Associative effects of feed mixtures have been studied looking at the differences (i.e. gas production) between single and mixed substrates, and significant differences have been found when fodder trees were mixed (Rosales et al., 1998). This may not be the more appropriate form to analyze mixtures, because small departures from simple additivity may result in significant differences as the sum of all its components equals one. Taking this fact into account mixture designs have been suggested as an adequate tool when dealing with mixtures (Mead, 1988). Mixtures design are a special class of response surface designs in which the sample under study is made up of several components or ingredients, thus the response depends on the relative proportion of the components. The objective of the present study was to asses the power of mixture designs to identify associative effects in feed mixtures.