10 



It should be stated, however, that this method of measuring 

 light is not adapted to an accurate measurem.ent of the total light 

 intensity, since the decomposition of the solution is not propor- 

 tionate from hour to hour; but as the results are relative, this 

 method answered our purpose fairly well. According to photo- 

 graphic light exposure tables, the light in June is from three to 



FIG. 1— SHOWING CLOCK ARRANGED FOR EXPOSING TUBES. 



a. Showing clock and tube in position, covered. 



b. Showing the lube exposed. 



The release of the tube covering is accomplished by using the mechanism 

 known as the alarm. 



five times as bright as in December or January; besides, the days 

 are match longer. Our records, therefore, do not give the exact 

 light relationships existing from month to month. A variation in 

 the decomposition of the solution would, however, tend to make 

 the difference in our percentages greater rather than less, since 

 the chemical solutions exposed to the stronger light would deterior- 

 ate more quickly and lose their sensitiveness sooner than the solu- 

 tions which had decomposed less. 



