During the past winter I have continued my studies on the spectrum of the night sky, and the connected subject of the auroral spectrum. The present paper reports the results obtained. The spectrographs used in this work are two of nearly identical construction. In designing them the paramount consideration was to obtain the greatest possible light gathering power, all other considerations being kept subordinate to this. It was accordingly decided to use the minimum number of optical pieces—one prism, one collimating lens, one camera lens, neither of the latter to be achromatic.