This study intends to determine the rational criteria (e.g., threshold value) for applying the weight variation test and to investigate the adequacy of the acceptance value for existing commercial products in Japan. The studied products were 489 lots (3 lots x 163 products) of compressed tablets (plain, film-coated, sugar-coated) and 42 lots (3 lots x 14 products) of hard capsules marketed in Japan. The individual drug content and the weight of 10 units in a lot were determined for each product and the acceptance values were calculated according to the Japanese Pharmacopoeia thirteenth edition (JP13) Content Uniformity Test (M=100.0, k=2.2). Product-specific intra-lot relative standard deviation of content (RSDD), weight (RSDW) and concentration (RSDC) were calculated by analysis of variance (ANOVA) using three lots of data per product. The RSDD and RSDC tended to increase with the decrease of the label strength for plain tablets, but not for film-coated and sugar-coated tablets, and hard capsules. A good correlation was found between RSDD and RSDC but not between RSDD and RSDW. These findings indicate that 1) it is difficult to rationally set the threshold level for weight variation, especially regarding the dosage forms except for plain tablets, 2) the application of weight variation tests should, in principle, be decided on the mixing homogeneity that is RSDC. 3) Most (99.6%) of the tablets and all the capsules investigated met the requirement of content uniformity test of JP13. Therefore the criteria of the JP13 content uniformity test are considered acceptable from the viewpoint of manufacturing capability.