I'm using SPSS 19 and am having problems with my KW test. I'm comparing differences of metal concentrations in fish tissue between 6 sites for 10 metals. The Kruskal Wallis output shows one test result for the first metal, which is what I expected and when I look at the pairwise comparisons, they show the post hoc differences between sites, again what I expected. When I look at the 3rd metal, there are three test results, before I even look at the pairwise comparisons. By the time I get to the 10th metal, there are 21 test results. I'm not sure why SPSS is giving me such a difference in number of tests between metals. Are they iterations? How do I know which test result to report?