396 



T. T. Beinnister and M. J. Vrooman 



D = (a-2)/p2 (10a) 



1-1? l-2a 

 p /p = (2b-l)( a ) (11a) 



1 2 



Comparison of models . From experimental values for maxlmxim 

 enhancement and for the ratio P-i/p„ needed to obtain maximiun 

 enhancement, the values of a ana b can be calculated from the 

 equations 9 and 9a and 11 and 11a, respectively. From our exper- 

 imental values E^j '^'2.5 and p,/pp'^'5, a = 0.20 and b = O.65 for 

 the spill-over model, and a = 6.29 and b = 0.57 for the separate 

 package hypothesis. 



When these values for a and b axe inserted into equations 7 

 through 10, or into equations 7a through 10a, one obtains, for 

 short-wave light limiting, 



E = 140.30 p^/p^ (12) 



D = 0.30 pj^ (13) 



and, for short-wave light greater than balancing, 



E = 2.5 (1^) 



D = 1.5 P (15) 



Thus, both models, fitted to the same experimental data, lead to 

 identical n\Americal equations for E and D as functions of illu- 

 mination. 



COMPARISON OF DERIVED AND EXPERIMENTAL FUNCTIONS 



The predicted functions E and D, given for both models by 

 equations (12) through (15), are shown in Fig. 3- Comparison 

 with the experimental functions (Figs. 1 and 2) shows that, for 

 smeill p and p , the predicted and experimental surfaces are in 

 good agi-eementf both with respect to the general shape of the 

 contours as well as to the positions of contours of a given 

 value. Despite these points of similarity, there are, however, 

 two important discrepancies. 



