In vivo animal models of primary brain tumors are necessary to advance knowledge related to the complex interactions between glioma cells and the adjacent brain. A cardinal feature of glioma growth, and a major reason why neurosurgical and adjunctive therapies ultimately fail in most patients is their invasive properties. We have adapted a previously described animal model developed by one of us to give better histological detail while preserving the identification of single infiltrating glioma cells. GL261 glioma cells were first transfected with the plasmid encoding green fluorescent protein (GFP) and then implanted into the brains of syngeneic C57BL/6 mice. Identification of GFP-positive tumor cells in paraffin sections of the brains of tumor-bearing animals utilized an antibody for conventional immunoperoxidase immunohistochemistry. This method is a more powerful technique compared with the prior use of frozen sections and fluorescence microscopy to identify GFP-tagged tumor cells. We find that this new method provides improved morphology and proves to be a sensitive and reliable system for detection of invading glioma cells. Using this methodology with other advanced technologies (eg, laser capture microdissection) holds out the promise of helping to elucidate the molecular mechanisms of glioma cell infiltration and invasion into the surrounding brain.