Recent studies have shown that available theoretical models for critical two-phase flows in simple nozzles are not able to predict observed data. To achieve agreement, it is customary to multiply the computed mass flow rates by empirically determined factors. In this paper, a two-dimensional theory is used to show that the deficiencies are associated with geometric effects not accounted for in the earlier models. The theory presented here is applied to both equilibrium and nonequilibrium situations with considerable success and supplies a rational basis for the use of break flow multipliers.