Mechanics of Luminosity. 249 



the determination of the brightness of the comparison-lamp 

 in absolute measure was the following: — At a definite distance 

 from the slit in the prolongation of the axis of the collimator 

 of a Glands photometer a platinum wire was placed, with its 

 axis of length parallel to the slit, which was then ignited by 

 means of a galvanic current. The source of light was placed 

 before the comparison-prism. 



The platinum wire d must be placed in a vacuum, since 

 otherwise its loss of energy would depend not only — nor even 

 chiefly — upon the radiation, but very largely also upon con- 

 vection-currents. Since the wire expands when heated, it 

 must not be clamped tight or it will become curved, and 

 would thus not remain parallel to the slit. It was therefore 

 fastened in the following manner (PI. III. fig. 1). Its upper 

 end was soldered to a strong brass wire m, at its lower 

 end a long thick copper rod k was fastened, provided below 

 with a point p. Above and below, the outer portions of two 

 ground pieces a and a' were melted on to a cylindrical glass 

 tube c of 38 millim. diameter. A brass wire m was cemented 

 into the inner portion s belonging to the upper ground portion. 

 The inner portion s f of the lower ground portion contained 

 mercury q, which was put into connexion with the source of 

 electricity by means of the copper wire r cemented in, into 

 which the point p of the copper rod k plunged. When the 

 platinum w T ire was heated by means of a current, the copper 

 wire was stretched by the copper rod so that it remained 

 perfectly straight. A side tube h connected the cylindrical 

 tube with Topler's mercury pump with SundelFs modifica- 

 tions and improvements*. It was exhausted as completely 

 as possible. In a discharge-tube melted on at the same time 

 and provided with a plate and a point 6 centim. from each 

 other, discharges were scarcely to be obtained, using an 

 induction-coil of moderate size. In fact the further action 

 of the pump did not diminish the brightness of the wire. 

 The heating of the glass wall by radiation was without in- 

 fluence, since, as far as the measurements of brightness and 

 the electrical measurements w r ere concerned, it was a matter 

 of indifference whether the wire w r as maintained at the high 

 temperature a longer or a shorter time. 



The double-image prism in the photometer was so placed 

 that the squares of the cotangents of the angles read were 

 proportional to the brightness J of the platinum wire. The 

 apparatus furnished to me differed from that described by 

 Herr Grlan in having the Wollaston's prism turned through 

 an angle of 180°. 



* Beibl. ix. p. 756 (1885). 



