Band 3 catalyzes the one-for-one exchange of monovalent anions across the red cell membrane. At least two anion binding sites have been postulated to exist on the transport unit: 1) a transport site that has been observed by saturation kinetics and by 35 Cl NMR studies of chloride binding, and 2) a 35Cl NMR-invisible inhibitory site that has been proposed to explain the inhibition of anion exchange at large anion concentrations. A number of independent studies have indicated that the transport site is alternately exposed to different sides of the membrane during the transport cycle. Yet the role, if any, of the postulated inhibitory site in the transport cycle is not known. Here it is shown that: 1) when the [Cl-], [Br-], or pH is varied, the band 3 transport sites on both sides of the membrane behave like a homogeneous population of simple anion binding sites in 35Cl NMR experiments, and 2) when the [Cl-] is varied, the outward-facing transport site behaves like a simple anion binding site. These results indicate that the postulated inhibitory site has no effect on chloride binding to the transport site. Instead, the results are quantitatively consistent with the ping-pong model (Gunn, R. B., and Frölich, O. (1979) J. Gen. Physiol. 74, 351-374), which states that the transport site is the only site involved in the transport cycle. Expressions are derived for the macroscopically observed characteristics of a ping-pong transporter: these characteristics are shown to be weighted averages of the microscopic properties of the inward- and outward-facing conformations of the transport site. In addition to supporting the simplicity of the transport mechanism, the high pH titration curve for chloride binding to the transport site provides insight into the structure of the site. The macroscopically observed pKA = 11.1 +/- 0.1 in the leaky ghost system indicates that an arginine must provide the essential positive charge in the inward- or outward-facing conformation of the transport site, or in both conformations.