Table 1.
Results of linear regression models for biodiversity of plants and arthropods in twelve traditional cornfields located in the Colombian Andes.
Coefficient values (b) of predictor variables and P-values are shown for each model. Alternative models were constructed using land use types measured at 100, 250, 500, 750 and 1000 m. Only significant models at 0.05 confidence level are included here.
Fig 1.
Relationships between the diversity of herbivores and plant diversity of order q = 2 in cornfields.
a) For all herbivore species 0D, b) for common herbivore species 1D, and c) for abundant herbivore species 2D. The plots indicate a positive relationship between the effective number of common plant species (1D) and all orders of herbivore diversity. Lines show fitted values from significant linear regression models.
Table 2.
Results of linear regression models for biodiversity of arthropods and plant diversity in twelve traditional cornfields located in the Colombian Andes.
Coefficient values of predictor variables (b) and P-values are shown for all models. Only significant models at a 0.05 confidence level are included here.
Table 3.
Effective number of plant and predator species (Mean and SD) according to previous land use of the cornfields.
Fig 2.
Beta diversity profiles of plants, arthropods and predators collected in twelve traditional cornfields.
The plot indicates that differences in species composition among cornfields were stronger among abundant plant species, whereas for arthropods these differences arose among rare species, particularly for predators.
Table 4.
True diversities of plants, herbivores and predators collected in twelve traditional cornfields in an area of 19 km2 in the Colombian Andes.
Table 5.
Pearsons’s r correlation for the Mantel tests between pair-wise beta diversity matrices and distance matrices for geographical location and environmental gradients in twelve traditional cornfields.
Environmental gradients included differences between fields in altitude, forest cover, percentage of soil organic matter and plant species richness. Bold values indicate significant correlations at the 95% confidence level.
Fig 3.
Non-metric multidimensional scaling ordination diagrams based on the Jaccard (left column) and Morisita-Horn (right column) indices for plants (a-b), herbivores (c-d), and predators (e-f) collected in twelve traditional cornfields.
Ellipses indicate the clustering of biological communities based on the land use context of the cornfield: Fields A1 to A6: 1–26% of native forest in a radius of 250 m around the crop; fields F1 to F6: 27–50% of native forest around the crop. The plots indicate that dissimilarity in plant species composition was related to the land use context around the crops, however for arthropods, the plots indicate a clear overlap in species composition between land use contexts. Stress values: a = 0.01, b = 0, c = 0.16, d = 0, e = 0.18 and f = 0.