PACS. 82.70.Dd -Colloids. PACS. 64.10.+h -General theory of equations of state and phase equilibria. PACS. 64.60.Cn -Order-disorder transformations; statistical mechanics of model systems.Abstract. -We describe a charge-stabilized colloidal suspension within a Poisson-Boltzmann cell model and calculate the free energy as well as the compressibility as a function of colloidal density. The same quantities are also calculated from the linearized Poisson-Boltzmann equation. Comparing nonlinear with linear theory, we test the quality of different linearization schemes. For concentrated suspensions, linearization about the Donnan potential is shown to be preferable to standard Debye-Hückel linearization. We also show that the volume term theory proposed earlier follows from a linearization about the Donnan potential. Using this linearization scheme, we find a gas-liquid phase coexistence in linear, but not in nonlinear theory. This result may imply that predictions of a spinodal instability in highly de-ionized colloidal suspensions are spurious.Suspensions of charged colloidal particles at low salt concentrations have attracted much attention recently, mainly because of experimental claims of attractive interactions between like-charged colloids and the possibility of a gas-liquid phase transition [1][2][3]. Such phenomena cannot be explained by the standard and well-accepted DLVO theory [4], which predicts that screened-Coulomb repulsions between the like-charged colloids completely mask the van der Waals attractions in the low-salt regime of interest here. On this basis it is often argued that there is no cohesive energy that can stabilize a liquid phase. It is important to realize, however, that such a consideration is based on the implicit assumption that the (effective) interaction Hamiltonian of the colloids is a sum of pair potentials, i.e. that three-and more-body interactions are being ignored. This assumption is expected to be valid at high salt concentrations, where the screening length λ is much shorter than the colloidal diameter a, but may well fail if λ a, i.e. at salt concentrations in the micro-mole regime. Recent attempts to effectively take into account these higher-body interactions (in the form of density-dependent "volume terms" that include the free energy of double layers with their "own" colloidal particle) are based on i) considering the total free energy of the suspension, including the "volume terms" and ii) including the counterion density to the screening, i.e. the screening becomes colloiddensity-dependent [5][6][7][8]. The combination of these two ingredients can explain (some of) the experimental observations qualitatively, since the resulting free energy gives rise to phase coexistence of a dilute (gas) phase with dense (liquid or crystal) phases; the cohesive energy that stabilizes the dense phases is provided by the Coulomb energy of the relatively compressed