Fig 1.
Results of the Principal Component Analyses applied to functional traits of tree species from Mexican tropical dry and wet forests.
(a) PCA of dry forest species (n = 51), (b) PCA of wet forest species (n = 81). Species (grey symbols) were separated based on their functional traits shown as arrows; LA = leaf area, SLA = specific leaf area, LD = leaf density, LT = leaf thickness, LDMC = leaf dry matter content, PL = petiole length, WD = wood density, LC = leaf compoundness (0 = simple, 1 = compound), Di = dispersal syndrome (0 = abiotic, 1 = biotic), De = deciduousness (0 = evergreen, 1 = deciduous). LA and PL were ln-transformed.
Fig 2.
Correlation coefficients (CC) of all pairwise trait combinations (11 traits, resulting in 55 pairwise trait combinations per forest type, see Table 2) of dry forest species plotted against those of wet forest species.
Correlation coefficients represent Spearman coefficients except when relating binary variables, then the Phi coefficient was used. The pairwise correlation coefficients of dry forest proved to be significantly correlated with those of the wet forest (Pearson product moment correlation [R], P < 0.001), indicating that trait spectra are consistent across the two different forest types.
Table 1.
Eigenvector scores of functional traits on the two main principal components for dry forest and for wet forest.
Table 2.
Spearman coefficients of the pairwise relations between variables and the principal components (Fig 1).
Fig 3.
Changes in the dominant plant strategies with succession.
Stand basal area was used to indicate succession; it increased asymptotically with successional age and reflects successional change in vegetation structure. Functional composition was calculated using the community-weighted mean of species scores on the principal component axes. (a) Dry forest succession (open symbols, broken regression line) was characterized by changes along the first PCA axis (Fig 1a) and reflected changes from deciduous species to evergreen species that invest in a secure reproductive strategy. (b) Wet forest succession (filled symbols, continuous regression line) was characterized by changes along the second PCA axis (Fig 1b) and reflected changes from an acquisitive strategy to a conservative strategy. Given is the r2, * P < 0.05; ** P < 0.01. See S1 Fig for the trends with fallow age as an indicator of succession.