The survival of cleaving embryos after freezing and thawing has been assessed. First, comparisons were made of the proportions of embryos in which all blastomeres were viable cells after thawing, following various forms of ovarian stimulation. A flare-up protocol using a GnRH-agonist (buserelin) produced significantly higher numbers of these embryos than a pituitary down-regulation protocol (P less than 0.05), though neither was significantly different from clomiphene citrate/HMG stimulation. Secondly, other parameters of embryo survival e.g. proportions with one or more surviving cells and pregnancy rates were assessed and were similar among stimulation protocols and treatments in the embryo replacement cycle. Survival of blastomeres in 2- to 8-cell embryos was inversely related to the theoretical total surface area of all blastomeres in the embryo. Thawed embryos with one or more blastomeres damaged during freezing had the same capacity to produce pregnancies as did those with all blastomeres intact. The survival of individual cells was clearly related to the stage at which the cleaving embryo is frozen, but moderate loss of cells does not significantly influence implantation.