Will the real iris data please stand up?

This correspondence points out several published errors in replicates of the well-known iris data, which was collected by Anderson (1935) but first published by Fisher (1936).