Ignition, of Gases by Hot Wires. 617 



the mixture being 30 per cent. of gas in air. The observa- 

 tions are represented by straight lines which cut the hori- 

 zontal axis at points between '03 and *05 millimetre diameter. 

 Ignition is abnormally difficult with thick silver wires. 

 The current i = a + bd, where a is negative and b increases 

 with the conductivity o£ the wire. The coefficient a is the 

 current which would give heat to the wire equivalent to that 

 caused by surface combustion in wires of negligible diameter. 



The coefficient b= j-j, the slope of the lines, is closely pro- 

 portional to the conductivity as shown in Table I., so that 

 i=a + kad, where k has a mean value of 1"67 . 10 -3 in 

 hydrogen, a is also proportional to er, having a mean 

 value 1*16 . 10 ~ 3 . k does not vary more than the wires 

 may do from their mean value, though hydrogen is 

 undoubtedly ignited by platinum more readily than by 

 other metals. 



Table I. 

 Hydrogen, 30 per cent, in air. 



Metal. Conductivity ft= U _ k= ^ 



a - 3d a 



Silver 60-10 4 1037 173. 10" 3 



Gold ... ' 41-2 812 1-93 



Molybdenum 25"0 345 1-38 



Tungsten 20-0 307 P53 



Platinum ..' 91 125 138 



Nickel 8-3 187 220 



Iron 110 166 1-51 



Mean 1*67 



From an extended series of measurements, the results 

 of which are collected in Tables II. and III., this type of 

 relation holds in all gases. 



Table II. 

 Platinum Wires (from three diameters). 



n Per cent. Calorific , 



^ as - in Air. Value. °' 



Hydrogen 30 58 I.e. 125 



Carbon Monoxide 33 68 177 



Methyl Alcohol 130 182 202 



Ethyl Alcohol 35 340 208 



Ethylene 7'0 333 197 



Ethyl Ether 3T> $60 208 



Benzene 4'5 799 205 



Pentane 3'5 850 204 



