THE STUDY EMPIRICALLY DETERMINED THE EFFECTS OF NONNORMALITY UPON SOME SAMPLING DISTRIBUTIONS OF THE PRODUCT MOMENT CORRELATION COEFFICIENT (PMCC). SAMPLING DISTRIBUTIONS OF THE PMCC WERE OBTAINED BY DRAWING NUMEROUS SAMPLES FROM CONTROL AND EXPERIMENTAL POPULATIONS HAVING VARIOUS DEGREES OF NONNORMALITY AND BY CALCULATING CORRELATION COEFFICIENTS FOR THESE SAMPLES. THE SAMPLING DISTRIBUTIONS OBTAINED FROM THE EXPERIMENTAL POPULATIONS WERE COMPARED WITH THE THEORETICAL SAMPLING DISTRIBUTIONS OF PMCC OF RANDOM SAMPLES FROM BIVARIATE NORMAL POPULATIONS AND WITH THE SAMPLING DISTRIBUTIONS OBTAINED FROM THE CONTROL POPULATIONS WHICH WERE APPROXIMATELY BIVARIATE NORMAL. THE RESULTS OF THE INVESTIGATION OF THE EFFECTS OF NONNORMALITY UPON THE SAMPLING DISTRIBUTIONS OF THE PMCC INDICATED THAT NONNORMALITY HAD AN AFFECT AND THAT IT WAS OF SUCH A MAGNITUDE THAT IT SHOULD BE TAKEN INTO ACCOUNT IN DEALING WITH STATISTICAL TESTS OF INFERENCE INVOLVING PRODUCT MOMENT CORRELATION. MORE RESEARCH IS NEEDED, ESPECIALLY IN REGARD TO OTHER TYPES OF NONNORMALITY AND POPULATIONS NOT HAVING IDENTICAL MARGINAL FREQUENCIES. (GD)