178 



INACTIVATION OF COMPLEMENT BY LIGHT 



by long exposure to the light before the first sample was placed in 

 the inner tube. If there were changes in transparency during photo- 

 inactivation successive samples should vary in their efficiency, and 

 this variation should be greater than that displayed by successive 

 samples in the control experiments. Examination of Table III will 

 show that such is not the case and that therefore the transparency 

 of the complement solutions remains constant during photoinactiva- 

 tion. In no case does the probable error of a single reading exceed 



TABLE III. 



The Variation in Efficiency of Portions of Complement Successively Exposed to 

 Light in the Inner of Two Concentric Quartz Tubes for Equal Lengths of 



Time. 



that to be expected as a result of errors inherent in the method of 

 titration, and although the mean calculated probable error is slightly 

 less in the controls the difference cannot be considered significant. 



We may sum up all these preliminary experiments with the state- 

 ment that the apparent order of the reaction is the true order, because 

 the effective light (probably that having a wave-length of 2536 

 Angstrom units) is absorbed by complement in the same degree 

 during all stages of photoinactivation. With the arrangement used 



