In many studies the relationship between a bond’s length and the angle formed with an adjacent bond has been shown to be inverse; larger bond angles result in shorter bond lengths. Mastryukov and co-workers, however, have found some bonds that lengthen at extreme angles. In this work, the bond length/bond angle relationship was reexamined to gain understanding of the normal behavior and this occasional turnover in behavior at large angles. The correlation between bond lengths and adjacent bond angles in CX(4), CXH(3), and CHX(3) (X = H, Cl, F) has been studied by extended Huckel and air initio Hartree-Fock/STO-3G techniques. In these molecules, the angle between three of the bonds and the fourth unique bond was varied, thus retaining C-3v symmetry. We examined the overlap populations between each of the atoms and the central carbon as a function of angle. In the case of the ab initio calculations, we also allowed all bond lengths to relax. We present arguments to explain the normal and direct behavior on the basis of overlap populations, Walsh diagrams, ab initio Optimized bond lengths, and consideration of the nonbonded, X ... X, interactions. We are able to explain the anomalous behavior at large angles in two ways : (a) from a decomposition of the molecular orbitals and the delicate balance of s and p character in them and (b) from the adjustments the electronic structure makes to the uncomfortably short X ... X contacts.