Table 2.
Analysis of variance testing the effects of species diversity in combination with temperature on single ecosystem processes
| anova term | d.f. | Leaf decomposition | Herbivory | FPOM production | Algal production | Ammonification | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| F | P | F | P | F | P | F | P | F | P | ||
| Temperature | 2 | – | – | – | – | – | – | – | – | – | – |
| Richness | 3 | 0.4 | 0.728 | 3.8 | 0.017 | 6.0 | 0.002 | 0.3 | 0.815 | 0.2 | 0.874 |
| Type | 3 | 37.8 | <0.001 | 19.3 | <0.001 | 150.7 | <0.001 | 30.0 | <0.001 | 7.9 | < 0.001 |
| Composition | 8 | 0.2 | 0.985 | 1.5 | 0.198 | 2.7 | 0.019 | 0.6 | 0.742 | 2.7 | 0.017 |
| Richness × Temperature | 6 | 1.5 | 0.190 | 1.0 | 0.428 | 1.4 | 0.221 | 4.3 | 0.002 | 0.3 | 0.937 |
| Type × Temperature | 6 | 3.8 | 0.004 | 3.4 | 0.008 | 6.0 | < 0.001 | 3.4 | 0.007 | 4.7 | 0.001 |
| Composition × Temperature | 16 | 2.3 | 0.017 | 1.4 | 0.191 | 1.4 | 0.169 | 1.7 | 0.092 | 2.6 | 0.007 |
| Blocks (Rooms) | 3 | ||||||||||
| Error | 42 | ||||||||||
Each row in the anova table corresponds to one of the hypotheses in Table1. In turn, this corresponds not to a model but to the difference between the model shown in the same row of Table1 and the sum of all simpler models (see Figure S3). For example, the small P-values observed for the row labelled ‘Type × Temperature’ indicate that, for these data, the larger model ‘Type × Temperature’ cannot be simplified to the smaller model ‘Type + Temperature’. There is no valid statistical test for the main effect of ‘Temperature’ on processes because ‘Temperature’ and ‘Rooms’ were the same in our study (one environmental-control room per temperature; see Materials and methods).