The aim of this thesis was to investigate the suitability and application of an in vitro gas production technique as a routine analytical tool in animal nutrition studies. Firstly, the effect of several biological, chemical and physical factors on the measurement of gas production were investigated. These factors included incubation temperature, head-space pressure, shaking movement of the in vitro cultures, the nature of the feedstuff (chemical composition, particle size and drying process), the source of the microbial inoculum and the apparatus used to measure gas production. Secondly, two potential applications for the technique were investigated; (1) as a routine feed analysis tool for the prediction of the digestible energy (DE) content of equine feeds and (2) as a screening method for investigating the potential use of novel feed additives. The first application was investigated by incubating sixteen feedstuffs of known DE. Gas production parameters, DM loss and VFA production were then used to derive prediction equations. The best prediction equation was DE = -0.68 + 0.01087 DML + 6.82 Z - 2.297 log L_T (R² = 0.878; RSD = 0.99; where DML is dry matter loss in vitro, Z is a rate parameter and L_T is the lag time for gas production). The second application was investigated using four antibiotics; monensin, avoparcin, penicillin G and chloramphenicol in the automated gas production system. Differences in gas production profiles were detected between antibiotic supplemented and control fermentations suggesting that the technique can be used to screen feed additives. Results presented in this thesis inform the development of standardised methodologies and procedures for use during in vitro gas production studies, thus enabling this technique to be adopted as a robust, reliable and routine analytical tool in animal nutrition studies.