Dear R colleagues, I applied a pca to 130 farms (sites) and 25 variables (species) using the R package vegan. Besides analyzing the variation explained by the different pca axes and the variables the pca axis are correlated with, I am also interested to analyze the correlation between the variables (e.g. groups within the pca biplot). I used rda(variables, scale=TRUE), so the angles between the variables and/or the pca axis represent their correlation (pearson correlation coefficient, loadings). pca output offers the scores to analyze the correlation between variables and pca axis. For the correlation between the variables I used cor (variables, method="pearson"). I expected that these correlation coefficient would be similar to the angles between the variables in the pca biplot. Indeed for most of the cases it is true, but why are there some evident differences? Is there a possibility to directly extract the correlation coefficients between the variables from the pca biplot? Thanks a lot Sibylle