different kindt of CoaUgaa Burnen^ ^e, 217 



light was as 150 to 100 ; in other words, there was a gain of 

 50 per cent. Beyond this, there was no farther gain by the 

 farther lengthening of the flame. 



The same was found to occur with argands, and in a greater 

 proportion. Thus, with a five-holed burner, varying from half 

 an inch to five inches, the light for equal consumpt of gas was, 

 for a flame of half an inch, 100 ; one inch, 282 ; two inches, 

 560 ; three inches, 582 ; four inches, 582 ; five inches, 604 : 

 so that, by raising the flame from half an inch to three inches, 

 the light for the same quantity of gas burned was nearly six: 

 times greater. Beyond this length of flame there was little 

 farther increase. In this last trial the standard was a four- 

 inch jet, and, as before, the light and consumpt were taken 

 as 100. 



These experiments I have myself verified repeatedly, as will 

 be afterwards noticed. 



It may be considered, then, as a rule with regard to argandsy 

 that the higher the flame is made, without smoking or ac- 

 quiring a dingy hue, the illuminating power becomes greater 

 in proportion than the increase in consumpt occasioned by the 

 lengthening, and this is to be accounted for on the principle 

 already stated ; for when the flame is low, too much air is ad- 

 mitted in proportion to the escape of gas, and consequently 

 most of it is burned without being decomposed; whereas, 

 when the flame is high, the air is in less proportion ; decom- 

 position is effbcted, and the carbon is separated from the hy- 

 drogen and then burned, and gives by its combustion in this 

 way the white light, and thus adds to the illuminating power. 



It is evident, however, that the flame cannot always be kept 

 at its highest elevation ; for as light draught or shaking of the 

 lamp is apt to occasion smoke, or a sudden change in the 

 flame, by which the chimney is apt to be broken ; and this is 

 more likely to occur in open apartments as in shops. 



The different circumstances detailed in the paper alluded 

 to, regarding the size of the apertures in argands, their proxi- 

 mity to one another, the width of the glass chimney, and its 

 distance from the sides of the burner, have left Httle farther 

 to be done with regard to the best means of consuming gas itt 

 this way. These experiments relate almost exclusively to ar- 



