4g DR. W. GEOFFREY D0FFIELD ON THE 



between X= 4000 and X= 4600 from the poles at ordinary pressures. But the brightness 

 of the arc increases with the pressure, and it is not impossible that the intensity of 

 the poles becomes sufficient to give a continuous spectrum. If tins is the case, we 

 must have, at 20 atmospheres, poles sufficiently hot to give a continuous spectrum 

 surrounded by an absorbing layer of silver vapour, and since, as the pressure increases, 

 the widths of the bright bands increase, this hypothesis requires that the amount of 

 absorption decreases relatively to the brightness of the poles. 



There is nothing inherently impossible in these conditions, but because the arc was 

 frequently quite free from a background of hot poles (as for instance when the bright 

 streams of vapour met some distance from them), and because the vapour is itself 

 intensely luminous and must emit a spectrum of its own, it is simpler to regard 

 the band spectrum as the true emission spectrum of silver vapour under these 



conditions. 



It should also be remarked that large white-hot molten drops do not form on the 

 silver poles as they do in the case of iron, but that the silver seems to be vaporised 

 without previous melting. There is, moreover, the additional evidence that the lines 

 a and c, which are intimately connected with the formation of these bands, remain 

 emission lines throughout. 



That the silver lines really disappear and are not hidden by the continuous spectrum 

 is borne out by the fact that on one photograph at a high pressure an impurity line 

 (of lead) remains visible against the continuous background. 



The disappearance of the line spectrum seems due to the replacement of the old 

 vibrating systems by new ones, which are perhaps new and complicated atomic 

 combinations. Whether these are aggregates of silver atoms, or combinations of silver 

 and oxygen atoms, remains to be tested a most valuable research would be the 

 examination of the silver spectrum under pressure in an atmosphere of hydrogen. 



In HARTLEY'S opinion the oxide of silver does not play any part in producing the 

 bands in the spectrum of the flame, whose temperature he considers too high for its 

 formation. The writer has, however, found that the brightness of the arc is 

 dependent upon the amount of air present in the cylinder, and since WHITTAKEK* 

 has pointed out that from theoretical considerations pressure is, in some cases, a more 

 potent factor in causing combination than is elevated temperature in producing 

 dissociation (and it is to be remarked that the bands under discussion do not appear 

 below about 20 atmospheres), the question of an oxide formation is not yet decided. 



The experimental evidence that a spectrum other than a line spectrum can exist at 

 the enormously increased temperature of the arc under pressure is of considerable 

 interest in the interpretation of sunspot spectra. The discovery of the identity of the 

 flutings of titanium oxide, magnesium hydride, &c., in these spectra has led to the 

 conclusion that sunspots are regions of low temperature, and, though this may be 

 actually the case, it is not necessarily so simply because of the existence of a fluted 

 ' WHITTAKER, ' Report British Association,' Dublin, 1908. 



