We show that parity symmetry is not spontaneously broken in the CP N sigma model for any value of N when the coefficient of the u term becomes u p ͑mod 2p͒. The result follows from a nonperturbative analysis of the nodal structure of the vacuum functional c 0 ͑z͒. The dynamical role of sphalerons turns out to be very important for the argument. The result introduces severe constraints on the possible critical behavior of the models at u p ͑mod 2p͒ [S0031-9007(97) The effect of a CP-violating u-term interaction on the infrared behavior of quantum field theories generated substantial changes in the low energy spectrum [1]. For this reason, phenomenological requirements impose severe bounds on the actual value of the u parameter. The analysis of the response of the systems to such a topological perturbation is, however, a very rich source of information on the vacuum structure of the quantum systems at u 0 [2]. This structure is very hard to analyze by direct methods because it dwells entirely on the strong coupling regime and involves deep nonperturbative behaviors. The most relevant changes appear at u p ͑mod 2p͒. In this case, the classical Lagrangian is also CP invariant as in the absence of the u term, u 0 ͑mod 2p͒, but the behavior of the system is completely different in these two regimes.Quantum chromodynamics (QCD), the archetype of such systems, is still inaccessible for analytic studies of its prominent physical effects: confinement and chiral symmetry breaking. CP N sigma models share many similar features, such as dynamical mass generation, confinement, and asymptotic freedom, but they are simpler to analyze. In particular, for some of those models there is exact analytic information on their quantum spectrum. It is, therefore, interesting to analyze the u-vacuum effects in those systems in order to gain some insight into the similar effects in QCD 311 .The simplest model, CP 1 , is integrable for u 0 and u p, the only two values of u for which the system is classically CP invariant. However, the behavior of the system is very different in those cases. At u 0 the system is confining and exhibits a mass gap; at u p the model is massless and its critical exponents are those of the SU(2) Wess-Zumino-Witten conformal invariant model at level 1 [3,4]. In both cases, the CP symmetry is not spontaneously broken. In the second case, this can be rigorously shown in a discrete regularization of the model, where it turns out to be equivalent to a spin 1͞2 chain by Haldane transformation [5]; for those chains the Lieb-Schulz-Mattis theorem [6] establishes that there are only two possibilities: Either parity is spontaneously broken and there is a mass gap or the theory is gapless and parity is not spontaneously broken. Since it is known that the mass gap is zero for u p [3-5,7], then parity is not spontaneously broken.For higher values of N the integrability of the CP N model is lost and the only nonperturbative information comes from either large N [8] and strong coupling expansions [9] or Monte Carlo numerical si...