Appendix A (Table A1). The full MANOVA and univariate ANOVAs for the principal components associated with the morphometric measurements of H. chrysoscelis tadpoles.
A. MANOVA: Principal Components |
||||
Source |
Wilks’ λ |
df |
F |
P |
Treatment |
0.5350 |
9, 362 |
11.81 |
< 0.0001 |
Block |
0.4182 |
6, 298 |
27.13 |
< 0.0001 |
Treatment × block |
0.7705 |
18, 422 |
2.26 |
0.0024 |
B. ANOVA: Overall size (PC 1) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment (tested with T × B MS) |
2.5224 |
3 |
0.83 |
0.5261 |
Block (tested with T × B MS) |
1.7961 |
2 |
0.26 |
0.7752 |
Treatment × block (tested with within tank MS) |
3.0557 |
6 |
0.43 |
0.8553 |
Within tank variation |
7.0406 |
151 |
||
C. ANOVA: Tadpole shape – D vs. L (PC 2) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment |
0.5899 |
3 |
3.56 |
0.0869 |
Block |
4.2552 |
2 |
102.41 |
< 0.0001 |
Treatment × block Within tank variation |
0.1656 0.0415 |
6 151 |
3.99 |
0.001 |
D. ANOVA Tadpole shape – tail vs. body (PC 3) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment |
1.0358 |
3 |
6.66 |
0.0245 |
Block |
0.0090 |
2 |
0.18 |
0.8338 |
Treatment × block |
0.1554 |
6 |
3.15 |
0.0061 |
Within tank variation |
0.0493 |
151 |
Notes: The interaction between treatments and blocks was used as the error term for tests of the significance of the main effect of treatments and blocks (see "B." as example) to prevent pseudoreplication. No morphometric data were used from block 1 (see methods).