Precision is given to the concept of electronegativity. It is the negative of the chemical potential (the Lagrange multiplier for the normalization constraint) in the Hohenberg–Kohn density functional theory of the ground state: χ=−μ=−(∂*E*/∂*N*)_{v}. Electronegativity is constant throughout an atom or molecule, and constant from orbital to orbital within an atom or molecule. Definitions are given of the concepts of an atom in a molecule and of a valence state of an atom in a molecule, and it is shown how valence‐state electronegativity differences drive charge transfers on molecule formation. An equation of Gibbs–Duhem type is given for the change of electronegativity from one situation to another, and some discussion is given of certain relations among energy components discovered by Fraga.

## REFERENCES

*Proceedings of the*1966

*Midwest Conference on Theoretical Physics*(Indiana U., Bloomington, IN, 1967), p. 14. The proof follows the lines of the proof for the nondegenerate case (Ref. 1), requiring only the additional observation that if ρ is a ground state density for two distinct potentials υ and $\upsilon \u2032,$ some elgenfunction ψ among the degenerate solutions for υ, and some eigenfunction $\psi \u2032$ among the degenerate solutions for $\upsilon \u2032,$ both produce the same ρ.

*E*, but this may not always be the case—see Eq. (22). This is like the situation in Hartree‐Fock theory, where the sum of orbital energies is almost always a minimum for the ground state. See

*Calculus of Variations*(Prentice‐Hall, Englewood Cliffs, NJ, 1963).

*Q*may be written in the form $\u222bP(1)d\tau 1,$ where

*P*is an appropriately defined local pressure in the electron gas. This point of view is being studied in this laboratory by Dr. L. Bartolotti.

*Die Statistische Theorie des Atoms und Ihre Anwendung*(Springer, Vienna, 1949).

*Tables of Hartree‐Fock Atomic Data*,” Tech. Rep. Dept. of Chem. Univ. of Alberta, Edmonton, 1973, p. 26.

*T*through a standard formula. Correct functionals $ENO[nk,\psi k]$ could be either explicit density matrix functionals or trivial transcriptions of the density functional using the relation ${nk,\psi k}\u21d2\rho .$ The latter viewpoint is emphasized in the present paper, the former in Ref. 28.

*The Nature of the Chemical Bond*(Cornell U.P. Ithaca, NY, 1960), 3rd ed. Eq. (3–10), p. 91.

*Chemical Bonds and Bond Energy*, (Academic, New York, 1976), 2nd ed.