NO. 9 DETERMINATION OF OZONE WULF 7 



an amount of ozone approximately equal to the average atmospheric 

 quantity was present in the cell. Failure to correct exactly for all 

 differences between the two sets of holograms led to the smooth curve 

 of the points in the absence of ozone absorption being below zero, but 

 the area due to ozone appears clearly and can be measured with 

 sufficient accuracy to be of much use. Several days' determinations 

 were made with larger quantities of ozone in the path, giving areas 

 which can be measured with greater percentage accuracy. On other 

 days smaller amounts were used in order to observe how well such 

 areas could be determined. Included in figure 3 are also three points 

 lying far oft" the curve which are illustrative of unsuitable points. 

 They lie in weak atmospheric water absorption in the red, which ab- 

 sorption may vary considerably over short intervals of time, rendering 

 the points evidently unsuitable for aiding in ozone determination. 



The area under the observed curve should evidently be limited 

 between ordinates whose values are still large compared with the un- 

 certainty in placing the base line. In the blue this has been taken as 

 the value at 4750 A. while in the red it was necessary to terminate 

 the area at 6135 A because of the uncertainty in the point at 6335 A. 

 This limitation was caused by instrumental circumstances which can 

 be altered in the future to include a greater area. 



These areas were determined for 10 independent sets of observa- 

 tions, the amount of ozone in the cell being known in each case from 

 the analytical work carried on at the time of the measurements. From 

 these results, shown graphically in Figure 4, a value sufficiently accu- 

 rate to be useful can be had for the amount of area per unit path 

 length of pure ozone at o"C and one atmosphere pressure, the 

 common meteorological form of expressing atmospheric ozone. The 

 least-squares solution, assuming the ozone concentration values essen- 

 tially free from error compared to values for the areas, which method 

 automatically weights the individual values in proportion to the area, 

 leads to the result 21.5 sq. cm. of area per mm. of ozone at standard 

 conditions of temperature and pressure when a plot is made to the 

 scale ordinates 0.001 per cm., abscissa 200 A per cm. This area can 

 thus be stated as 4.30 A independent of the scale to which it is plotted. 

 Plotting the data and planimetering the area is a procedure which has 

 the decided advantage of giving a visual record of the amount of 

 ozone which can be judged approximately at a glance. From the 

 results of Colange's data on the absorption coefficient of ozone one 

 can compute this same area, and it may be estimated directly from his 

 published curve of the absorption coefficient. One finds thereby that 

 the value obtained in the present work is about 4 per cent higher than 



