Colchicine has been demonstrated by ion-exchange and by gelfiltration assay to bind to a protein fraction derived from the higher plant Heracleum mantegazzianum. Colchicine-binding protein from a plant source was much more unstable than tubulin from animal preparations. The tissues of Heracleum vary in their content of colchicine-binding activity. No activity was obtained from non-vascular tissue. Phloem has at most, twice as much activity as xylem. The significance of these results is discussed in relation to a proposed degree of homology between P protein of phloem and microtubule protein.