When frequencies of the occurrence of some phenomenon are available on a monthly or other cyclic basis a simple test for possible periodic effects is obtained by comparing totals over 2 halves of the cycle. For example, totals for 6 consecutive months are contrasted with totals for the remaining (not necessarily consecutive) months. Often the line of demarcation between the 2 halves is arbitrary and the test will then be based on the largest difference which can be constructed. Allowance can be made for this selection effect by means of the upper 5% and 1% points of the largest contrast, which are presented in this paper. A numerical example is given and other methods of analysis are mentioned.