40 TRANSACTIONS OF THE A5IERICAN INSTITUTE. 



of limitiBg this in No. 6, by increasing the upper surface, is note* 

 worthy. The difficulty in lighting is a serious drawback to No. 

 1, involving much care and time ) indeed so long a time that 

 it might be a serious difficulty to have no other lamp in 

 case of fire. No. 2 is only slightly more effective, Nos. 4 and 6 

 may be relighted instantly night after night. It is to be observed 

 also that while these may be blown out with comparatively little 

 smell, Nos. 1 and 2 require to be turned down to be extinguished. 

 They therefore require readjustment for lighting. The quantity 

 of heat withdrawn from the flame proportionably retards light? 

 ing. It seems to have been assumed by those inventors who use 

 pinions for setting the wick, that they would also serve for alter- 

 ing the quantity of light at will ; but it is not so, the extent of 

 exposed surface is then changed with the effect before noted, and 

 direct experiment shewed that great variation in the character 

 of a flame is produced by small changes in the height at which 

 air is deflected upon it. In the field which seems open to these 

 lamps — the desk and the secluded work table — it is of great 

 advantge that a lamp should be self-supplying — the weary eye 

 says that the light is burning, but affords no standard for adjust- 

 ment, and turns dissatisfied from even destructive glare. 



The pinions without racks afford an uncertain guide where a 

 small want of parallelism may disarrange the whole effect. 



The principle applied in Nos. 1, 2, 4 and 6, viz: drawing up a 

 heated column of air below the flame, seemed to deserve exami- 

 nation. Perfect combustion being due to a proper admixture of 

 the combustible and air, at a sufficient temperature, it is appa- 

 rent how much velocity of issue, and extended conductors of heat 

 in contact with the vapor, may impede this. 



The characteristic of a lamp with a chimney is the concentra- 

 tion of flame and heat, where the air is made to impinge. Con- 

 trast with this the dispersion of the flame upwards and heat 

 downwards. In proportion as this principle is developed, (in the 

 reverse order of the numbers,) there results a larger flame from 

 the same sized wick, at a cost of radiated heat and of oil; the 

 light I ot varying equally with the size of the flame, and com- 

 plete combustion being obtained by withdrawing heat from a 

 much smaller portion of the flame than in No, 1. Another fea- 

 ture of lamps with chimneys, as commonly made, is excess of air 

 beyond what is necessary for perfect combustion. This is always 

 ready for an irregular supply of vapor, and with this form of 



