The folded native state of a globular protein is noted to be marginally stabilized over the structureless unfolded state by various atomic/group interactions. Quantitative enumeration of these factors remains to be a difficult task to workers in the field. In this work we collect experimental information about stability factors, such as, hydrogen bonds, ion-pairs, disulfide bridges, and semi-empirical information about hydrophobic and van der Waals interactions from the knowledge of the crystal structures of a large set of globular proteins, and translate them into free energy contributions to the stability of the folded state. Taking the experimental conformational stability (the free energy difference between the folded and unfolded states) of 14 globular proteins, and computed free energies for the stability factors, we set up multiple regression equations to predict protein stability. The outcome is an excellent agreement between the experimental and theoretical predictions for each of the 14 proteins, when they do not form part of the input in the regression analysis, and the stability prediction of 24 other proteins for which there are no experimental results on stability. Interestingly, many proteins are noted to have (\Delta G) values in the range of (30-40 \mathrm{kcal} \mathrm{mol}^{-1}), much more than the usually stated range, ie. 5-20 k kal mol ({ }^{-1}). There is no relationship between (\Delta G) and (N), the number of residues in a protein. Though hydrophobic factor is a dominant force, hydrogen bonds do contribute significantly to the stability of the folded state, the other factors contributing only marginally. The relative contributions from the different factors point to a picture of protein stability in the following way: The hydrophobic force increasingly drives the polypeptide chain to the folded state overcoming the entropic factor, while the other factors, especially hydrogen bonds and van der Waals attraction, define the shape and keep it from falling apart.