We present a new determination of the parity of the neutral pion via the double Dalitz decay π 0 → e + e − e + e − . Our sample, which consists of 30 511 candidate decays, was collected from KL → π 0 π 0 π 0 decays in flight at the KTeV-E799 experiment at Fermi National Accelerator Laboratory. We confirm the negative π 0 parity, and place a limit on scalar contributions to the π 0 → e + e − e + e − decay amplitude of less than 3.3% assuming CPT conservation. The π 0 γ * γ * form factor is well described by a momentum-dependent model with a slope parameter fit to the final state phase space distribution. Additionally, we have measured the branching ratio of this mode to be B(π 0 → e + e − e + e − ) = (3.26 ± 0.18) × 10 −5 .PACS numbers: 14.40.Aq, 13.40.GpThe parity of the neutral pion has been determined indirectly by studying negative pions captured on deuterium [1,2]. The observed reactions imply that the π − is a pseudoscalar and that the parities of the π − and the π 0 are the same. It has long been known that the decay π 0 → γγ in principle offers a direct means of determining the π 0 parity through the polarizations of the photons [3,4]. Given that there are no available methods for measuring the polarization of a high-energy photon, this measurement has never been performed. However, it was soon noted that the double Dalitz decay π 0 → e + e − e + e − , which proceeds through an intermediate state with two virtual photons (see Fig. 1), is sensitive to the parity of the pion since the plane of a Dalitz pair is correlated with the polarization of the virtual photon [5,6]. This process was studied in a 1962 hydrogen bubble chamber experiment using stopping negative pion capture (π − p → nπ 0 ). That group observed 206 π 0 → e + e − e + e − events and reported that the observed distribution of the e + e − planes was consistent with a pseudoscalar pion and disfavored a scalar pion at the level of 3.6 standard deviations [7]; this experiment also produced a measurement of the branching ratio of this decay, which remains the most precise result to date. Using a sample of more than 30 000 π 0 → e + e − e + e − decays, we report new precise measurements of the properties of this decay. Our modeling of the decay includes for the first time a proper treatment of the exchange contribution to the matrix element, and consideration