Experience accumulated in several years of routine work on liquid scintillation counting is presented, together with an analysis of the main sources of error and the mention of some means to reduce them. A precise method based on comparing analytical samples with known standards is described.