690 POPULAR SCIENCE MONTHLY. 



experience of florists, who always seek to obtain a white variety 

 from which to develop the desired hue. 



Red flowers are much rarer than blue, and both are seldom com- 

 mon in the same family. For instance, in the pink family red and 

 white blossoms prevail, and there are no blue shades. The pinks 

 are crimson and scarlet, often with elegant markings and a strong 

 aromatic odor. The honey is deeply concealed, and they are vis- 

 ited almost exclusively by butterflies and millers. Twenty-eight 

 species of diurnal Lepidoptera have been collected upon a single 

 variety of Saponaria. Of the eighty species of Rosacece, thirteen 

 are red and two purple, but the forty-four white flowers are very 

 generally tinged or tipped with red. The two purplish-flowered 

 species, Geum rivale and Potentilla palustris, belong to genera in 

 which yellow predominates, and this primitive color is still evident 

 in both their calyx and corolla. There are no blue or violet flowers. 

 This family exhibits a marked tendency both in stem, leaf, bud, 

 flower, and fruit to develop reddish coloration, a tendency which 

 is probably due to the chemical constitution of the sap. There 

 are no flowers in this family adapted to Lepidoptera, but they are 

 visited by a mixed company of flies, beetles, and Hymenoptera. The 

 smaller and less specialized Rosacece are yellow and white and are 

 visited by a variety of short-lipped insects. "With the increase of 

 the flower in size and conspicuousness the number of insect visitors 

 greatly increases, and the enlargement of the flower is attended by 

 red coloration. Owing to the chemical constitution of the nutri- 

 tive fluid, probably to its acidity (for when the petals of a rose are 

 treated with ammonia they become blue), there has been no oppor- 

 tunity for the development of blue coloration by insects. With the 

 enlargement of the perianth and the increased flow of sap, red tints 

 have tended to appear by process of oxidation. 



The correlation of red coloring with an increased flow of sap 

 is well illustrated by the galls of the wild-rose tree, which are often 

 " as rosy as the rosiest apple." An abnormal flow of sap is caused 

 to the part stung by the insect, and red coloration is due to the 

 action of light, for it is of no service to the plant. Again, when the 

 flowers of Cratcegus coccinea are stung by the gall-fly the different 

 organs all become bright red, and the change in coloring is accom- 

 panied by an increase in size. In some instances red colors, accord- 

 ing to Darwin, indicate greater vigor on the part of the plant, and 

 I have also observed that the dwarfing of red flowers under culti- 

 vation may cause them to revert to white. 



It was long, indeed, believed that the same species could not 

 produce yellow, red, and blue flowers. But this doctrine, to use the 

 words of Dr. Lindley, " must now be laid up in the limbo of pleasant 



