On p. 229, on top of the left column, the phrase which reads: "the number of nodes used for the results presented here was lo-12" should instead read "the number of nodes used for the results presented here was 5-7". On p. 302, Section 5, the phrase "a similar potential energy surface" should read "another potential energy surface". In Table 6, "Ref.
[ 19 ] " should read "Ref.
[ 20]", and "Refs. [ 17,181" should read "Refs. [ 18,19 ] ". In Ref. [ 61, "1993" should read "1990".
An error was found in the scheme used to construct the symmetry-adapted basis sets. This error resulted in incorrect eigenenergies. The scheme for restricting the quantum numbers of the basis functions is described on p. 299 of the original Letter. We also take advantage of the identical-particle symmetry of the ( HF)l system, as follows. The primitive basis functions .X&r, d) (Eq. ( 6), p. 298) are symmetric with respect to simultaneous interchange of the vibrational and rotational quantum numbers associated with monomer 1 and monomer 2, respectively, i.e. Xvv,uuti,21(~, &) and Xvti zvyu, z,(~, 2) are identical within a sign. Thus, within the confines of the scheme on p. 299 (and the parity and interchange symmetry requirements), we generate all allowable values of vI and v2, but we only generate those j, and j, which satisfy j, >j,. When j, =j, one needs only those basis functions with v, > v2. We have now corrected an error in the way the symmetry-adapted basis functions were generated, and we recomputed the ten lowest eigenvalues from each symmetry block; the corrections to Tables 2-5 are given here in Table 1.
The table shows that the corrected calculation of the tunneling splitting for the Bunker et al. surface studied in this Letter is 0.47 cm-', and this value agrees within 0.01 cm-' with our most well-converged rigid-rotor results. This value agrees within 30% with the experimental value of 0.66 cm-', and it agrees excellently with recent full-dimensional calculations by Zhang et al. [ 11, which gave 0.46 cm-'. We have also confirmed that our corrected code agrees well with the results of Zhang and BaEiC for other energy levels [ 21 on the Bunker et al. surface and for many energy levels [ I] on the SQSBDE potential surface of Quack and Suhm and with new calculations [ 3 ] of the tunneling splitting and Do on the latter surface by Quack and Suhm [ 3 1. These checks confirm the accuracy of our corrected code. Thus our conclusion that the potential energy surface of Bunker et al. yielded an erroneously large tunneling splitting was incorrect. * SSDI of original article: 0009-26 14(94)00505-K.