We describe software for aligning protein or nucleic acid sequences based on the concept of match density. This method is especially useful for locating regions of short similarity between two longer sequences which may be largely dissimilar (e.g. locating active site regions in distantly related proteins). Our software is able to identify biologically interesting similarities between two sub-regions because it allows the user to control the matching parameters and the manner in which local alignments are selected for display. Furthermore, the collection and ranking of alignments for display uses a novel, highly efficient algorithm. We illustrate these features with several examples. In addition, we show that this tool can be used to find a new conserved sequence in several viral DNA polymerases, which, we suggest, occurs at a functionally important enzymatic site.