No.423.] NORTHERN POLYPETALOUS FLOWERS. 237 
inclined to believe that it is due rather to the chemical constitu- 
tion of the soil and nutritive fluids and the action of light. The 
terminal flower receiving the greatest amount of sap would be 
the first affected, then those of the umbellets, and finally the 
entire umbel. The color changes of many flowers appear to be 
due to such causes rather than to the influence of insects. In 
Low Germany Pimpinella magna is white, and in the more 
intense light of the Alps pinkish. Both rose-colored and yellow 
flowers have been found on plants of Eriogonum ovalifolium 
growing in silver-ore localities. Chemical analysis showed that 
the rose flowers contained arsenic, which was not present in the 
yellow. As a stimulus in nutrition may intensify the color, so 
a check in growth may cause it to revert to white. 
Increased conspicuousness of the umbel is gained in some 
genera by the enlargement of the outer petals of the marginal 
flowers, as in Heracleum. White bracts subtend the white 
flowers of Astrantia, and yellowish bracts the dull-yellow flowers 
of Bupleurum, and in Eryngium the flower stalk is colored. 
The Cornaceze include many ornamental shrubs valued for 
their bright-red bark, their variegated leaves, the masses of 
handsome flowers, and the coral and blue berries. The flowers 
are chiefly white, but in the European Cornus mascula are 
yellow. The forty or more small white flowers of the herba. 
ceous bunchberry Cornus canadensis, gain conspicuousness by 
an involucre of four white bracts, sometimes tinged with red. I 
have also seen the leaves partially white. This species yields 
very little nectar, which can be detected only by close examina- 
tion. The blooming season lasts for over a month, and the 
flowers are produced in the greatest profusion. I have watched 
these plants long and carefully, and have enumerated thirty-six 
visitors ; but not once have bumblebees been seen to visit the 
blossoms, — a fact I attribute to the scarcity of honey. The 
shrubby species contain a more abundant supply of nectar. 
The cymes are large and very numerous, and not infrequently 
are sought by bumblebees. On C. alternifolia I have collected 
twenty-eight insects. There. can be no doubt that the quantity 
and flavor of nectar is an element in the limitation of insect 
visits, but one that has received much less attention than 
