EFFECT OF PRESSURE UPON ARC SPECTRA. GOLD. 57 



remarkable because there is not the slightest trace of either of these lines in the 

 comparison spectrum which was obtained by a divided exposure, one-half before the 

 pressure photograph, the other half after it. 



Faint traces of aluminium (3944*29 and 3961 '85) and iron also occur on the same 

 photograph, but they are feeble compared with the strong H and K lines. Except 

 for a faint trace of the line K at 50 atmospheres (H being absent) there is no 

 evidence of calcium on any of the other photographs taken under pressure, though 

 they occur in the comparison spectrum at 1 atmosphere, in the next photograph at 

 30 atmospheres. Above 50 atmospheres there is no sign of this element in either the 

 pressure or comparison spectra ; it appears that the calcium which had by some 

 means got on to the pole at 10 atmospheres was dissipated as time went on. 



The remarkable nature of the phenomenon demanded that further photographs 

 should be taken at low pressures, and so photographs were obtained at 5 and at 15 

 atmospheres. The former contained strong evidence of the presence of calcium, the 

 H and K lines being, as on the previous photograph at 10 atmospheres, five or six 

 times as strong as the strong line f2 of gold, and, again, there was only the smallest 

 evidence of this substance in the comparison spectrum, only a faint trace of the line K 

 being discernible, though the exposure was, as always, a divided one. Upon the 

 photograph at 1 5 atmospheres there is, on the other hand, no evidence of either of 

 these lines under pressure, though there is a faint indication of both H and K in the 

 comparison spectrum. 



An explanation that seems possible to the writer is that some calcium impurity 

 either pre-existed in one of the gold rods and formed a tiny pocket of metal, or that 

 some particle of dust rich in calcium (and containing aluminium) was introduced into 

 the cylinder with the air from the gasholder, and settled upon the pole while molten. 

 In either case, we can imagine that a small quantity of calcium becomes lodged in one 

 particular spot upon the face of one of the poles. It is known that the arc wanders 

 from point to point over the surface of the poles, and when it happened to spring from 

 the spot rich in the impurities they would manifest their presence. We may suppose 

 that the locality occupied by the calcium happened to be on an irregularity on the 

 pole which infrequently formed the starting place of the arc. Under these circum- 

 stances, when the arc burned badly and required a' very prolonged exposure, as 

 it did most especially at 5 and 10 atmospheres, this spot would be more frequently 

 visited by a fluctuating arc than on those occasions when the arc burned steadily 

 and kept to one spot. 



It is worth noting that the spectrum of gold at atmospheric pressure was 

 photographed by a small 1 -metre grating spectrograph before the gold was subjected 

 to pressure, and afterwards upon the same plate after the arc had been investigated 

 at high pressures. The two photographs show not even the slightest difference. 



Another point of interest is the order of appearance of the H and K lines when 

 only small amounts of calcium are present. When only one of them occurs either 



VOL. CCXI. A. I 



