Appendix B. The full MANOVA and univariate ANOVAs for the principal components associated with the tail conspicuousness data from H. chrysoscelis tadpoles. 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.
| A. MANOVA: Principal Components | ||||
Source |
Wilks’ λ |
df |
F |
P |
Treatment |
0.6449 |
12, 379 |
5.69 |
< 0.0001 |
Block |
0.9557 |
8, 286 |
0.82 |
0.5845 |
Treatment × block |
0.7466 |
24, 500 |
1.82 |
0.0105 |
B. ANOVA: Overall tail conspicuousness (PC 1) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment (tested with T × B MS) |
28.2830 |
3 |
7.66 |
0.0179 |
Block (tested with T × B MS) |
1.1085 |
2 |
0.54 |
0.5828 |
Treatment × block (tested with within tank MS) |
3.6945 |
6 |
1.81 |
0.1017 |
Within tank variation |
2.0454 |
146 |
||
C. ANOVA: Tail spottiness (mottling) (PC 2) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment |
1.6347 |
3 |
2.66 |
0.1421 |
Block |
0.8019 |
2 |
0.85 |
0.4289 |
Treatment × block Within tank variation |
0.6145 0.9418 |
6 146 |
0.65 |
0.6881 |
D. ANOVA Tail conspicuousness / tail melanin outline (PC 3) |
||||
Source |
Mean Square |
df |
F |
P |
Treatment |
0.6181 |
3 |
1.44 |
0.3212 |
Block |
0.5841 |
2 |
1.84 |
0.1621 |
Treatment × block |
0.4292 |
6 |
1.35 |
0.2372 |
Within tank variation |
0.3170 |
146 |