The purpose of this report is to form a basis for the ongoing revision of physical performance tests. Important theoretical prerequisites and considerations and results from investigations of similar nature performed in the US and Canada are reported on very briefly. Directions for developing new tests of muscle strength are suggested. Results from different test of muscle strength correlate poorly. Considerably higher correlations were found when comparing the sum of results from different muscle tests with the sum of performance in actual working tasks. Using a three- level scale, "low", "medium" and "high" for the test, "low"-rated persons would rarely show a high work performance and vice versa. The suggestion regarding a new test of muscle strength consists of mean 3 to 5 different subtests. Passing requires not only a minimal performance in each of the subtests but also that the sum of these should exceed the sum of the minimal levels.