Abstract
This paper introduces an exact test for testing H 0: F is exponential against H 1: F is harmonic new better than used in expectation (HNBUE) or F is harmonic worse than used in expectation (HNWUE) on complete data. Selected critical values are tabulated for sample size n = 5(1)60. An asymptotic test is made available for large values of n. The standard normal table can be used for the above testing problem. Power of the test is estimeted by simulation. The problem when right censored data is available is also handled. An example of 81 patients of melanoma disease demonstrates the practical application of the proposed test in medical Sciences.