Chewing performance can be defined in terms of the reduction in food particle size after 15 chewing strokes. In this study, the relationship between chewing performance and electromyographic activity was investigated to develop optimal values of electromyographic v ariables, based on their ability to predict chewing performance. Electrognathographic and electromyographic recordings from surface electrodes over the digastric (abductor), masseter, and temporalis (adductors) muscles were made from 24 subjects while they chewed a hard fruit gum. A moderate negative correlation was found between the food particle size and the root mean square calculation for masseter activity (-.48; P .01). Weaker postive correlations were found between particle size and the asynchrony of ipsilateral and contralateral anterior temporalis muscles. (.36; P .05). A multiple regression model of electromyographic and electrognathographic varioables was able to predict chewing performance with an R(to the 2nd power) value of .66. If chewing performance is used as an output measure of masticatory function, it may be possible to determine optimal ranges for electromyographic variables and jaw movements.