Although relativity can play an important role in the prediction of atomic or molecular spectra for all elements, the chemical consequences of relativity typically become evident only in the fifth row, and occasionally in the fourth row. In the fifth row, the relativistic effects are so large as to make even quantitative assessment of the atomic or molecular electronic structure and chemical behavior impossible without considering the orbital contraction, orbital expansion, and spin-orbit effects. The major differences in chemical behavior predicted by the relativistic model in comparison to the non-relativistic model are bond contraction and changes in valency. The bond contraction has been found to be directly related to the s and p orbital contraction[37,38], in so far as both are manifestations of the relativistic lowering of the kinetic energy in the core and intermolecular region. The changes in valency are rooted in the stabilization of s and p orbitals, the destabilization of d and f orbitals, and spin orbit splitting.
In the first and second columns, where the valence occupancy is not
dependent upon the relative energetics of the outer shells, the primary
relativistic effect is bond contraction.
The relativistic
contraction of the outer s and p shells of Fr and Ra is so great, that
these species actually exhibit smaller atomic radii than the lighter Cs and
Ba, respectively[39].
The actual extent to which 
-bonds involving the valence s orbital of
these elements are contracted relative to the non-relativistic case has not
been established.
The third column of the main group elements exhibits some of the most
interesting relativistic effects of all the elements in the main group.
The valence occupancy of these elements is 
.
For all the elements up to and including In, the 
shell is relatively
close in energy and spatial extent to the np shell, and these shells can
hybridize to allow for trivalent bonding. For Tl, however, two effects are
manifest: the ns shell is dramatically stabilized, relative to the more
modest stabilization of the 
shell, while the
and 
shells experience a strong spin-orbit splitting.
The result is that the s electrons become relatively inert, and that Tl
appears to be mono-valent.
Main group columns four through seven exhibit progressively less pronounced relativistic effects. Lead exhibits a preference for a valency of 2, in contrast to the lighter elements which exhibit a valency of 4. This change in valency can be explained in much the same fashion as was the valency of 1 for Tl, except that for lead, the spin-orbit splitting of the two p shells is not as prominent. Excellent explorations of the importance of relativistic effects in the chemistry of the group IV di- and tetra-hydrides have been performed by Dyall[40,41] and others[42].
Heavy transition metals exhibit the strongest relativistic effects of all
the elements.
One of the most dramatic examples is that of the contrast between
silver and gold.
When the non-relativistic predictions of the electronic spectrum of silver and
gold atoms are compared, no striking differences appear, and one would
expect Au to behave in an extremely similar manner to Ag, both in its
electronic spectrum and in its chemical interaction with other species.
What is observed, however, is that gold absorbs strongly in the visible
region, thereby giving the metal its yellow luster.
An explanation of this difference can be found in the relativistic
prediction of the electronic structure. Gold has a ground state occupation
of 
. The relativistic destabilization of the 5d
orbitals and the stabilization of the 6s narrows the energetic gap between
the two orbitals, and allows for excitation of the d electrons
within the visible range in the bulk metal. Another interesting
manifestation of the large relativistic effects experience by the gold atom
is found in the bond length of AuH. The Hartree-Fock method predicts an AuH
bond length which is significantly longer than the AgH bond
length[43,44] in contrast to the
experimental observation that AuH is shorter. The Dirac-Hartree-Fock
method, on the other hand, predicts an AuH bond that is 0.2Å shorter
than the predicted AgH bond, in qualitative agreement with experiment.
This dramatic evidence of the necessity for relativistic treatment in heavy
element chemistry has given coinage metal containing compounds
a prominent role in relativistic quantum chemistry, both as textbook
examples and as test systems for approximate methods.