The methods of either image or flow cytometry applied to developing bladder tumors in rats requires satisfactory cell samples and a system for cell classification that is related to the lesions from which the cells are derived. Seven- to eight-week-old male Fisher 344 rats were fed 0.05% of the carcinogen N-butyl-4-N-hydroxybutylnitrosamine (BBN) in drinking water for 10 wk and then returned to tap water. Animals were killed at 14, 26, 34, 45 and 62 wk after the start of carcinogen feeding. Age-matched untreated animals were controls. Bladders were fixed, embedded, sectioned, and histologically evaluated, or were dissociated with a trypsin/EDTA solution into single cells that were Papanicolaou stained and evaluated for cytopathologic changes. Overnight urines were collected before killing; urine sediments were Papanicolaou stained and evaluated. Histologic features were hyperplasia at 14 wk, followed by slowly progressing papillary transitional cell tumors that eventually led to invasive carcinoma and were similar to those reported for this animal model. Treated animals had cytologic features of repair at 14 and 26 wk and neoplastic features at 45 and 62 wk. Both reparative and neoplastic changes were found at 34 wk. Cells were much more numerous in urines from treated rats (greater than 1,000 per sample) than in urines from controls (less than 1,000 per sample). Although exfoliated cells in urine samples were generally of poor quality, as many as 11% of cells were adequately preserved.(ABSTRACT TRUNCATED AT 250 WORDS)