224 



PEOCEEDINGS OP THE AMERICAN ACADEMY 



TABLE IV. 

 Diameter of wire = .006 in. 



In order to determine the possible accuracy of setting the photometer 

 disk, several series of readings (Table V.) were taken, in which two 

 apertures through which a gas-flame shone were balanced against each 

 other. These readings wei'e taken with apertures of different sizes, iu 

 order to test the accuracy of setting in different parts of the scale ; that 

 is, for different relative brilliancies of the lights. Series 3 and 4 

 were taken so as to bring the disk in the same position as that which 

 it occupied with the .004 in. and .006 in. platinum wire. 



TABLE V. 



It appears that the average deviation of the different readings when 

 the opposite illuminated apertures were balanced against each other 

 is in general considerably less than that of the readings with the melt- 

 ing wire. The maximum average deviation of I. P. is 0.8 per cent 

 of its total value ; the minimum value is about half as great. 



Marked differences in the illuminating power of the platinum wire 

 of the same diameter on different days were observed, contrasting 

 strongly with the uniformity of the results of successive measurements. 

 This undoubtedly arose from changes in the quality, and hence in the 

 illuminating power, of the gas, as the height of the flame was kept 

 constant. In future experiments I hope to obtain a more constant 

 standard by using a moderator lamp with aperture, or else by employ- 

 ing the pentane burner. 



A consideration of the results given in Tables III. and IV. shows 

 that the light emitted by a filament of commercial platinum at the 



