The aim of this study was to obtain mice, hopefully identical multiplets, from single diploid blastomeres isolated at the 4-cell stage, or from pairs of sister blastomeres isolated at the 8-cell stage. To this end isolated blastomeres were aggregated with one or two tetraploid carrier embryos produced by electrofusion of 2-cell embryos. Diploid embryos were albino and homozygous for the "a" allele of glucose-phosphate isomerase (GPI-1a1a) and tetraploid embryos were pigmented and GPI-1b1b. The aggregates were cultured in vitro up to the blastocyst stage. Each quartet (occasionally triplet or doublet) of chimaeric blastocysts was transplanted to the oviduct of a separate pseudopregnant recipient. Altogether 62 blastocysts were transplanted to 17 recipients. Eight full-term foetuses (two singletons and three pairs of twins) were rescued by Caesarian section on day 19, 20 or 21 of pregnancy. Three young (one singleton and twins) were successfully reared by foster mothers and proved to be normal and fertile females. All foetuses and animals were albino. In five individuals only the 1-A form of GPI (characteristic for 2n blastomere) was found. In one adult female traces of the 1-B form of GPI (characteristic for 4n carrier blastomeres) were detected in the heart and the lungs while 4 other organs contained only the 1-A form. These observations strongly suggest that the majority of foetuses/animals produced according to our experimental system are 'pure' diploids rather than 2n/4n chimaeras, and that the described method can be used in future to produce twins, triplets and quadruplets in the mouse. Our study confirms earlier work by Kelly (1975, 1977) that 'quarter' blastomeres of the mouse are still totipotent.