1904] On the Group IV Lines of Silicinni. 297 



tube, the top that from the capillary, both photographed during the same 

 exposure. The transverse white line represents the junction of the 

 spectra of the bulb and txipillary. In the top portion the silicium lines 

 of Groups II, III, and 1Y,' are well seen, and also the ordinary lines of 

 oxygen, these being indicated in the reproduction. The triplet 

 marked is one of the" most conspicuous features of the spark 

 spectrum of oxygen, and the individual lines are, according to Neovius, 

 the strongest in the oxygen spectrum. It will be seen that the Si IV 

 line 4089'! is stronger than any of the oxygen triplet lines, while the 

 line 4116-1 is about as strong as the latter. As the oxygen and nitrogen 

 lines of Neovius which occur near these lines are of intensity (1) and 

 (< I) respectively, it is evident that if the lines ascribed to Si IV in the 

 Kensington spectra are really due to air, as M. de Gramont contends, 

 they must undergo a remarkable transformation as regards intensity in 

 passing from the conditions of the ordinary spark spectrum to those of 

 the vacuum tube. There is, however, no indication of the selective 

 enhancement of these faint air lines in the Kensington spectra of air 

 under the vacuum tube conditions. 



It will be instructive to trace the behaviour of these different sets of 

 lines in the two portions of the spectrum shown in Plate 11. It will be 

 noticed that in the bottom or bulb portion of the spectrum the oxygen 

 triplet survives, though it is very weak. The Si IV lines, however, 

 one of which is stronger in the top spectrum than the oxygen lines, 

 have entirely disappeared. As to the silicium lines of Groups II and III, 

 they are represented in the bulb spectrum ; both sets, however, being 

 weaker than in the capillary spectrum, the weakening in intensity 

 being more marked in the case of Group III than Group II. It is thus 

 seen that the conditions appertaining to the gases in the bulb are 

 conducive to the existence of the lines of Group II ; less so to those of 

 Group III ; and not at all to those of Group IV. 



In B of Plate 1 1 the Si F 4 spectrum is reproduced alongside that of 

 Orionis. The identity of position of the oxygen triplet and Si IV 

 lines with lines in the stellar spectrum is there denoted. It may here 

 be stated that the stellar and terrestrial spectra do not exactly fit 

 throughout their whole length, owing to the fact that they were 

 obtained with different prisms. 



In A of Plate 12 the spectrum shown in the top part of A Plate 11 is 

 compared with the ordinary spark spectrum of air. The lines 4089 '1 

 and 4116-4 of Si IV, which are strongly marked in the vacuum tube 

 spectrum, are entirely lacking in the air spectrum, although the oxygen 

 triplet previously referred to is common to both spectra and of about 

 equal intensity in each. 



In B of Plate 12 the spark spectrum of sodium silico-fluoride, 

 volatilised between platinum poles, is compared with the spectrum 

 of air, also made incandescent between platinum poles. In each 



