Much effort has been invested in seeking to understand the thermodynamic basis of helix stability in both peptides and proteins. Recently, several groups have measured the helix-forming propensities of individual residues (Lyu, P.C., Liff, M.I., Marky, L.A., Kallenbach, N.R. Science 250:669-673, 1990; O'Neil, K.T., DeGrado, W.F. Science 250:646-651, 1990; Padmanabhan, S., Marqusee, S., Ridgeway, T., Laue, T.M., Baldwin, R.L. Nature (London) 344:268-270, 1990). Using Monte Carlo computer simulations, we tested the hypothesis that these differences in measured helix-forming propensity are due primarily to loss of side chain conformational entropy upon helix formation (Creamer, T.P., Rose, G.D. Proc. Natl. Acad. Sci. U.S.A. 89:5937-5941, 1992). Our previous study employed a rigid helix backbone, which is here generalized to a completely flexible helix model in order to ensure that earlier results were not a methodological artifact. Using this flexible model, side chain rotamer distributions and entropy losses are calculated and shown to agree with those obtained earlier. We note that the side chain conformational entropy calculated for Trp in our previous study was in error; a corrected value is presented. Extending earlier work, calculated entropy losses are found to correlate strongly with recent helix propensity scales derived from substitutions made within protein helices (Horovitz, A., Matthews, J.M., Fersht, A.R. J. Mol. Biol. 227:560-568, 1992; Blaber, M., Zhang, X.-J., Matthews, B.M. Science 260:1637-1640, 1993). In contrast, little correlation is found between these helix propensity scales and the accessible surface area buried upon formation of a model polyalanyl alpha-helix. Taken in sum, our results indicate that loss of side chain entropy is a major determinant of the helix-forming tendency of residues in both peptide and protein helices.