Introduction. Since Block et al., 1953 introduced the term “preleukemia” for certain states of hemopoietic insufficiency preceding acute leukemia, discussion about the nature of this disease has never ceased. Systematic research on this question should be based on a better understanding of the process of leukemogenesis, for which studies in an animal model could be useful. So far, observations have rarely been made during the phase preceding clinical manifestation of induced leukemia, and hematological changes comparable to human “preleukemia” have not been reported. The development of such changes during the induction phase of acute leukemia is described in this paper. Material and methods Leukemia was induced according to the method of Huggins et al., 1966 by pulse-doses of 7,12-DMBA. Leukemia is reported to develop in about 80% of animals within a hundred days. Fifteen male rats of the Wistar-strain, 25 days of age, received up to five times 35 mg/kg body weight, every ten days. Animals were killed or died spontaneously during the following time-spans: six before day 40, five between day 40 and 60, four after day 60. Experiments were done with controls. Blood for cell counts and smears was obtained every ten days up to day 65 and on day 95. Bone marrow was examined during life up to four times by femoral puncture and iliac crest biopsy as well as extensively post mortem. Cyto(histo)chemical reactions included peroxidase, acid phosphatase, and naphtol-AS-Dchloracetate-esterase. Morphometric analysis of bone marrow sections was done with the Zeiss integration plate II at 250 × magnification. Results express bone marrow constituents as a percentage of whole volume. Results. Four of 15 DMBA-injected rats developed nodular blastic infiltrates in bone marrow and/or spleen and liver (days 47, 65, 85, and 156). In the peripheral blood, few blasts were seen even during the terminal stage of disease. One animal showed only small aggregations of blast cells in splenic sinuses when killed on day 104. None of the control animals developed leukemia. In all DMBA-treated rats the bone marrow was aplastic up to day 40, presenting first with signs of “acute myelitis“, later with compensatory growth of fat tissue. Varying degrees of peripheral cytopenia, especially a macrocytic, hyperchromic anemia, are in accordance with this finding. After day 40 in five of nine rats still alive, predominantly erythropoietic marrow hyperplasia evolved while anemia still persisted. These five rats later developed leukemia. In addition, dysplastic signs of blood and bone marrow cells were found. Peripheral giant thrombocytes, nuclear deformities and atypical mitoses of erythropoiesis and siderin-pigment deposits in reticulum cells appeared in nearly all DMBA-treated rats from day 10 onwards. The following dysplastic changes, however, were seen only in rats in which leukemia was detected later, and were most numerous during the phase of marrow hyperplasia. Thirty and 50 days prior to the onset of leukemia, increasing aniso- and poikilocytosis as well as polychromatophilia of erythrocytes and massive expulsion of erythroblasts were noticed in two rats. Most interesting was the occurrence of lymphocyte-sized mononuclear cells in peripheral blood with a granular positive peroxidase-reaction (3/5), and of larger monocytoid cells with an intensity of peroxidase-reaction stronger than in monocytes and weaker than in myelocytes (4/5). . Moreover, a coarsening of monocytic and neutrophilic granulation was striking in the peroxidaseand acid phosphatase- reaction as well as in the Pappenheim-stain (5/5). None of the above-mentioned changes could be detected in control animals. Discussion. The described myelopoietic dysplasias are interpreted as signs of disturbance in bone-marrow proliferation, differentiation, and maturation, which must be located at the stem cell level since all myelopoietic cell lines are affected. They are therefore comparable to human “preleukemia”. Correlation of regeneration, dysplasia, and leukemia in our experiment is discussed and related to some clinical observations in human “preleukemia”. It is felt that further studies with this experimental design can serve as a model for human “preleukemia”.