We report the results of the 1995 and 1996 Paternity Testing Workshops of the English Speaking Working Group of the International Society for Forensic Haemogenetics. In 1995, 18 laboratories participated and in 1996, 21 laboratories participated. Each year, blood samples from three persons (child, mother and alleged father) were sent to participating laboratories which performed paternity testing according to their usual protocols. The results and answers to questionnaires concerning methods were compiled and are presented in this report. From the questionnaires, a general tendency to a more frequent use of polymerase chain reaction (PCR) based methods was seen. In 1996, 62% of the laboratories used PCR based methods. Ten per cent of the laboratories used only PCR based methods. The remaining 90% of the laboratories performed restriction fragment length polymorphism (RFLP) investigations of variable numbers of tandem repeat (VNTR) loci with single locus probes (SLPs) either alone or in combination with PCR based typing, multi locus probing, classical systems (ABO etc.), or serological HLA typing. In 1996, typing with classical systems was used in 29% of the laboratories. The majority of the laboratories performed RFLP typing of VNTR loci using very similar methods. The results and the inter-laboratory variations of the measured lengths of the DNA-fragments of the VNTR regions D2S44, D7S21, D7S22, and D12S11 of the trios were analysed. The overall coefficient of variation was 2.15% in 1995 and 1.43% in 1996. During the period 1991-1996, the inter-laboratory variation has decreased, most probably due to the fact that the methods have now been optimised and the majority of the participating laboratories have adopted the standardised method for RFLP typing with SLPs which was agreed upon for investigations in crime cases by the European DNA Profiling Group. In 1996, eight laboratories reported the results of PCR based typing of the short tandem repeat (STR) locus HumTH01, six laboratories reported results of HumVWA31A typing, and five laboratories reported the results of typing of the STR locus HumF13A1 and the VNTR locus D1S80. The results were concordant although the nomenclature was slightly inconsistent concerning the classification of an irregular repeat of the HumTH01 system.