The formalism of Ursell operators provides a self-consistent integral equat
ion for the one-particle reduced operator. In three dimensions this techniq
ue yields values of the shift in the Bose-Einstein condensation (BEC) trans
ition temperature, as a function of the scattering length, that are in good
agreement with those of Green's function and quantum Monte Carlo methods.
We have applied the same equations to a uniform two-dimensional system and
find that, as we alter the chemical potential, an instability develops so t
hat the self-consistent equations no longer have a solution. This instabili
ty, which seems to indicate that interactions restore a transition, occurs
at a non-zero value of an effective chemical potential. The non-linear equa
tions are limited to temperatures greater than or equal to Tc, so that they
do not indicate the nature of the new stable state: belt we speculate conc
erning whether it is a Kosterlitz-Thouless state or a "smeared" BEC, which
might avoid any violation of the Hohenberg theorem, as described in an acco
mpanying paper.