Combination of evidence over samples, each of which is too small to be conclusive, is the central problem in complex inheritance. There are three approaches: meta-analysis, prospective collaboration, and retrospective collaboration. Our experience with the first and last, which are the most cost-effective, is discussed.