The gas production (GP) technique has been developed at a number of laboratories throughout the world as a means of providing information on the rate and extent of the rumen fermentation of foods. However, each laboratory has developed its own method for obtaining a GP profile. The incubation may occur in gas-tight syringes (e.g. Hidayat et al., 1993; Khazaal and Ørskov, 1994) or in culture bottles using pressure transducers that are either manually operated (Theodorou et al., 1994) or automated (e.g. Pell and Schofield, 1993; Cone et al., 1996). Clearly, the method adopted to measure gas production could have quite a significant effect on the profile obtained. In a previous study (ring test 1; J. A. Huntington, unpublished) identical samples of five different foods (barley, soya-bean meal, ‘poor’ hay, ‘good’ hay and cellulose) were supplied to eight different laboratories. Each laboratory obtained GP profiles of each food using their own particular technique. These were then fitted to the model proposed by France et al. (1993). As might be expected there was quite substantial variation between laboratories. It was therefore decided to repeat the study (ring test 2) but to adopt a more standard procedure to determine the extent of variability between laboratories that is inherent in this technique.