The modified histologic grading system for solid ovarian teratomas used by Norris et al. which attempts to quantitate the presence and amount of primitive neuroepithelium, is in our opinion, an improvement over the original system. In our experience, however, some difficulties in its application remain. Grade 0 and 1 tumors were easily identified, but there was considerable confusion in attempting to separate grade 2 tumors in which neuroepithelium does not exceed three low-power microscopic fields in any one slide from grade 3 tumors in which neuroepithelium occupies four or more low-power fields. The difficulty arose in deciding whether the amount of neuroepithelium required to qualify as "occupying" a low-power field was sufficient if a single focus was present, or if the entire microscopic field should contain neuroepithelium. In the former case, four small foci of primitive neural tissue present in separate locations in a single slide would qualify for diagnosis as a grade 3 tumor, whereas if the same amount of tissue was found close together (i.e., within the space of a single low-power field), it might be read as a grade 1 tumor. This system is also potentially dependent on a fortuitous sampling technique, since a grade 3 tumor could be diagnosed as a lesser grade if a particular section in which four low-power fields containing neuroepithelium is not chosen for microscopic examination. One of the most important steps in attempting to resolve some of the conflicts surrounding the correlation of tumor histology with natural history in solid ovarian teratomas is to assure adequate sampling of the tumor. Norris et al. suggest that one block per centimeter of the maximum diameter of the tumor should be evaluated, others stress the importance of directing sampling toward areas of hemorrhage or necrosis, in an attempt to disclose immature foci. Because of these potential problems, we suggest as a possible alternative using the estimated percentage of neuroepithelial tissue present in all the tissue examined microscopically as a basis for determining the histologic grade. In our small series, it appears that tumors without neuroepithelium would be equivalent to a grade 0, tumors with up to 10 percent neuroepithelium would be graded 1, those with approximately one-third neuroepithelium would be grade 2 tumors and those with approximately one-half or greater of their tissue as a neuroepithelium would be grade 3 tumors.(ABSTRACT TRUNCATED AT 400 WORDS)