68 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 68 



of the research involved the use of a different spectroscope using 

 a rock-salt prism of 60°. In this spectroscope at 10 /a with a devia- 

 tion of 150', (instead of the 30' with the 15° prism), practically no 

 field light was observable upon the insertion of the ^ cm. thick quartz 

 plate. Now line 6 of table 16 shows that at 10 fx the deflection, which 

 (with a rock-salt prism in both cases) should be dependent only on 

 the wave-length, would be 400/100,000; whereas the field light, if 

 dependent only upon the deviation, would be perhaps 4/100,000 or 

 only I per cent of the true deflection. The deflection in the spectrum 

 was so small at this wave-length (see fig. 6, upper part) that a de- 

 flection of I per cent of it would not have been detectable. 



Two idiosyncrasies of the field light may be noticed.: (i ) Mention 

 has been made of the erro-r produced in the determination of the 

 diaphragm values supposed to be caused by diffraction with the long 

 wave-lengths and narrow slits. If this explanation were true, it 

 would be expected that the values obtained at the same deviations 

 with the field light would not show this error, since the field light is 

 principally due to the more intense short wave-length energy. This 

 was found to be so. (2) Another cause besides the natural decrease 

 of energy with the greater wave-length tends to make the field light 

 increasingly disturbing with the longer wave-lengths. For the 

 absorption in the rock-salt of the true spectrum energy gets more 

 and more effective as the wave-length increases ; whereas the field 

 light, being of short wave-lengths, passes through the prism almost 

 unhindered. 



