The experimental part of this research involved measurements of the brightness of a hollow cathode discharge in the light of various wavelengths of the first spark spectrum of aluminum. The hollow cathode tube was constructed in accord with modern specifications for a light source for high resolution spectroscopy. The line brightnesses were measured under various conditions of operation of the discharge tube. The external parameters which were varied were discharge current, cathode temperature, carrier gas pressure and cathode diameter;The motivation for the experiment was two-fold. Firstly it was desired to obtain empirical information on the optimum conditions for the operation of this type of source for the production of spectra, and, secondly, date were sought which might help to clarify the atomic excitation processes. Both of these objectives have been attained to some extent;In general it has been found that the brightness of the source is roughly proportional to the square of the well current density. There is an optimum pressure for operation of the discharge which is dependent on the cathode diameter and, to some extent, on the cathode temperature and on the particular atomic lines to be excited;The average line brightness at constant current varied approximately inversely with the square of the cathode diameter. From the quadratic dependence of brightness on well current density noted above, this implies that cathode diameter has little or no effect on the brilliance of the source. The total light output from the discharge should then vary directly with the square of the cathode diameter, a result which is not inconsistent with the observations of Winger37 of the spectrum of Pb II in a hot hollow cathode;It has been found possible to explain in a rough way the features of the brightness variations by making a number of approximations concerning the physical processes within the discharge, and some rather crude estimates of the magnitudes of various important but unknown constants of the atomic interactions. In addition a mechanism has been proposed for the preferential excitation of certain excited states of the aluminum ion, which lead to the peaks observed in the brightnesses of several series of atomic lines in its spectrum;In addition the observations of Sawyer38 on the "limit of excitation" of the Al II spectrum in the type of source have been confirmed.