The majority of in-vitro-derived human preimplantation embryos are chromosomally abnormal but whether the same pattern exists in vivo is unknown. This would be impossible to demonstrate in humans. Hence we chose murine embryos to study this difference owing to their ease of manipulation and compared the incidence of mosaicism between in-vivo- and in-vitro-cultured embryos. Two groups of embryos were analysed. Group A (in vitro) were obtained 48h following superovulation and cultured in vitro until the blastocyst stage. Fluorescent in-situ hybridization (FISH) was performed at different stages that included the cleavage, morula and blastocyst stage. Group B (in vivo) were obtained on day 2 or day 5 and FISH was performed immediately without culture. There was an increase in chromosomal mosaicism seen from the cleavage stage up to the blastocyst stage in the in-vitro culture group. Overall chromosomal abnormality from day 3 to day 5 was found to be 30% (28/94) in group A. The incidence of chromosomal abnormalities in blastocysts from group B was significantly lower than group A blastocysts (8% (3/40) and 31% (20/64) respectively; P<0.05). These data show that in-vitro cultured embryos had a significantly higher incidence of mosaicisim in comparison with the in-vivo group. Cultured human embryos show high levels of chromosomal abnormalities but whether this is a pattern seen in all embryos or is the result of culture is unknown. To study this pattern we used mouse embryos and carried out chromosome analysis by fluorescent in-situ hybridization. We compared embryos that were cultured (in vitro) with those that were not (in vivo, i.e. grown exclusively in the mouse). We found that cultured embryos showed significantly higher chromosomal abnormalities as compared with in vivo embryos. This suggests that certain culture conditions are responsible for the high level of chromosomal abnormalities seen in these embryos, which should be investigated further.