The effect of hypoxia on micronucleus formation in mice was evaluated using special "shoebox" cages containing a silicone rubber membrane that effectively excludes the replacement of cage oxygen as it is spent by group-housed mice. In five separate trials using between 13 and 17 mice per cage, the oxygen level in the cage was reduced from the normal 21% to an average of 7.5% in 3-7 days. Compared to control animals, which exhibited a spontaneous micronuclei (MN) frequency of 0.08%, mice living under hypoxic conditions exhibited a statistically significant (p < or = 0.01) elevation in MN frequency of 0.15%. These findings may be relevant to in vivo studies in which drug or chemical treatment results in hypoactivity or complete sedation accompanied by varying degrees of hypoxia. It is proposed that structurally unexpected positive clastogenicity findings in animals exposed to these conditions may be due to hypoxia rather than to drug-DNA interaction. The likely mechanistic basis for this effect, decreased fidelity of cell division under conditions of elevated erythropoiesis, is discussed.