The origin of the unoccupied nuclear oestrogen receptor (R(n)) was studied. Three working hypotheses were investigated. (a) R(n) is a dissociation product of the oestrogen occupied nuclear receptor (ER(n)). (b) ER(n) is only partially occupied, so that additional binding may occur at 0 degrees C (the temperature at which oestradiol saturates unoccupied sites). (c) R(n) is derived from the penetration of unoccupied cytoplasmic receptor (R(c)) into the nucleus. The MCF-7 cell line was used as a model in the present investigation. The amount of unoccupied receptors was measured by saturation with 7.5nm-[(3)H]oestradiol at 0 degrees C, whereas the occupied receptors were measured by exchange at 30 degrees C. The cells at preconfluency were exposed to a medium fortified with 10nm-[(3)H]oestradiol for 1h, washed and cultured up to 5 days in fresh growth medium. The distribution of oestradiol receptors was determined before exposure and during the following 5 days. After 1h exposure only ER(n) was found in the nuclear fraction. Thereafter ER(n) declined continuously so that on day 5 it approached 15% of its value measured 1h after exposure. Although after 3 days about 80% of ER(n) disappeared no R(n) appeared, which contradicts hypotheses (a) and (b). On day 4 R(n) and R(c) appeared simultaneously. The appearance of R(n) and R(c) was not prevented by culturing the cells in an oestrogen-free medium, supporting hypothesis (c). Exposure of cells to increasing concentration of [(3)H]oestradiol (0.1-10nm) for 1h resulted in a parallel increase in ER(n) without increasing the amount of unoccupied binding sites, which contradicts hypothesis (b). The present study supports the hypothesis (c), i.e., R(c) may also penetrate the nucleus without binding to oestradiol.