The effects of solution variables on the in vitro reconstitution of calf brain tubulin, purified by the method of Weisenberg et al. (Weisenberg, R. C., Borisy, G. G., and Taylor, E. W. (1968), Biochemistry 7, 4466-4479; Weisenberg, R. C., and Timasheff, S. N. (1970), Biochemistry 9, 4110-4116), as modified by Lee et al. (Lee, J. C., Frigon, R. P., and Timasheff, S. N. (1973), J. Biol. Chem. 248, 7253-7262), were investigated at pH 7.0. Reconstitution of microtubules was successful in a variety of buffer systems, the free energy of the propagation step of microtubule formation being little dependent on the buffer. Microtubule formation is promoted by magnesium ions and guanosine triphosphate, but inhibited by calcium ions. The dependence of the apparent association constant for microtubule formation on ligand concentration was analyzed by the linked function theory of Wyman (Wyman, J. (1964), Adv. Protein Chem. 19, 224-286), leading to the conclusion that the formation of a tubulin-tubulin contact involves the binding of one additional magnesium ion per tubulin dimer. Microtubule formation is also accompanied by the apparent binding of one additional proton and the release of water molecules, as suggested by the thermodynamic parameters determined. The reaction is entropy driven with an apparent heat capacity change, deltaCp, of -1500 +/- 500 cal/deg-mol. The enhancement of tubulin reassembly by glycerol is most likely due to nonspecific protein-solvent general thermodynamic interactions.