Pure metals with nearly the same elastic properties as hard, covalent ceramics are often very soft, even at low temperatures. Why the difference? “Directed bonds” is the common answer. Or, “the Peierls-Nabarro stress is large”. However, the former is only qualitative, while the latter does not distinguish between covalent and metallic bonding, and is quantitatively incorrect. The rules of quantum chemistry, as developed to account for chemical reaction rates, indicate why dislocation mobilities are low in crystals bonded through hybrid sp and spd orbitals. Resistance to the motion of kinks on dislocation lines arises from the way in which the initial electronic states of these kinks are correlated with the final states (after a unit of motion has occurred). The initial bonding states correlate with the final antibonding states, and vice versa. Thus, the correlation lines cross, creating substantial barriers to the motion. The barrier magnitudes can be calculated from the energy gaps (e.g., for Sic, the HOMO-LUMO gap); together with the quadratic dependencies of bond energy on bond-bending. The calculated results agree with observations. A second prototype, TiC is also discussed.