The sharp peaks in the photo-electron spectra have heights which are, on the whole, a monotonically increasing function of the atomic number for a given shell. If it is assumed that the number of electrons counted is proportional to the number of photons absorbed by the shell, the intensity is proportional to the number of electrons in the shell, to its average value of r–2, and to a third factor representing the square of an effective nuclear charge. As a consequence, the signals achieve intensities some five times the intensity of fluorine 1s for the highest ionization energies which can be measured with 1486 eV photons, viz., 1s of magnesium, 2p of gallium to arsenic, 3d of tin to barium, and 4ƒ of thorium and uranium. This agrees with independent measurements by Wagner. On the other hand, most valence shells and 1s of lithium and beryllium are very weak. These statements are modified by broadening of 4d signals by radiative half-life below 10–15 s in lutetium and heavier elements, and of 4p signals between rhodium and cadmium containing many 4d electrons, by the splitting of the signal in several components due to differing interelectronic repulsion in many paramagnetic compounds, and by the occurrence of satellites in copper(II)(possibly connected with the anti-bonding 3d(x2–y2) electron) and in lanthanum(III) and uranium(VI)(due to inter-atomic electron transfer to the empty 4ƒ or 5ƒ shell). The precision of repeated measurements of intensities is not better than 30 %. Differing shells may vary by a factor of 1000, the valence electrons frequently not observed at all. On the other hand at the end of the transition groups, the d and ƒ shells can produce signals half as strong as F 1s.