They've done it again! Those friendly people who brought you DNA fingerprinting have now revealed genetic variation on an extraordinarily fine scale. Not only that: in a recent paper describing DNA sequence differences between alleles at a minisatellite (fingerprint) locus, Alec Jeffreys, Rita Neumann and Victoria Wilson of Leicester University have also answered fundamental questions about the rate and mode of generation of this variation I. DNA fingerprints-',.*, those unevenly spaced ladders of bands, which in many species are unique for an individual, are the result of screening several extremely variable loci simultaneously. Each band represents an allele at a particular locus and is thus one of a pair. Its partner is usually elsewhere in the fingerprint profile (heterozygous state) but may occasionally occur at the same point (homozygous state). The DNA sequences detected by fingerprint probes are very short (e.g. 16 base pairs for the Jeffreys probe X33.15) and often occur in strings, known as tandem repeat arrays or minisatellite DNA. Variation in allele position on a fingerprint profile is due to differences in the length of the DNA fragments detected, which in turn result from variation ,7:~ the number of repeats within an array. At minisatellite loci, repeat units appear to be easily gained or lost and different alleles can have anything from a handful to hundreds of repeats. Although DNA fingerprints are useful for de termining close kin, for many other purposes (e.g. population surveys) it is useful to have locus-specific information. Unfortunately, sorting out which bands