Our methodology for producing a spectroscopically determined force field (SDFF) (i.e., a molecular mechanics energy function) that, in addition to structures and relative energies, reproduces vibrational frequencies to spectroscopic standards, has been extended from its previous implementation on the linear to include branched saturated hydrocarbon chains. To the ab initio force fields of the 14 stable conformers of n-pentane and n-hexane, we have now added those of the 7 stable conformers of isopentane, 3-methylpentane, and neopentane, plus specific force constants from a secondary set of branched molecules, to optimize the parameters of the SDFF. This SDFF reproduces 791 ab initio non-CH stretch frequencies with a root-mean-square deviation of 6.2 cm(-1). When applied to other molecules not in the optimization set, viz., cyclobutane, cyclohexane, isobutane, tri-tert-butylmethane, and tetra-tert-butylmethane, not only are ab initio as well as experimental geometries and frequencies well reproduced, but the correct (reassigned) tertiary CII stretching frequency in tri-tert-butylmethane is satisfactorily predicted. The larger frequency deviations for other modes of this molecule provide an unusual insight into sensitive features of the nonbonded interaction terms in the potential function.