We present breakdown voltage measurement data in air taken with a technique we have recently developed and present calculations that show that the common assumption that one protrusion is responsible for the observed current is not warranted. The data suggest that below 10μm of electrode separation the dominant effect is field emission of electrons. Analyzing the data in that regime, using the theory of standard field emission and field amplification on the surface of a conductor, we come to a prediction about the geometry of the surface protrusions on the electrodes. We look for these protrusions using an atomic force microscope. We find several reasons why the standard theory does not appear to explain our data and conclude that the standard theory is not sufficient.