When mixed gases are under examination, it frequently hap pens that a particular discharge will not produce spectra of all the gases present. Thus, an uncondensed discharge through a vacuum tube containing residual air will show the bands of nitro gen but will give no indication that there is any oxygen in the tube. Both gases, however, are revealed by their line spectra when the condensed discharge is passed through the mixture. In analysing a mixture of gases it is accordingly necessary to observe the spectrum under varying conditions of electrical dis@harge, and variations of pressure are also desirable.
Interesting results have also been obtained by King with arcs in air using a current of the order of i,000 amperes at ioo volts. Similar experiments by J. W. Ryde, with currents up to 25o amperes, have been of special value in revealing with great in tensity the line spectra of neutral carbon and nitrogen, which were previously only partially known. Our knowledge of these spectra had previously been considerably extended by T. R. Merton by observations of carbon compounds, or of nitrogen, when mixed with a large proportion of helium. A further appli cation of Merton's method, though better results were obtained with argon admixture, was made by J. C. McLennan and G. M. Shrum in experiments on oxygen which led them to their important discovery that a line of neutral' oxygen observed under these conditions was identical in position with the green line which is most characteristic of the spectrum of the aurora.