In addition to the intrinsic proclivities of various types of amino acids for different backbone conformations, a detailed description of secondary conformations in globular proteins should account for the important effects of position within the regular secondary regions. These position effects are expected to be largest for ionized and polar amino acids. We have formulated a simple method to estimate the free energies of regular secondary regions for polypeptides of specified sequence. For all possible regions of regular conformation, intraregion electrostatic interactions are calculated explicitly, in the approximation of fixed side chain positions. The free energies of backbone-backbone and backbone@ atom interactions are treated as a single conformation-dependent parameter. These parameters are taken to be identical for all amino acids except glycine and proline. The principal parameter, the energy difference between a helix and @ strand, is related in a simple way to the experimental ratio of the number of helical residues to the number of p strand residues. An energy minimum for the molecule is determined by a dynamic programming scheme which is rigorous if the total molecular energy is given in the form of a sum of energies of the independent secondary regions. Four standard backbone conformations are considered. The neglect of specific solvent effects and long-range interactions means that results will correspond to early stages of folding. For six proteins, results compared with reported X-ray conformations are correct, on average, for 65% of all residues and for 83% of residues within regular secondary regions. Backbone-backbone and backbone-side chain interactions are most important in determining secondary structures, with side chain-side chain interactions appearing to make a relatively small contribution. Two other methods utilized to select a best set of molecular conformers are: (1) “conformational stability” selection and (2) a priori conformational probabilities calculated with a partition function representing an equilibrium mixture of all combinations of independent secondary regions. All methods yield similar results. Inspection of conformational probabilities reveals that a single conformation is highly favored for many residues; for these residues that conformation is chosen, regardless of the selection method. Apparently many of these conformations are so stable as to be retained at later folding stages, after specific solvent and long-range interactions are imposed. The similar quality of these and other reported secondary structure prediction methods, in spite of their diversity, implies that the native conformations of globular proteins are most likely maintained through redundant interactions, with significant electrostatic intraregion stabilization.