Geometries predicted by various theoretical methods

["Mr. Mingzuo Shen" <SHEN%CCQC (+ at +) uga.cc.uga.EDU>]

Hello,
     I hope someone can help me or point to a reference on the
 nature of predicted geometries by semi-empirical SCF methods
 (such as MOPAC): that is, is a distance re (equilibrium distance),
 or some sort of thermally averaged value?
     I know that ab initio methods give re, equilibrium values.
     I know that MM2 and MM3 give rg, usually called thermally
 averaged nuclear distance. When a bond is more anharmonic,
 (rg - re) is more positive. I found a reference on this sort of things:
 K. Kuchitsu and S.J. Cyvin, Representation and experimental
 determination of the geometry of free molecules. in "Molecular
 Structures and Vibrations." S.J. Cyvin, ed. Elsevier (New York) 1972.
 pp. 183-211.
      This article lists up to 8 different geometries. I wonder
 what geometry the MOPAC predictions correspond to.
 Thanks in advance for any help!
 Mingzuo Shen
 ---