388 Prof. W. G. Duffield : A Comparison of the 



Table II. 



Class. 



Lr s . 



II. r v , 



III. s ... 



IV. g b 3 



V.gb r 



Spark (Column A) 

 agreements with 



B. ! C. 



18* 

 1 



Spark (Column A) 

 disagreements with 



2* 

 1* 

 1 

 1 



* The figures in Classes I. and II. are subject to alteration if the argument 

 in par. 5 is conceded. 



The disagreements relate only to lines not included in the 

 agreement columns. 



3. Accepting the classification given by Bilham, we note 

 from Tables I. and II.: — 



(a) The spark displacements per atmosphere of lines 



reversed both symmetrically and un symmetrically 

 under 10 atmospheres pressure are smaller than 

 those observed in the arc under the same pressure. 

 For 14 lines the ratio of spark to arc displace- 

 ments is 1*9 to 4*0. 



(b) The spark displacements per atmosphere of lines 



which are unreversed, whether slightly broadened 

 or greatly broadened towards the red under 10 

 atmospheres pressure, do not differ appreciably 

 from those observed in the arc under the same 

 pressure. 



For 6 slightly broadened lines the ratio of spark 



to arc displacements is 4 # 8 to 5*0. 

 For 7 lines greatly broadened to the red, the 



ratio of spark to arc displacements is 9*7 to 



11-6. 



4. It was found in the case of the nickel arc that the 

 displacements per atmosphere under 10 atmospheres pressure 

 were higher than those measured when the arc was subjected 

 to higher pressures (see columns B and 0, Table I.), — the 

 ratio being approximately 1*8 to 1*0 for all lines compared, 

 but smaller for those of large wave-lengths and unsymmetrical 

 broadening. 



The displacements of reversed linps in the spark under a 

 pressure of 10 atmospheres are in good agreement with the 

 arc readings derived from higher pressures. 



