A well-known theorem states that the maximal correlation between a pair of bivariate normal random variables equals the absolute value of their ordinary (Pearson) correlation. This work provides a short new proof, besides establishing some results regarding the maximal correlation coefficient in general.