Controlled ovarian stimulation during an in vitro fertilization cycle usually produces large numbers of oocytes and, consequently, it is likely that more embryos will be generated than can be transferred in a given cycle. It is desirable to freeze-bank surplus oocytes before insemination to avoid the ethical and legal complications of disposing of or storing embryos. Although many attempts have been made to cryopreserve human oocytes, to date, post-thaw survival has been poor, and viable pregnancies after in vitro fertilization have been rare. A possible explanation for the lack of success is that the freezing methods have been adapted from animal studies but have not been optimized for the human oocyte. In this study, video microscopy was used to determine the volumetric responses of mature human oocytes to changes in osmolarity during preparation for freezing. A Boyle van't Hoff plot of data collected in static experiments with fresh human oocytes gave a value of 0.19 +/- 0.01 (mean +/- SEM) for the osmotically inactive volume. Dynamic measurements during exposure to dimethyl sulphoxide at room temperature (22 degrees C) were analysed by a two-parameter transport model and produced values of 1.30 x 10(-6) cm atm-1 s-1 for the hydraulic conductivity of the plasma membrane and 3.15 x 10(-5) cm s-1 for dimethyl sulphoxide permeability (chi-squared = 0.43, df = 20) of fresh human oocytes. Oocytes that had failed to fertilize had a slightly lower hydraulic conductivity and dimethyl sulphoxide permeability and, after exposure to 1.5 mol dimethyl sulphoxide l-1, these cells appeared to become permeable to normally impermeable solutes. These permeability properties have been used to design a protocol for the addition and removal of dimethyl sulphoxide to control the magnitude of volumetric changes.