io SCIENCE PROGRESS 



lamp, however, on account of its longer burning hours affords 

 some saving in the cost of carbons and a considerable saving in 

 the expense of trimming — the recarboning of the lamps — which 

 has only to be done once every 90-100 hours instead of once 

 every 15-20 hours. The question as to which lamp is the better 

 is therefore largely a matter of locality : where the lamps are 

 scattered and labour is dear the economy in upkeep in the case 

 of the enclosed lamp more than balances the extra cost of 

 energy due to its comparatively low efficiency. So powerfully 

 have these factors, especially the latter, operated in America 

 that the open lamp was almost unused and the introduction of 

 the flame lamp was delayed for a long time. 



Various attempts were made from time to time to improve 

 the efficiency of the arc by adding to the carbons compounds 

 which would volatilise into the arc and cause it to emit light in 

 addition to the light obtained from the carbons but no success 

 was attained till in 1 898-1900 Bremer succeeded in making satis- 

 factory flame carbons by using calcium and other fluorides as 

 an addition to the carbon. The original Bremer carbon was a 

 solid rod composed of a mixture of carbon, calcium fluoride 

 and other salts. Later practice has developed the flame 

 carbon as a cored carbon having a very large core compared 

 with the ordinary carbon: the actual diameter of the core is 

 from one-third to half that of the carbon. The outer shell 

 is pure carbon ; the colouring matters are added only to the 

 mixture used for coring. The composition of the colouring 

 mixture has been simplified and the core now consists simply 

 of carbon, potassium silicate and a fluoride : for the golden- 

 yellow flame arc, the one in by far the most general use, calcium 

 fluoride is used ; a white flame is given by cerium fluoride and a 

 red flame by strontium fluoride. The fluoride volatilises into the 

 arc, converting it into an intensely luminous flame and greatly 

 lowering its resistance. As a consequence of the lowered 

 resistance it is possible to work with a very much longer arc, 

 an arc in fact about ten times as long as the ordinary open arc. 

 The flame is therefore large and contributes a large percentage 

 of the total light obtained from the lamp. It is however possible 

 to obtain a second advantage from the long arc. In the open 

 type arc the shortness of the arc necessitates the use of two 

 coaxial carbons, one vertically above the other; the lower 

 carbon obstructs about 30 to 40 per cent, of the light emitted 



