Some years ago Dr. H y 1 1 e r a a s suggested to us to measure the intensity-ratio of heliumlines, in order to check the results of a planned calculation of probabilities of transition. The comparison of theory and observation is simplest when the lines correspond to transitions from a common initial-level. In this case the conditions of excitation have no influence on the intensity-ratio. The latter is then only determined by the properties of the atom.
We have measured the ratio of the pairs of lines 23P-43D (?,= 4472 A) and 33P--43D (X = 17003) and of 21S-41P ()~ = 3965) and 31S-41p (~, --15088). There are in the heliumspectrum no pairs of lines with the same initial level, both of them lying in the visible part or in the ultraviolet. We were therefore compelled to make measurements in the infrared part of the spectrum. As in that case, the method does not allow the measuring of weak lines, we have restricted ourselves to the lines mentioned above.
The lightsource was a tube with a hot cathode and with a thinwailed capillary tube of quartz with a diameter of 1 ram. This capillary tube was surrounded by a cooling mantle through which air was blown. Liquid-cooling had to be rejected, in order to avoid absorption in the infrared. The tube was connected to a direct tension of about 2000 Volt. After a prolonged use, the wall of the capillary tube was somewhat blackened by the sputtering of the electrodes. To avoid an error in the intensity-ratio, introduced by this effect the tube was from time to time cleaned with dilute hydrofluoric acid.
The capillary tube was put in front of the entrance slit of a Wadsworth-monochromator. The cooling mantle was so narrow that