APPENDIX I. 



SOURCES OF RADIATION. 



In previous investigations the radiator used was a zircon lamp, or a 

 platinum strip covered with iron oxide and heated to redness by means 

 of an electric current. In the foregoing work a light portable radiator 

 was desired which could be mounted upon the spectrometer arm and 

 be moved with it. For this purpose the Nernst lamp was found to be 

 the most serviceable because it is light, compact, and has no products 

 of combustion, such as CO2 and water vapor, which would contaminate 

 the room (which was small), thus endangering the prism. The 

 " glower " causes some trouble, and since the distribution of the energy 

 in the spectrum of the " heater " is more uniform, the latter was used. 

 This is a very satisfactory radiator, since it is not affected by air cur- 

 rents,, while it can be maintained at a uniform temperature by using 

 current from a battery of storage cells. As is well known, the "heater" 

 consists of a hollow cylinder of clay wound with a fine platinum wire, 

 over which is a thin coating of a more refractory clay. When the plati- 

 num wire is heated to incandescence the clay gives out the desired radi- 

 ation. By using such a heater requiring no volts on a 90- volt circuit 

 it lasted for at least three months, when the platinum wire had volatil- 

 ized sufficiently to cause it to break. This was easily repaired and cov- 

 ered with clay. The volatilized platinum condenses upon the surface 

 of the heater in small, flat, triangular and hexagonal crystals. Joly^ 

 found that on covering a strip of platinum with topaz dust and heating 

 it to redness microscopic crystals of platinum were formed on the par- 

 tially decomposed topaz, the prevailing form being the octahedron. 



The extraordinary distribution of energy from these heaters is shown 

 in fig. 125. The minima lie close to the well-known absorption bands 

 of COo and water vapor. The maxima are not so easily explained. 

 The first one, at 2.5 ju, is no doubt due to the hot platinum wire. The 

 third maximum lies close to the absorption band of quartz found^ at 

 5.3 jx. The suppression of this apparently selective radiation at 5.2 /* 

 is well illustrated in curve h, which was obtained from a " heater " cov- 



^Joly : Nature, 43, p. 541, 1891. 



^E. F. Nichols: Physical Review, 4, 1897. 



119 



