The effect of route of administration on the outcome of the micronucleus test was evaluated in 2 laboratories by administering the model chemical, 1-beta-D-arabinofuranosylcytosine (Ara-C), by intraperitoneal injection (i.p.) and oral gavage (p.o.) to 2 mouse strains, MS/Ae and CD-1. On the basis of a small-scale acute toxicity study and a pilot experiment for the micronucleus test, a full-scale test was performed with a 24-h sampling time at doses of 12.5, 25, 50, and 100 mg/kg i.p. and 25, 50, 100, and 200 mg/kg p.o. In both strains, MNPCEs were induced at lower dose levels by the i.p. treatment, as determined not only on the basis of mg/kg but also as a ratio of the LD50. When compared with other chemicals tested in this collaborative study, the effective dose levels of this chemical based on the LD50s were exceptionally low by both routes and in both strains, e.g., less than 0.3% of the LD50 by the i.p. treatment. The maximum frequencies of MNPCEs induced were, however, identical (MS/Ae) or even higher (CD-1) by the p.o. treatment.