Not long ago Rankin (1990) presented strong evidence in favor of a low altitude (r ≈ R*) dipole geometry for the site of the core component of pulsar radio emission. Arons (1993) gave evidence that spun up millisecond pulsars must have a substantially dipolar large scale field at low altitute. Electron-positron pair creation at low altitude above the polar caps has long been hypothesized to be an essential ingredient of pulsar radio emission. If so, all observed pulsars must lie in the region of P − Ṗ space where polar cap acceleration has sufficient vigor to lead to the copious pair production. Yet, to date, all internally consistent theories of polar cap pair creation have required hypothesizing a large scale (eg, quadrupole) component of the magnetic field with strength comparable to that of the dipole, in contradiction with the evidence in favor of an apparently dipolar low altitude geometry. The internally consistent theories also violate other observational constraints. The discharge models of Ruderman and Sutherland (1975), Gurevich and Istomin (1985), and Jones (1977, 1978, 1979) all accelerate equal, counterstreaming flows of electrons and positrons, thus putting one half of the particle acceleration energy into high energy particle and photon bombardment of the polar caps. The heating causes pulsed thermal X-ray emission from hot spots in excess of what is seen (Ogelman 1993). While the Arons and Scharlemann (1979) model does not have this problem, since the space charge in the starvation zone above the polar cap is made up almost entirely of the outbound beam, in star centered dipole geometry it dramatically fails to account for pulsar emission over most of the P − Ṗ diagram and predicts radio polarization variations in contradiction to the observations (Narayan and Vivekanand 1982).