* | 3469.76 | 22.72** |
Nutrients = N | 1 | 12.72 | 0.16 | 254.89 | 0.25 | 308.27 | 1.14 |
H×R | 1 | 4710.42 | 151.68** | 44154.98 | 29.59* | 1392.02 | 2.15 |
H×L(R) | 2 | 31.05a | 1.49 | 1492.25 | 3.70* | 647.04 | 4.37* |
H×G | 2 | 82.59 | 1.08 | 1190.89 | 1.58 | 67.76 | 0.61 |
H×N | 1 | 26.83 | 0.43 | 222.98 | 0.78 | 504.60 | 5.45 |
R×G | 2 | 489.84 | 10.69* | 2672.99 | 3.05 | 1182.88 | 7.75* |
R×N | 1 | 4.61 | 0.06 | 1341.11 | 1.30 | 43.35 | 0.16 |
G×L(R) | 4 | 45.82a | 2.20 | 875.06 | 2.17 | 152.72 | 1.03 |
N×L(R) | 2 | 81.31a | 3.90* | 1033.38 | 2.56 | 270.01 | 1.82 |
G×N | 2 | 90.85 | 1.57 | 1540.87 | 1.92 | 690.40 c | 4.66* |
H×R×G | 2 | 115.50 | 1.51 | 446.37 | 0.59 | 2383.88 | 21.33** |
H×R×N | 1 | 33.94 | 0.55 | 363.42 | 1.27 | 74.82 | 0.81 |
G×H×L(R) | 4 | 76.33 a | 3.66** | 754.07 | 1.87 | 111.75 | 0.75 |
H×L(R)×N | 2 | 62.02a | 2.97 | 285.21 | 0.71 | 92.64 | 0.63 |
H×G×N | 2 | 29.85a | 1.43 | 467.18 | 1.16 | 270.24 | 1.12 |
R×G×N | 2 | 112.78 | 1.96 | 2356.81 | 2.93 | 18.24 | 0.12 |
N×G×L(R) | 4 | 57.68 a | 2.76* | 804.28 | 1.99 | 138.12‡ | eliminated |
H×R×G×N | 2 | 90.97 a | 4.36* | 2166.67 b | 5.37** | 11.00 | 0.05 |
H×L(R)×G×N | 4 | 18.96‡ | pooled | 569.39† | eliminated | 240.35 | 1.62 |
Residual | 192 | 20.91 | 403.52 | 148.02 |
Pooling procedures were used according to Underwood (1997). * = P<0.05; ** = P<0.01, *** = P<0.001. Higher order interactions relevant for testing proposed hypotheses are reported in bold.
Tested against the pooled term: Residual+H×L(R)×G×N (df = 196; MS = 20.87).
Tested against the Residual;
not significant at P = 0.25;
not significant at P = 0.23.