Figure 4
From: Mycorrhizal fungal community structure in tropical humid soils under fallow and cropping conditions
Venn diagrams of the RDA-based variation-partitioning results indicating the proportions (in percentage, %) from (a) the quantitative and factor estimators, fallow types (F, three levels; forest fallow, short fallow and cropping conditions), soil origins (S, three levels; Typic Kandiudox, Typic Kandiudult and Rhodic Kandiudult) and soil charactrisitics (SC, eight parmeters; soil-pH, pH measured in soil suspension; soil-N, nitrogen content; soil-C, carbon content; soil-Mg, magnesium content; soil-Al, aluminium content; soil-K, potassium content; soil-Ca, calcium content; and soil-Pi, inorganic phosphate content measured in soil), and their combinations; and from (b) the quantitative and factor estimators S, F and sampling depths (D, two levels; 0–10 cm and 10–20 cm), and their combinations on the arbuscular mycorrhizal (AM) community composition. (c) Multigroup structural equation modeling displaying the pathways by which fallow practices showed influences on the soil properties and the AM community composition in two soil depths (0–10 cm/10–20 cm). The various pathways drawn from the models used the dataset as single group. Path coefficient values shown above to the pathways are for the soil depths and listed in their order (0–10 cm/10–20 cm). Only the significant path coefficient values (p < 0.05) from the models are presented. The thick and thin arrows indicate the direct and indirect effects, respectively. Dotted frame indicates the soil properties, namely soil-C, -N and -Ca contents that strongly influenced the model groups. Values in parentheses indicate path coefficient values for the main models using the dataset as single group. ns, not significant.
