Restrictions on the hyperbolic trigonometric (HT) transformation are imposed to guarantee that a probability density function is obtained from the maximum likelihood estimation. Performance of the restricted HT transformation using data generated from normal, beta, gamma, logistic, log-normal, Pareto, Weibull, order statistic, and bimodal populations is investigated via sampling experiments. Results suggest that the restricted HT transformaltion is sufficiently flexible to compete with the actual population distributions in most cases. Application of the restricted HT transformation is illustrated by characterizing uncertain net income per acre for community-supported agriculture farms in the northeastern United States.