Hello, I am trying to use the biplot function on an object (called 'acp') resulting from a prcomp call. However, the scale I get when calling the biplot function differs according to how I call: biplot(acp) gives me a scale of -0.10 - 0.10 for the variable loadings and a scale of -15 - 15 for the rotated data for the first component. However, if I look at the individual columns of the matrix, I get the following: > max(acp$rotation[,1]) [1] 0.4108393 > min(acp$rotation[,1]) [1] -0.2117583 > max(acp$x[,1]) [1] 5.845884 > min(acp$x[,1]) [1] -4.932297 i.e. a scale of -0.21 - 0.41 for the loadings and -4.93 - 5.85 for the rotated data. These values are also used in the plot when I do: biplot(acp$x[,1:2],acp$rotation[,1:2]) Graphically both look exactly the same, just the depicted scales are different. Am I misunderstanding something, or is there something wrong with the default biplot(acp) call ? Moritz