Attempts at in vitro fertilization can fail at two points: during the fertilization itself or during embryo transfer. In the case of implantation failures, two causes can be invoked: the condition of the endometrium and the quality of the embryo. It can be assumed that after ovarian stimulation or after the loss of large quantities of follicular fluid during oocyte recovery, the postovulatory phase, and consequently the condition of the endometrium, may be atypical, or that the synchronism and the menstrual cycle is no longer perfect. To date, experimental, biochemical, and morphological observations have not confirmed this hypothesis. The condition of the embryo appears to be the predominant factor. The criteria for evaluating its quality are not all perfectly objective. Those which are truly objective are obtained a posteriori by invasive methods after the death of the embryo, the other criteria, obtained with the living embryo, entail a certain element of subjectivity. Non-invasive methods: By observation, the size and form of the embryo can be evaluated, as well as the equality of blastomere size, the presence or absence of anuclear cellular fragments, the number of nuclei per blastomere, the granular or clear appearance of the egg or the blastomeres, and the appearance of the cumulus. The rate of development is an important criterion: the best chances for pregnancy are obtained when the 2-pronuclei stage occurs no later than 20 h, after insemination, the 2-blastomere stage no later than 35 h., and the 4-blastomere stage no later than 45 h. Invasive methods: The intensity of fluorescence using fluorescein diacetate. Karyotype containing numerous chromosome anomalies.(ABSTRACT TRUNCATED AT 250 WORDS)