1886.] 



Radiation of Liglit and Heat. 



209 



This filament was then mounted on platinum electrodes, as the 

 other had been, enclosed in a similar glass globe, and exhausted of 

 air, the vacuum being carried to about the xooVoo" °^ an atmosphere. 



The remeasurement of these carbon filaments subsequent to >the 

 redaction of the carbon on their surfaces, showed no perceptible 

 increase in their diameter, the deposit of carbon which had been 

 added being in all likelihood less than the ten thousandth of an inch 

 in thickness, and the surface areas of each filament still remained 

 practically equal in all respects. 



Having now two carbon filaments with which a comparative test 

 might be made, and in which the conditions were in all respects 

 identical, except in that of surface condition or polish, the one being 

 like soot and the other like silver, I passed a series of known electrical 

 currents through each in turn, registering the light produced against 

 a standard candle burning 120 grains per hour in a good photometer 

 provided with a sliding screen. 



In Table No. I, Carbon A, are shown the testings of the blackened 

 filament, and in Table II, Carbon B, those of the filament which was 

 made bright. In diagram No. 1 may be seen the plotting of these 

 results and their relative curves. The dotted curve marked Carbon A 

 shows the testings of the black filament, and the curve marked with a 

 plain hard white line, Carbon B, gives the testings of the filament 

 which was bright. The horizontal divisions in the diagram give the 

 watts or volt amperes of energy passing through the filament, and 

 the perpendiculars mark the corresponding candle powers. From 

 these tests it may be noticed that with two carbon filaments identical 

 in all respects but in that of surface polish or brightness, the 

 blackened filament required no less than 100 watts to keep its surface 

 at an incandescence yielding 20 candles, whilst the filament with the 

 bright surface was kept at the same incandescence, and gave an equal 

 light with 74 watts only, also that each filament when consuming an 

 energy of 4 watts per candle, that which was blackened required no 

 less than 113 watts of energy to effect this (besides having its surface 

 incandescence strained to yield 28 candles), while the bright filament 

 with 71 watts only effected the same economy, viz., 4 watts only per 

 candle, and had to give from its surface only 17f candles. 



These results satisfied me that the condition of the carbon surface 

 was wholly the cause of the large differences shown by these curves, 

 and I determined therefore to carry out a more extended series of 

 tests with carbons about which I knew nothing. 



For this purpose, therefore, I procured two carbons of foreign 

 manufacture, but by whom made I did not know. The following 

 were their chief characteristics. The carbons were very nearly 

 square in section, and appeared before carbonising to have been 

 sliced from some homogeneous material like parchment paper. 



