82 BROOMALL : 



Dividing the flowers into three color groups and testing the 

 alcoholic solutions made from the petals of these three classes 

 of flowers we find the following reactions : — 



( 1 ) Red and pink flowers — Green with alkali, 

 red with acid ; 



(2) Yellow flowers — Green with alkali, yel- 

 lowish with acid ; 



(3) Purple and blue flowers — Green with 

 alkali, purple or blue with acid. 



These reactions the writer has found to occur in solutions 

 made from many kinds of flowers, both wild and cultivated, 

 and the color changes are strong and delicate and capable of 

 reversal back and forth a number of times. As regards Class 2, 

 however, care must be taken not to add an excess of reagent. 



While these reactions of course may not be true for all 

 flowers, 5'et in no case did they fail to appear in the writer's 

 extended tests. However that may be, for the sake of the 

 interesting speculations that result, let us assume them gener- 

 ally true and see where we land. 



In the first place we see that in all cases the color is green 

 with alkali, while with acid the solution takes more or less 

 the color of the mature flower as we ordinarily see it. Bearing 

 in mind, however, the fact that purple is not a simple color, 

 but a combination of red and blue, the reactions lead us to 

 suspect that we are dealing with but three pigments, which 

 are red, yellow and blue in the acid condition respectively. 

 These, however, are the three primary pigments from which, 

 by various combinations, all colors may be produced. Hence, 

 granting the existence of one or more of these three pigments 

 in the flower petals, we see how it is possible to account for 

 all the varied hues of the flowers around us. So much for 

 the physical side of the subject. 



But a still more interesting thought results from a consid- 

 eration of the physiological relations of the pigments. Thus, 

 we know that all ])lants lix'c, as it were, a dual existence. 



