Behaviour of Metallic Vapours in Flames. 25 



atmospheres, and kept burning there for an hour or more. 

 During experiments the cock in the cover must, of course, 

 be partly open and a constant supply of fresh air maintained. 



The appearance of the flame changes considerably with 

 rise of pressure. The flame-piece used in these experiments 

 was of quartz with a rectangular opening 11 mm. by 2 mm.; 

 the hollow space of the flame (usually called the cone, though 

 in this case it was nearly rectangular in shape) was always 

 adjusted to the same height, about 1 cm. As the pressure is 

 raised the outer edge of the flame becomes brighter than the 

 rest of the flame, taking on the appearance of the bright blue 

 line which defines the cone. At an excess pressure of about 

 two atmospheres this outer line is very plainly marked, and 

 the space bounded by it has somewhat shrunk together 

 above, while outside the blue line the flame is seen in two 

 horns. At higher pressures this appearance becomes more 

 definite ; the outer blue line defines a more or less pointed 

 space, and the two horns of pale blue flame become very 

 definite : the flame has thus three points above. The inner 

 part of the flame above the cone appears faintly reddish at 

 high pressures (3-5 atmospheres) and gives a continuous 

 spectrum. If a round flame-piece be used the same pheno- 

 mena are observed, modified by the shape of the flame. 



When the electric field is applied, the flame loses this 

 distinctive appearance ; it appears to be to some extent 

 depressed from above and broadens out : the different parts 

 so clearly seen before become merged. The flame presents 

 a uniform appearance, without projections, and though on 

 the whole a little more deflected towards the negative than 

 the positive electrode, it seems, by the broadening, to be 

 deflected in both directions. 



On bringing a platinum wire bearing a little salt into the 

 flame, it was observed that with a fixed amount of salt on the 

 wire the streak of coloured vapour appeared somewhat fainter 

 at high pressures. It seemed possible that the same amount 

 of salt vaporizing at high pressure might not radiate so 

 intensely as at atmospheric. The other alternative is that 

 the salt was not vaporizing so freely at high pressure. To 

 test this point measurements were made of the amount of 

 salt vaporized for a given luminosity at various pressures. 



§ 7. Measurements of the Luminosity of the Streak, and the 

 Weight of Salt vaporized. 



The metal used was strontium, and to measure the bright- 

 ness of the red streak of strontium vapour an optical 



