Ecological Archives E096-168-A5
Kunio Takatsu and Osamu Kishida. 2015. Predator cannibalism can intensify negative impacts on heterospecific prey. Ecology 96:1597–1608. http://dx.doi.org/10.1890/14-1616.1
Appendix E. Summary of statistical analyses comparing the large-variation and large-variation non-giant-removal treatments.
We established the large-variation giant-removal and large-variation non-giant-removal treatments as additional treatments in the second step. Because we expected that the occurrence of salamander giants would be key to establishing predator-prey interactions between the salamanders and the frog tadpoles, we conducted preliminary analyses of the data of the two treatments including salamander giants (large-variation and large-variation non-giant-removal treatments) to determine whether for any measured variables the results of the two treatments were similar. We tested whether (1) mortality of the frog tadpoles (Table E1), (2) behavior (i.e., activity, Table E2), morphology (i.e., degree of the defensive morphology, Table E3), or life history (i.e., mean, Table E4, and median, Table E5, metamorphosis timing and size at metamorphosis) of the frog tadpoles, or (3) mortality of the salamander larvae (Table E6), or (4) morphology (i.e., largest body length, Table E7, and largest gape width, Table E8) or life history (i.e., earliest timing of metamorphosis, Table E9, and size at metamorphosis of the first metamorphs, Table E10) of the salamander larvae differed between the two treatments. We used repeated-measures ANOVA to compare activity of the frog tadpoles, conducted a Kaplan-Meier analysis of the earliest timing of metamorphosis in the salamander larvae, and used Kruskal-Wallis tests to compare the other measurements. We did not find any significant differences in any traits between the two treatments (Tables E1–10).
Table E1. Kruskal-Wallis test results for mortality of frog tadpoles.
Measurements |
χ²1 |
P |
|
Mortality of frog tadpoles |
Day 19 to day 31 |
0.49 |
0.49 |
Day 31 to day 150 |
1.10 |
0.30 |
|
Table E2. Repeated-measures ANOVA results for frog tadpole activity.
Factors |
df |
F |
P |
Treatment |
1 |
3.41 |
0.076 |
Time |
2 |
185.69 |
<0.001 |
Treatment*Time |
2 |
1.88 |
0.17 |
Table E3. Kruskal-Wallis test results for the degree of the defensive morphology.
Measurements |
χ²1 |
P |
Degree of defensive morphology on day 31 |
0.0086 |
0.93 |
Table E4. Kruskal-Wallis test results for mean metamorphosis timing and mean size at metamorphosis of the frog tadpoles.
Measurements |
χ²1 |
P |
Mean timing of metamorphosis |
0.078 |
0.78 |
Mean size at metamorphosis |
0.054 |
0.82 |
Table E5. Kruskal-Wallis test results for median metamorphosis timing and median size at metamorphosis of the frog tadpoles.
Measurements |
χ²1 |
P |
Median timing of metamorphosis |
0.75 |
0.39 |
Median size at metamorphosis |
0.36 |
0.55 |
Table E6. Kruskal-Wallis test results for mortality of the salamander larvae.
Measurements |
χ²1 |
P |
|
Mortality of the salamander larvae |
Day 19 to day 31 |
0.056 |
0.81 |
Day 31 to day 151 |
0.051 |
0.82 |
|
Table E7. Kruskal-Wallis test results for largest body length.
Measurements |
χ²1 |
P |
Largest body length on day 31 |
0.0022 |
0.96 |
Table E8. Kruskal-Wallis test results for largest gape width.
Measurements |
χ²1 |
P |
Largest gape width on day 31 |
1.57 |
0.21 |
Table E9. Kaplan-Meier analysis results for earliest timing of metamorphosis in the salamander.
Measurements |
χ²1 |
P |
Earliest timing of metamorphosis |
0.33 |
0.57 |
Table E10. Kruskal-Wallis test results for size at metamorphosis of the first metamorphs.
Measurements |
χ²1 |
P |
Size at metamorphosis of the first metamorphs |
0.36 |
0.55 |