The binding of 9-hydroxyellipticine to calf thymus DNA, poly[d(A-T)]2, and poly[d(G-C)]2 has been studied in detail by means of CD, linear dichroism, resonance light scattering, and molecular dynamics. The transition moment polarizations of 9-hydroxyellipticine were determined in polyvinyl alcohol stretched film. Spectroscopic solution studies of the DNA/drug complex are combined with theoretical CD calculations using the final 50 ps of a series of molecular dynamics simulations as input. The spectroscopic data shows 9-hydroxyellipticine to adopt two main binding modes, one intercalative and the other a stacked binding mode involving the formation of drug oligomers in the DNA major groove. Analysis of the intercalated binding mode in poly[d(A-T)]2 suggests the 9-hydroxyellipticine hydroxyl group lies in the minor groove and hydrogen bonds to water with the pyridine ring protruding into the major groove. The stacked binding mode was examined using resonance light scattering and it was concluded that the drug was forming small oligomer stacks rather than extended aggregates. Reduced linear dichroism measurements suggested a binding geometry that precluded a minor groove binding mode where the plane of the drug makes a 45 degrees angle with the plane of the bases. Thus it was concluded that the drug stacks in the major groove. No obvious differences in the mode of binding of 9-hydroxyellipticine were observed between different DNA sequences; however, the stacked binding mode appeared to be more favorable for calf thymus DNA and poly[d(G-C)]2 than for poly[d(A-T)]2, an observation that could be explained by the slightly greater steric hindrance of the poly[d(A-T)]2 major groove. A strong concentration dependence was observed for the two binding modes where intercalation is favored at very low drug load, with stacking interactions becoming more prominent as the drug concentration is increased. Even at DNA: drug mixing ratios of 70:1 the stacked binding mode was still important for GC-rich DNAs.