A table is presented here giving the part-whole correlations to be expected for selected test reliabilities and relative sizes of the part. Two uses for such a table are also given: 1) to decide how many items to use in a common subtest in order to equate two forms of a test; and 2) to estimate the reliability of a test after the test has been given and the part-whole correlation is known. The theoretical derivation of the part-whole correlation is also discussed briefly.