An ability to distinguish individual strains of Helicobacter pylori with sensitivity and efficiency is valuable for studies of the epidemiology, population genetic structure, and evolution of this gastric pathogen. The arbitrarily primed polymerase chain reaction (AP-PCR), or random amplified polymorphic DNA (RAPD) method (1-4), provides one of the most sensitive and efficient means for distinguishing individual strains, and has been particularly useful for H. pylori (5-7). In overview, the method entails PCR amplification with an oligonucleotide primer of arbitrarily chosen sequence and no known match to sequences in the target genome. This allows initiation of DNA synthesis from genomic sites to which the primer is fortuitously, and usually only partially, matched (Fig. 1). The method detects DNA sequence diversity throughout the genome, rather than just at individual loci; less DNA is needed than in most other DNA fingerprinting methods; the DNA need not be very large nor be double-stranded; and no DNA labeling or hybridization, nor information about target DNA sequences, is needed. There are two principal variants of the AP-PCR protocol, one using oligonucleotide primers of about 10 nucleotides (nt) (3,4), and a second using longer primers, which often may have been constructed for other purposes, such as conventional PCR or DNA sequencing (1,2). Fig. 1. Strategy for DNA fingerprinting by the arbitrarily primed (AP) PCR or random amplified polymorphic DNA (RAPD) method. In the top left are diagrammed the genomes of related but genetically distinct strains of H. pylori that may have diverged from a common ancestor by mutation and/or gene transfer from other strains. Pairs of thick half-arrows indicate primer annealing to pairs of sites that result in AP-PCR products; thin half-arrows indicate the same primer annealing to individual sites that are not near enough to other potential primer binding sites in opposite orientation to yield AP-PCR products. The annealing of primers to pairs of incompletely matched sites, which is postulated to be responsible for many AP-PCR bands from prokaryotic genomes, is diagrammed in the DETAIL section (lower left), and the array of products that would be generated from the two strains compared here is diagrammed at the right.