In a series of previous papers, the detection of arcing on GPS, Van Allen Probe, and GEO satellites with the Arecibo Observatory Wm. E. Gordon 305 m radiotelescope has been reported and elaborated. In this paper, characteristics of the radiofrequency spectra of two true GEO satellites are reported as they were measured with the Arecibo Mock spectrometer every 9.6 microseconds, and are contrasted with spectra of impulsive RFI (RadioFrequency Interference) events. It is shown that the arc spectra in the 300–350 MHz range are not continuous, but consist of a number of narrow spectral lines. These lines are consistent within the duration of single arcs, but change between arcs, so that when averaged over many arcs, the appearance of a continuous spectrum is approximated. This behavior is replicated in the laboratory within charging-induced vacuum arcing on small solar array samples. In contrast, the spectrum of impulsive RFI (including, but not limited to lightning strokes) is essentially random from one 9.6 microsecond sampling interval to the next. Since the spectrum of a true arc changes little throughout many samples, it cannot be due to the risetime of the arc current, which is much less than one sampling interval. Furthermore, in laboratory experiments done at very high time resolution (< 1 ns), there is little correlation between the arc current spectra and the arcing RFI spectra, indicating that even arc current fluctuations within each arc can have little to do with the UHF spectra observed.