Acoustic emission (AE) signals that are generated by different microscopic processes during flaw growth in graphite-epoxy specimens have measurably different characteristics. In particular, the amplitudes of the emissions and a parameter that describes their frequency spectral content seem to give the most information about the processes. These parameters have a range of values for a given process which can be described by certain types of analytical distribution functions. When several processes occur simultaneously during flaw growth, such as epoxy crazing, fiber fracture, fiber-matrix disband and interlaminar cleavage, the distributions in the values of the AE parameters generally overlap so that identification of an individual AE signal as being caused by a particular process is not possible. However, statistical evaluation of the data for a few hundred events in terr.ls of the analytical distributions, once the shape and modal value of these distributions are defined for each process, should provide a quantitative measure of the relative amounts of the various processes which occurred. Analyses of many data sets are required to develop confidence in the decomposed distributions as being descriptive of the individual processes. The ultimate purpose for this determination is to provide a description of the stage of flaw growth from the quantitative knowledge of the types and the amounts of the microscopic processes which occurred.