Booster phenomenon (recall effect) of tuberculin skin test, which disturbs diagnosis of tuberculous infection, is prevalent among BCG vaccinated population. We retested 34 nurse students whose initial tuberculin reaction was smaller than 30 mm by erythema (Group A) and 53 hospital employees whose initial reaction was smaller than 20 mm by erythema (Group B). Among the people whose diameter of erythema was less than 10 mm by the first test, 88 percent (8/9) of group A and 43% (6/14) of group B showed reaction 10 mm < or = by erythema and among those whose induration was < 10 mm, 54% (6/11) of group A and 48% (12/25) of group B showed reaction 10 mm < or = in the second testing. Mean and standard deviation of [the difference between the diameter of the 2nd and the 1st testing] was +7.3 +/- 11.8 mm in group A, +9.8 +/- 11.1 mm in group B by erythema and +2.6 +/- 5.9 mm in group A, +2.9 +/- 5.1 mm in group B by induration. These results indicate that booster phenomenon is highly prevalent among the tested group and there can be no appropriate criteria to distinguish new infection and booster phenomenon. Though two-step tuberculin skin test is recommended to get rid of booster phenomenon. Only a little is known about the value of this test to diagnose new infection in Japanese population, majority of whom being BCG vaccinated. Further investigations are required to apply two-step tuberculin skin test for diagnosis of new infection among hospital employees and health care workers in Japan.