Hartley — On the Constitution of Electric SjiarJcs. 365 



range between 10,000 and 15,000° C. These measurements are 

 based upon the thermal value of each spark, together with the 

 volume of the same (Proc. Roy. Soc. Edinburgh, vol. vii. p. 699, 

 1872). Sir William Thomson calculated the temperature of the 

 sun to be 16,000° C; it is therefore, comparatively speaking, not 

 much hotter than the spark. 



The Relation between the Properties of the Metals and their 

 Spectra. — Sparks which yield the most characteristic spectra may 

 be described as follows : — 



I. Sparks from metals which are good conductors : 



(a) Metals easily vapourized. 



(b) Metals difficult to vapourize. 



II. Sparks from bad conductors : 



(c) Metals easily vapourized. 



(d) Metals vapourized with difficulty. 



I. Sparks from good conductors yield spectra with long lines, 

 usually isolated, or distributed in definite groups. The lines 

 are generally strong and continuous from point to point. If the 

 metals are easily vapourized, the lines extend above and below the 

 points of metal to a distance which not infrequently is quite the 

 length of the line which occupies the distance between the points. 

 When the metals are difficult to vapourize, the extension of the 

 lines is very slight, and a number of lines are so short as not even 

 to stretch from point to point. 



II. Sparks from bad conductors show few long lines, but 

 many short ones. The lines are generally of feeble intensity, 

 and they are not extended above and below the extremities of 

 the points of metal. If they are easily volatilized, a band of 

 continuous rays is seen at the extremities, and this joins lines of 

 different wave-lengths. 



Metals which are bad conductors, and are difficult to vapourize, 

 show lines which are short and of feeble intensity. As the points 

 get heated, the lines lose their intensity unless they are cooled by 

 immersion in water. Under this latter condition these lines are 

 prolonged, but never extended, beyond the upper and lower points. 



The Appearance of Impurities in Spectra. — As a general rule it 

 is best for the negative electrode to be placed below, so that sparks 



