The increasing volume of electronic mail communication threatens to cause a state of “e-mail overload” where the volume of messages exceeds individuals' capacity to process them. To address this problem, this study extends the application of hierarchical clustering to the domain of e-mail. We report on the design and development of a system that applies a multi-weight, multi-attribute clustering approach to a collection of messages. We found strong evidence that clustering messages improves users' ability to locate messages compared to an ordered list, and promising (though weaker) evidence of even greater improvement when given the ability to adjust attribute weights.