The authors evaluated the predictability of MIB-1 immunohistochemistry for growth and recurrences of meningiomas using two different counting methods: 1) in the area of the highest MIB-1 labeling (HL method) and (2) in randomly selected fields (RS method). The MIB-1 staining indices (SIs) determined by the HL method were approximately twice as high as those by the RS method, and the correlation coefficient between them was high (R = 0.86) in 139 meningiomas when transformed logarithmically. The differences in SIs in histologic grades were significant with either method. Tumor doubling time (Td) was calculated in 22 meningiomas from serial radiologic examinations. The RS method yielded a slightly higher correlation coefficient between log Td and log SI than the HL method. When the authors examined the predictability of recurrence in 112 totally removed meningiomas, the RS method distinguished the recurrent group more definitively. Several benign meningiomas with low SIs by the RS method exhibited focal accumulation of MIB-1-positive cells. Although they were assigned high MIB-1 values by the HL method, these meningiomas did not recur, and therefore obscured the prognostic importance of the MIB-1 value with the HL method. Focal accumulation of MIB-1-positive cells in meningiomas is not likely to correlate with their biologic aggressiveness.