To evaluate the separation efficiency of Ericsson's two- and three-layer albumin separation methods, semen samples from 21 healthy males were studied. Seven patients already had two or more sons, another seven had two or more daughters and the other seven had primary infertility due to female factors. The semen samples were divided into three aliquots: one remained unprocessed initially, the other two aliquots went through two- and three-layer albumin separation methods respectively. All samples were then stained with X-Y double staining probes. In each group, four or five samples were processed at room temperature, and two or three at body temperature (37 degrees C). The labelling efficiency of X-Y double staining probe was over 99%. The X:Y sperm ratios were even in the original samples. The ratios of the X and Y spermatozoa were altered slightly but significantly after the two-layer (P < or = 0.05) or the three-layer (P < or = 0.005) separation. The alterations occurred only at room temperature. The X spermatozoa increased and the Y spermatozoa decreased, both to a small degree of difference (1.4-3.5%). Double fluorescence in-situ hybridization analysis therefore showed that albumin separation methods do not enrich Y spermatozoa.