A large number of bacterial species have been identified in fetal membranes after preterm labour (PTL) associated with intrauterine infection by microbiological culture. In this study, we have investigated a molecular and bioinformatic approach to organism identification which surmounts the need for specific and diverse microbiological culture conditions required by conventional methods. Samples of fetal membranes were taken from 37 preterm infants, and 6 normal term controls delivered by caesarean section, in which bacteria had been detected by in situ hybridization of 16S ribosomal RNA using a generic probe. Degenerate primers were designed to amplify bacterial 16S ribosomal DNA by PCR and used to amplify bacterial DNA from human fetal membranes. Amplicons were cloned, sequenced and bacteria were identified bioinformatically by comparison of sequences with known bacterial DNA genomes. In situ hybridization using an organism specific probe was then used to confirm the presence of the commonest identified organism in tissue samples. Bacterial DNA amplified from 15/43 samples, all from preterm deliveries, and the bioinformatic approach identified organisms in all cases. Multiple bacteria were identified including Mycoplasma hominis, Pasturella multocida, Pseudomonas PH1, Escherichia coli and Prevotella bivia. The commonest organism Fusobacterium nucleatum was found in 9/15 (60%) of samples. Ten of the 12 samples obtained after prolonged membrane rupture were positive for bacterial DNA, and 7 of these (70%) contained DNA from F. nucleatum. Bacteria from fetal membranes may be identified by molecular and bioinformatic methods. Further work is warranted to investigate the apparent linkage between F. nucleatum, fetal membrane rupture and preterm delivery.