Hello, I've got a question concerning kruskal.test: Is there any post-hoc test for kruskal.test like there is one for aov (TukeyHSD) and is it possible to plot this post-hoc test (if available). I need to find out, which of my sample-groups differ significantly from each other... If not, is there any other approach to achieve the desired information? Thanks in advance and have a nice day Allan -- Allan Buras Dipl.-Geograph & PhD-student Landscape ecology and geobotany University of Greifswald, Germany Tel.: 03834 864182