1: \begin{abstract}

2: The performance of two recent {\em ab initio} computational thermochemistry

3: schemes, W1 and W2 theory [J.M.L. Martin and G. de Oliveira, \JCP{111}{1843}{1999}], 

4: is assessed for an enlarged sample of thermochemical data consisting 

5: of the ionization potentials and electron affinities in the G2-1 and 

6: G2-2 sets, as well as the heats of formation in the G2-1 and a subset 

7: of the G2-2 set. We find W1 theory to be several times more accurate

8: for ionization potentials and electron affinities than commonly used

9: (and less expensive) computational thermochemistry schemes such as 

10: G2, G3, and CBS-QB3: W2 theory represents a slight improvement for 

11: electron affinities but no significant one for ionization potentials.

12: The use of a two-point $A+B/L^5$ rather than a three-point $A+B/C^L$

13: extrapolation for the SCF component greatly enhances the numerical

14: stability of the W1 method for systems with slow basis set convergence.

15: Inclusion of first-order spin-orbit coupling is essential for accurate

16: ionization potentials and electron affinities involving degenerate 

17: electronic states: inner-shell correlation is somewhat more important

18: for ionization potentials than for electron affinities, while scalar

19: relativistic effects are required for the highest accuracy. The 

20: mean deviation from experiment for the G2-1 heats of formation is 

21: within the average experimental uncertainty. W1 theory appears to

22: be  a valuable tool for obtaining benchmark quality proton affinities.

23: \end{abstract}

24: