In conclusion it should be noted that the practical use of these classification techniques will usually require the use of high speed computing facilities. This is especially true if the dimension of the problem is at all large or if one must empirically generate the conditional istribution of the statistic being used by utilizing the individual observations available in the samples. There are many unresolved problems associated with the use of many of these techniques, but it is felt that the systematic exploration of their applicability in many practical problems cannot help but advance the general state of the art. Although the discriminating power of the set of variables currently being accumulated can be determined, the characteristics of the underlying distributions and the relative effectiveness of the competitive procedures must in many respects be tackled pragmatically. Attention must be given to the problem of estimating both the underlying a priori probabilities associated with the populations being considered along with the misclassification cost factors. Individuals may feel that such refinements are inappropriate to their particular classification problem, but it can be argued that until one addresses himself to the problem in some such systematic and scientific way, no real improvement can be expected. The criterion of worth of any system is its operational effectiveness and thus one should not only feel challenged to obtain estimates of the operational effectiveness of the “system” he is now using, but he should also investigate how the effectiveness may be improved by using one of the above statistical classification techniques.