FSH mRNA is transcribed after the onset of high FSH secretion during proestrus and estrus. Pituitary cell fractions separated by size and density were studied to determine if expression of FSH mRNA activity was predominantly in one subset during the estrous cycle and to determine the source and significance of "silent FSH" cells that secrete FSH, but store too little for detection. Pituitary cells were separated by centrifugal elutriation, plated, and then exposed to 0.1-1 nM [D-Lys6]GnRH for 3 h. Media were assayed for FSH by RIA, and the cells were fixed for immunocytochemistry or in situ hybridization. The percentages of immunoreactive FSH cells in unseparated populations increased from 8% at metestrus to 12% during proestrus. Percentages of cells with FSH beta mRNA showed the same rise; however, peak levels were higher (17%) during proestrus and estrus. Small cells with FSH beta mRNA were more frequent than those with antigens early in the cycle. The largest cell fractions contained 38-44% immunoreactive cells. Only 8-21% of these cells had FSH beta mRNA, except during the morning of proestrus (33%). The distribution analyses showed that the increment in immunoreactive FSH cells during diestrus initially stemmed from smaller subsets; however, over half of immunoreactive FSH cells were large by the evening of proestrus. During the time of active transcription of FSH mRNA, more than half of the cells with FSH beta mRNA were small or medium-sized. Thus, early in the cycle, FSH beta mRNA is transcribed in the smaller cells, which may be the source of the silent FSH cells reported in previous studies. During proestrus, smaller FSH cells also secreted as well if not better than those in the unseparated population or large fractions. When they secreted more than expected from their percentages of FSH cells, this response was interpreted to be due to either the presence of cells that are immunoreactively silent or the possible removal of autocrine or paracrine regulatory factors.