Table 3. Significance of fixed effects and contrast tests in analyses of variances of the effects of ecosystem (forest vs savanna), species within ecosystem (Ceiba and Cynometra within forest, Bridelia and Piliostigma within savanna), shading (shade vs light), and fire (burnt vs protected) on all raw plant size variables (Fig. 1) and allometric ratios.
| fixed effects (F tests) | size variables | ratios | RGR | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| W | Wabg | WR | Wstr | WL | LA | LP | nL | Hmax | nA | Dmax | VP | HL | LD | LMAP | rs | ls | ||
| (Intercept) | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** |
| treatment | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** |
| age | — | — | — | — | — | — | ** | *** | *** | — | — | *** | — | — | — | — | — | |
| spatial effects | *** | *** | ** | ** | *** | *** | * | * | ** | ** | ** | . | ** | |||||
| contrasts for treatment (t tests) | ||||||||||||||||||
| forest – savanna | ** | *** | . | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | |||
| Ceiba: 3 – 9 | *** | *** | *** | *** | *** | — | — | — | — | — | — | — | — | — | — | *** | ** | *** |
| Ceiba: 9 – 15 | *** | *** | *** | *** | * | . | — | — | — | — | *** | *** | — | *** | ** | *** | *** | |
| Cynometra: 3 – 9 | *** | *** | *** | *** | *** | — | — | — | — | — | — | — | — | — | — | * | *** | |
| Cynometra: 9 – 15 | ** | * | *** | * | . | — | — | — | — | — | — | — | — | . | ||||
| Bridelia: 3 – 9 | *** | *** | *** | *** | *** | — | — | — | — | — | — | — | — | — | — | * | *** | |
| Bridelia: 9 – 15 | *** | *** | *** | *** | *** | * | — | — | — | — | *** | *** | — | * | . | *** | ||
| Piliostigma: 3 – 9 | *** | *** | *** | *** | *** | — | — | — | — | — | — | — | — | — | — | *** | *** | |
| Piliostigma: 9 – 15 | *** | *** | *** | *** | *** | *** | — | — | — | — | *** | *** | — | ** | *** | *** | ||
| forest:shading | ** | ** | ** | ** | *** | *** | ||||||||||||
| savanna:shading | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | *** | ** | *** | *** | *** | ||
| forest:fire | *** | ** | *** | *** | — | — | * | — | * | * | ||||||||
| savanna:fire | . | . | *** | *** | ** | *** | *** | * | *** | * | ||||||||
| light: Ceiba – Cynometra | *** | *** | * | *** | ** | * | *** | *** | . | *** | *** | *** | *** | |||||
| light: Bridelia – Piliostigma | ** | ** | * | * | ** | *** | *** | *** | *** | *** | * | |||||||
| shade: Ceiba – Cynometra | *** | *** | * | *** | ** | *** | * | * | *** | *** | *** | *** | *** | ** | *** | *** | *** | *** |
| shade: Bridelia – Piliostigma | *** | *** | *** | *** | *** | . | *** | |||||||||||
| burnt: Ceiba – Cynometra | ** | ** | ** | ** | *** | *** | *** | * | *** | *** | — | — | * | — | * | * | ||
| burnt: Bridelia – Piliostigma | *** | *** | . | *** | — | — | *** | — | ||||||||||
| protected: Ceiba – Cynometra | — | — | — | — | — | — | . | . | . | *** | — | — | — | — | — | — | — | |
| protected: Bridelia – Piliostigma | — | — | — | — | — | — | *** | *** | *** | *** | — | — | *** | — | — | — | — | — |
Non-destructive variables (LP, nL, Hmax, nA, Dmax, VP, HL, LD) were analysed through linear mixed models, all other variables through simple linear models. All variables were log-transformedbefore analysis. Date was treated as a covariate for some variables, and as a factor for others, depending on thenumber of measurement dates for each variable (see Appendix 4 for details of statistical models and contrast tests). Variables named as in Fig. 1. —: term not present in specific model; ***: P < 0.001; **: 0.001 ≤ P < 0.01; *: 0.01 ≤ P < 0.05;. : 0.05 ≤ P < 0.1.