ABSTRACT
The aim of performing quantum chemical calculations may be twofold. There
are cases when we are just interested in getting a number, for instance to de-
termine what conformer of a molecule is more stable or to estimate the barrier
of a chemical reaction. Another purpose may be understanding or interpret-
ing one or another property of the system studied-may be discussing the
question of why the number just mentioned is as large or small as it is, and
how it can depend on a substitution or other changes in the system. It is true
that, in principle, any possible information concerning a molecular system
is included in its wave function, but any direct interpretation of the latter is
out of range of our capabilities in most cases of practical interest. This is
the case because the wave function of a molecule is a multidimensional-
may be complex-valued-function depending on a large number of variables
which is very difficult to visualize even in the simplest cases. In the prac-
tical (necessarily approximate but more and more accurate) calculations the
wave functions are represented by enormously big sets of numbers of dif-
ferent types-it is again hard to do anything directly with them. A further
disturbing factor, discouraging the direct analysis of wave functions, may be
connected with the fact that often the same wave function can be given in
several, quite different but fully equivalent, forms; for instance it can be writ-
ten down either by using the delocalized “canonical” molecular orbitals or in
terms of some localized ones which can be assigned to different fragments of
the molecule.