Why is the molecular formula of a gas not necessarily the same as its empirical formula?

Question

At higher temperatures, those that produce a gas, say, from a liquid, the kinetic energy of the molecule is large enough to break those bonds holding more than one molecule together. Thus, H2O, in the liquid state and at low temperatures, occurs in units of six, at higher temperatures the units are groups of 4 and at the boiling point of water, in the vapor state, it is H2O in unimolecular form.