The proliferative capacity of brain-tumor cells was analyzed in vitro and in situ using monoclonal antibody (MAb) against deoxyribonucleic acid (DNA) polymerase alpha. For the in vitro studies, two cultured human glioma cell lines were investigated using MAb against DNA polymerase alpha, the MAb Ki-67, a serum against proliferating cell nuclear antigen (PCNA/cyclin), bromodeoxyuridine (BUdR), and an anti-BUdR MAb. During exponential growth of the cells, the percentage of polymerase alpha-positive cells (the "polymerase alpha score") ranged from 72.0% to 77.1%, the Ki-67-positive cells (the "Ki-67 score") ranged from 43.4% to 59.4%, the PCNA/cyclin-positive cells from 30.9% to 41.4%, and the BUdR labeling index from 28.6% to 39.3%. For the in situ studies, tissue from 60 human brain tumors and from two normal human brains was investigated and the polymerase alpha scores and Ki-67 scores were compared. In normal brain tissue, no immunostaining was found by either method. In brain tumors, both the polymerase alpha scores and the Ki-67 scores correlated with the histological grade of malignancy. Polymerase alpha scores were generally higher than Ki-67 scores in the same specimen, especially in malignant brain tumors. These findings suggest that immunostaining of DNA polymerase alpha is a convenient and important new method by which to estimate the cellular proliferation rate of brain tumors. Polymerase alpha scores may be closer to the growth fraction of the individual tumor than the MAb Ki-67 or other scores.