Yellow. Green. Blue. Dr. Vogel’s oxygen-lines 5603 5189 4829 Dr. Watts’s carbon-tube bands or lines 5602 5195 4834 Now these wave-length differences are so small that they raise a presumption of the possibility of the spectra being identical. On the other hand, assuming the spectra are not identical, the comparison tells the other way, viz. that the differences are so minute as to escape detection in instruments of moderate dispersion. With my own instrument I found the O spectrum too faint to increase the dispersive power with advantage. Considering the extremely different character of the two discharges, the identity of all the O tubes, and the presence of the line found between γ and α in the O spectrum, I think the two spectra are independent, though I admit there is room for doubt. [Sidenote: O and CO₂ spectra photographed.] _Note._—Since this examination I have photographed both spectra side by side (see ‘Photographed Spectra,’ Plate XXXI., text, pp. 69, 70). The pictures include, of course, only the blue and violet parts of the spectrum; but they are widely different in aspect, and show that, photographically at least, in this part of the spectrum there is a complete want of identity. Subsequent investigations, however, by Schuster and others (detailed later in this Chapter), go to establish that the principal lines shown in mine and Dr. Vogel’s tubes were due to (probably hydrocarbon) impurity. The exception is the single line common to mine and Dr. Vogel’s tubes, but absent from the coal-gas spectrum. This line proves to be oxygen. Compare oxygen-tube spectra (Plate XIV. spectra 3 and 4) with Schuster’s oxygen-tube spectrum (Plate XVIII. fig. 15). The line in question is found identical in the three tubes. The tube OH₂ was found to give the principal lines of the O and H spectra combined on a faint continuous spectrum. _Geissler Mercury-tube_ (Plate X. fig. 7) _and Barometer Mercurial vacuum_. [Sidenote: Mercury- and barometer-tubes examined. Mercury-tube described. Barometer-tube.]