The emission spectrum of SiCl, produced in a tube discharge through a continuous flow of SiC1, vapour, has three systems of doublet bands degraded towards the further ultra-violet, the 0, o bands being at M 2942, 2924, M 2436, 2424, and M 2232, 2 2 2~. They are attributed to transitions B + X, C + X and D + X, where X is a double level with electronic separation 207'9 cm?, we= 531.0 and x,w,=2.2; B is a single level 34186.0 cm:l above X with wa =698*7; C is a close doublet 41234.8 and 41247.8 cm;l above X with U , = 674.2 and x,we = 2.2 ; and D is a single level 45005.9 cm;l above X, with w X = 663. These figures are from band-head measurements only. Datta's analysis of the-B + X system is confirmed, except in his identification of the 2,o and 3,o bands and his interpretation of observed isotope band-heads. With Datta's 2, o and 3, o bands are two groups of prominent bands between h 2830 and X 2770 not fitting into any of the three systems; their emitter may be a chloride of silicon other than SiC1.
I. I N T R O D U C T I O NP A R T of the spectrum described in this paper was observed in 1913 in the glow produced by the interaction of SiCl, vapour and active nitrogen(').A The whole of it was observed in 1923 in the uncondensed induction-coil discharge in SiC1, vapour flowing continuously through a discharge tube('). Lowdispersion spectrograms of the discharge were taken with a small quartz-prism instrument (Hilger's E 6) and some of the stronger bands were also photographed in higher dispersion (about 7-5 A. /". ) with a 2-4-m. concave grating. The spectrum was seen to consist of two somewhat similar parts A and B, one on either side of the AlCl band-system that resulted from the use of aluminium electrodes. Each part consisted of bands degraded to the further ultra-violet, i.e. opposite in direction to the S i 0 bands, which were present if residual air had not been entirely displaced by the SiC1, vapour. Parts A and B were regarded as two related band-systems, each having progressions of bands with common separations of about 525 cm? and regular variations of intensity. A constant separation, amounting to about 205 cm? in A and 195 cm? in B, was also recognized, but this did not suffice for the make-up of the usual Deslandres' arrangement of the band-heads; it was, in fact, the interval between the heads of sub-bands such as had been observed in other doublet band-systems but were not understood. The observed numerical relations have now proved to be significant for the vibrational analysis, which was not apparent at that time.