242 



SCIENCE 



[N. S. Vol. XLIX, No. 1262 



substance for whicli /' is tlie transmitted beam 

 and I the corresponding incident beam. If 

 tbese two beams, I and I', be equalized pboto- 

 grapMcally by means of interposed rotating 

 sectors, S and <S", tben some relation exists be- 

 tween 7/7' and S'/8. 



A previous determination by H. E. Howe^ of 

 this relation made 



L -§1 

 r ~ S 



where 8 is the variable sector cutting down the 

 beam 7 till it balances the beam I'. The equal- 

 ity was established by measuring with a sector 

 photometer the transmissions of neutral smoke- 

 glass plates, the transmissions of which had 

 previously been measured on a visual spectro- 

 photometer. 



In further establishing the validity of this 

 equality the method here employed has been 

 to use likewise two beams of known relative in- 

 tensities determined geometrically as a func- 

 tion of the relative distances of their sources 

 from the slit of the spectrograph. 



Using the Hilger quartz spectrograph, size 

 C, in front of one half of a slit 3 mm. long 

 there was placed a total reflecting prism whose 

 face was illuminated through a variable sector 

 by a fixed Nernst lamp. The illumination of 



such slight variations in voltage would be 

 equivalent. 



A series of exposures were made on succes- 

 sive portions of a plate with the movable lamp 

 set at distances increasing by small increments. 

 Such a series of exrposures would then be ex- 

 pected to contain one at which the illumina- 

 tion by the movable lamp would balance that 

 by the fixed lamp as judged by equal blacken- 

 ing of the spectrum bands on the plate. 

 I The average distance for such a balance with 

 the sector at rest, as determined by making 

 settings with increments of half a centimeter, 

 was about 40 cm. It was possible to estimate 

 the true distance to tenths of a centimeter. 

 With the sector running corresponding settings 

 of the movable lamp were made for various 

 openings of the sector, the principal openings 

 being 7.3°, 10° and 19°. The following table 

 gives the results for these openings, the figures 

 being all reduced to a common denominator by 

 dividing the distance of the movable lamp from 

 the slit by the initial distance at which the two 

 beams balance with the sector at rest. The 

 figures are arranged in rows according to the 

 sector opening and in columns according to 

 the individual plates and set ups of the appa- 

 ratus. 



the other haK of the slit was accomplished by 

 another Nernst lamp arranged on a runway in 

 the axis of the collimator so that its distance 

 from the slit could be varied at will. 



The lamps were on the same 110-volt alter- 

 nating current, city circuit, and plates were 

 always made under such conditions that the 

 voltage did not vary more than two volts dur- 

 ing the course of an experiment. The lamps 

 being of the same construction it seemed rea- 

 sonable to suppose that the simultaneous varia- 

 tions in candle power of the two lamps due to 



iH. E. Kowe, Phys. Sev., VIII., 6, 1916, 674. 



The apparatus was taken down and realigned 

 several times in order to eliminate any syste- 

 matic error. Considerable care was taken to 

 have the spectrograph and runway level, and to 

 have the runway so that the Nemst filament 

 remained in line with tlie collimator axis. 

 This was accomplished in some instances by 

 halving the distance between the two points at 

 which the spectriim disapjieared in shoving the 

 lamp from side to side, and in some instances 

 by placfing the filament at one end of the run- 

 way in contact with the slit, at the other end 

 making a symmetrical shadow of the collimator 



