108 PROCEEDINGS OF SECTlOiV A. 



Scliuster, with the object of repeating and extending the work of 

 Humphreys and Mohler, which then only concerned pressures as high 

 as 14:^^ atmospheres. In 1906 the writer showed to the Britisli 

 Association! photographs of the iron arc under pressures of +3, +4, 

 + 5, +10. +15, +20, +25, +30, +40, +50, +60, +80, and 

 4-100 atrns., and presented in 1907 a detailed account of the work.| 

 At the same time (1907) Humphreys§ published the displacements of 

 lines of iron and various metals under pressures of 42, 69, and 101 

 atmospheres. Photographs have now been obtained by the writer of 

 the spectrum of the copper,]] silver, iron, gold, and metal arcs up to 

 200 atmospheres, and alloys have also been investigated up to the 

 same pressure. The apparatus consists of a drawn steel cylinder 

 (designed by Dr. J. E. Petavel, F.R.S.), 2 ft. long, 3 in. internal, 

 5 in. external diameter. Heavy covers are bolted to the top and 

 bottom, and these are furnished with insulated stuffing-boxes through 

 which pass steel rods, to whose ends can be clamped the electrodes 

 whose spectrum is required. These were connected with the terminals 

 of tlie corporation mains at 100 A'olts, and an arc formed between 

 them o])posite a glass windoAv in the side of the cylinder. Tlie light 

 was examined by means of a large Rowland grating spectograph 

 (21^ ft. radius), which, in the second order, gave a dispersion of 

 1 nim. = 13 A. Us. 



Iron presents a spectrum wliose lines exhibit most of the pheno- 

 mena associated with pressure changes. The effects first noticeable 

 as the pressure is increased are the broadening of the lines and their 

 tendency to reverse, but careful examination also shows that they are 

 displaced from their original positions by an amount which is greater 

 as the pressure is increased. To compare the original and displaced 

 lines it is usual to employ a comparison shutter, which allows the 

 central strip of the plate to be exposed to the normal arc, and then 

 covers that part of the plate anci exposes the rest of it to the spectrum 

 given by the arc under pressure. The two are then in close juxta- 

 position, and the displacements are easily seen, and are measureable. 



Careful measurements of the displacements of the iron lines show 

 that three groiips exist with displacements in the ratio of 4:2:1. 

 The unsymmetrically reversed lines are anomalously displaced, the 

 reversal being displaced half as much as the non-reversed part. 

 From this we learn that the outer envelope of an arc does not neces- 

 sarily absorb the most intense vibration emitted by the central core. 

 Forty thousand measurements of the displacement of sixty lines were 

 necessary for the determination of the relationship of the displace- 

 ment to the pressure up to 100 atjnospheres. The plates taken 

 between 100 and 250 atmospheres are now being measured by an 

 assistant at the Manchester University, and, as far as has at present 

 been ascertained, there is no discontinuity throughout the whole 

 range of pressure. The relation between the pressure and the dis- 

 placement is in general a linear one, though some anomalous readings 



t Brit. Assocn. Report, York, p. 481, 1906. 



J Roy. Soc. Proc, 1907; Phil. Tr.ins. Roy. Soc, 208, iii., 1908. 



§ A.strophysioal Journal, 26, IS, 1907. 



!! Roy. Soc. Proc, A., 81, 378, 1908. 



