In the summer of 1971 we conducted new measurements of the primary electron spectrum from 1 to 20 Gev with our electron spectrometer telescope at Fort Churchill. This telescope has been calibrated by using electrons and pions of 0.5–15 Gev at Slac (Stanford Linear Accelerator) in early 1971 and again in 1972. The differential primary electron spectrum over the energy range 3–20 Gev is given bydj/dE= 100/E3±0.2el/m² ster sec Gev, the integral spectrum byJ(E)= 50/E2±0.2. These intensities are considerably lower than many earlier measurements above ∼5 Gev. It is observed that many showers have anomalous lateral distributions as determined in our telescope and must be rejected. This directly leads to the lower intensities. Our calibration results suggest that these events are due to interactions of high‐energy protons. Our measured primary electron spectrum is uncomfortably close to that calculated for secondary electrons produced by the interaction of cosmic ray nuclei with interstellar hydrogen in the galaxy, and therefore the question of the origin of the high‐energy electrons is reopened. The problem of the age of cosmic ray electrons must also be reexamined if our measured spectrum is found to extend to still hig