A tube such as is usually employed in examining the spectra of gases at low pressures was next filled with the gas of density 19. Such a tube, called a Pflücker’s tube, after its inventor, contains wires of platinum sealed through at each end, where it is about half an inch in width; the middle portion of the tube is about 3 inches long, and its bore is a fine capillary. When the platinum wires are connected with the secondary terminals of a Ruhmkorff’s coil, and the tube is partially exhausted, a brilliant glow appears in the capillary portion. If viewed through a glass prism, different gases show different sets of coloured lines crossing the usual gradation of colours of the spectrum. Thus hydrogen exhibits three striking lines, one bright red, one peacock blue, and one violet; nitrogen shows a large number of somewhat hazy bands, red, orange, yellow, and yellow-green in colour, besides a number of bands of a violet colour; but the new gas, while exhibiting the bands characteristic of nitrogen, showed in addition certain groups of red and green lines which did not appear to belong to the spectrum of any known gas. [Illustration: FIG. 2.] While these experiments were in progress, Lord Rayleigh was occupied in preparing nitrogen from other sources, and in determining its density; and in every case it was evident that nitrogen from all sources except the atmosphere weighed somewhat less than atmospheric nitrogen. He therefore proceeded to repeat Cavendish’s experiment, and like Cavendish, he obtained a small residue of gas which would not disappear on sparking with oxygen, in presence of caustic soda. The sparks, as they passed, could be observed through a spectroscope (which consists of an arrangement of prisms and lenses so designed as to examine the components of the light emitted by the sparks), and he, too, was struck with the unusual character of the spectrum. His experiments proved, besides, that the amount of residue was roughly proportional to the amount of air taken; thus, beginning with 50 cubic centimetres of air, the residue was 0·32 cubic centimetre; and from 5 cubic centimetres of air, only 0·06 cubic centimetre of gas was obtained. These small amounts are not proportional to the quantities of air taken; but, as will afterwards be seen, the discrepancy is owing to the solubility of the new gas in water. Still they served to show that from a comparatively large amount of air, more of the new gas could be obtained than from a smaller amount.