A number of short tandem repeat (STR) loci are currently being examined for their usefulness as markers of identity; HUMVWA31A is one such locus. We used a high-sieving agarose technique to type 200 British Caucasians for this locus. Comparison of the resultant allele frequencies with other published databases showed that their distributions were similar. Observed heterozygosity was similar to that reported in other population studies but significantly lower than expected. A goodness of fit analysis of observed and expected genotypes was highly significant, suggesting a deviation from Hardy-Weinberg equilibrium, although reports from other populations appeared not to show such differences. The possibility that mistyping had led to an erroneous statistic was investigated by detailed examination of samples within our own laboratory and in two other laboratories, involving verification of the agarose typing by both automated fluorescent detection techniques, and sequencing. Other reported deviations from Hardy-Weinberg equilibrium in STR systems are discussed.