In a recent investigation an attempt was made to discover, by this empirical method, a set of mental tests which would aid in the selection of efficient workers in a specific field. Thirty workers who were already employed under fairly comparable conditions of work were taken as subjects in a preliminary search for tests of value. These thirty people were each put through a series of "association tests," of the familiar laboratory form, naming opposites, naming colors and forms, completing mutilated passages, following hard directions, giving responses bearing specified relations to stimulus words, cancellation and number checking, etc. While these tests were in progress, during a period of several days, the thirty workers were rated by three supervisors who were familiar with their work at the actual task, and who had for some time been observing their performance with a view to making subsequent judgments. Each supervisor arranged the thirty workers in an order of merit, according to his or her impression of their relative efficiency. The judgments of these three supervisors were then averaged and each worker assigned a final position on the basis of these averages. This was believed to be as accurate a measure of actual ability as could be secured under the complex conditions of work. The results of these ratings were then compared with the results of the mental tests. Some of the tests were found not to correlate with the ratings for actual working efficiency. Three tests showed definite and positive correlations, as follows: _Color-Naming_ (thirty-seven per cent), _Hard Directions_ (forty per cent), _Completion_ (thirty-three per cent). When results from these three tests were combined, the records correlated with the ratings by a coefficient of fifty-five per cent. These three tests were then accepted as having value in the selection of good operators, and search was continued for further tests which might also yield positive correlations. This investigation is again an illustration of the purely empirical method.