A model system of 15% exfoliated tumor (Ehrlich ascites) cells and 85% exfoliated cervicovaginal human squamous cells and acellular viscous mucuslike material was used to investigate the distribution of tumor cells on smear preparations. The smears were taken by the same individual, with plastic spatulas, using traditional smearing motions. The presence or absence of normal squamous cervicovaginal cells and tumor cells was recorded in 400 randomly distributed areas. Smears done with circular motions contained the largest number of tumor cells and those done with back-and-forth strokes, the least. Circular motions apparently should be preferred if the purpose is to transfer to slides the largest possible number of tumor cells. Independently of the motion employed, the largest total number of tumor cells was found in ring 9 and the lowest, in ring 5. When the various motions were taken into account, the largest number of tumor cells was found in ring 9 for smears done with zigzag strokes; the lowest, in ring 5 with clockwise circular motions. Despite these differences, comparative studies indicate that plastic spatulas may induce a more even distribution of tumor cells onto slides than wooden spatulas or cotton swab applicators. We have been unable to confirm that plastic spatulas induce better cytologic preparations than do wooden spatulas or cotton swab applicators.