Two monoclonal antibodies reactive for alpha-tubulin but not for beta-tubulin have been prepared, characterized in terms of their relative binding to tubulins from different sources by a solid-phase binding assay, immunoautoradiography, and indirect immunofluorescence, and utilized to study flagellar motility. Our results demonstrate that alpha-tubulins from different species, and even from different tissues of the same species, are nonidentical. Especially interesting was the observation that one of the antibodies, Ab2, immunofluorescently stained microtubules of chick embryo fibroblast cells, but was completely unreactive for microtubules of rat kangaroo (PtK2) fibroblasts; a different antibody, Ab1, stained both cell types. Results of these and additional experiments clearly show that Ab1 and Ab2 recognize discrete and different epitopes on alpha-tubulin. Monoclonal antitubulins Ab1 and Ab2 each inhibited the bend amplitude of reactivated sea urchin spermatozoa without affecting beat frequencies or the ability of the outer doublet microtubules to slide past each other in elastase-digested models. These results, together with those obtained previously using rabbit polyclonal antitubulin antibodies [Asai and Brokaw, 1980], demonstrate that inhibition of bend amplitude is a common property of antitubulin antibodies and is not due to the binding of antibodies to one specific site on the axoneme. Our results suggest that tubulin subunit conformational changes may occur on the outer doublet lattice and may be integrally involved in the mechanism and control of flagellar bending.