Tritiated thymidine has been widely used as a nuclear marker of cell birth. The true diameters and packing densities (nuclei/microns 3) of such radioactively labeled nuclei cannot be measured directly from tissue sections. Here we show that existing stereological corrections cannot be applied to data from radioactively labeled nuclei. We empirically measured the number of silver grains exposed by nuclei containing tritiated thymidine. The nuclei were separated from the photographic emulsion by known thicknesses of fixed, embedded avian telencephalon. The results of this experiment were used to develop an equation that estimates the number of silver grains exposed by a cell nucleus of any given diameter, containing a given amount of radioactive label, and located at any given distance from the photographic emulsion. The equation also allows one to calculate the probability that a label-containing nucleus will be correctly classified as labeled. Simulations of the equation revealed that not all label-containing nuclei are correctly classified by using commonly employed identification procedures and that larger nuclei are less likely to be correctly classified than smaller nuclei, given the same amount of label. The equation can be used to modify one class of existing stereological equations so as to be applicable to measurements of radioactively labeled nuclei. Finally, we discuss the assumptions and limitations of this modification.