To conclude this section, let us inquire whether we can say anything rigorous about
the occurrence or not of (simple) BEC in an extended system in thermodynamic
equilibrium, or failing that at least about the condensate fraction as a function of
the system parameters (density n, temperature T, . . .). Actually, results in this area
are rather few. More than 40 years ago, Gavoret and Nozi`eres (1964) showed in a
classic paper that BEC persists in an interacting gas in three-dimensional free space
at T = 0 provided that perturbation theory in the interaction converges (a premise
which excludes solid 4He as a counterexample); however, this method cannot be used to
set a nontrivial limit on the transition temperature Tc, nor on the condensate fraction
at T = 0. More recently Kennedy et al. (1988) proved the existence, at T = 0, of BEC
in a “lattice gas” at half filling without relying on perturbation theory. However, the
existence of BEC has not been proven even at T = 0, to my knowledge, for any other
extended system with short-range interactions.