His book, now a botanical classic, attracted but little attention; his publisher did not even send him a copy of it, and in disgust he turned from the study of plants to that of languages. The title of the work, The Secret of Nature in the Form and Fertilization of Flowers Discovered, affords us the pleasure of knowing that he rightly estimated the importance of his observations, Sprengel clearly states that the bright hues of flowers, as is now well established, serve as signals to attract the attention of nectar-loving insects flying near by. He was led to this conclusion very fitly by the study of Myosotis, the "forget-me-not." He has not been forgotten. His name and theory were rescued from obscurity by Darwin; his book a few years ago was reprinted at Leipsic, and is now universally recognized, says H. Müller, as having "struck out a new path in botanical science."
A day's stroll through the fields and woodlands is sufficient to show that yellow and white blossoms are in Nature more common than red or blue. From an examination of 741 New England and Eastern species belonging to 48 families (see table) it appears that 164 are yellow, 283 white, 71 red, 136 blue and purple, and 87 green. Greenish flowers occur in 25 families, yellow in 29, white in 32, red in 16, purple and blue in 22.
Yellow appears to have been the first color developed, and flowers with this coloration are usually simple and regular in structure, as the buttercups and five-fingers. But why, it will be asked, should yellow have been the primitive color? The spores and spore-cases of the club mosses, and the pollen of all cone-bearing trees, and, in fact, of most plants, are yellow, and the yellow coloration of the first petals is doubtless correlated with this fact. Flowers of this tint are peculiarly attractive to yellow-banded flies, and when dull are avoided by beetles. Yellow flowers vary greatly in size, but pale yellow flowers are usually small, and bright or orange-yellow are large. Ranunculus ahortivus and R. sceleratus, which grow in wet places, are small and pale, while R. bulbosus and R. acris, the familiar buttercups of our meadows, are an inch broad. An apparent exception to the above rule is offered by the globe-flower (Trollius laxus), found in dense swamps, which has solitary, very large, pale greenish-yellow flowers. As the cultivated European and Asiatic species have bright yellow flowers, the coloring of the sepals of T. laxus, for the petals are wanting, has probably retrograded from growing in dense shade.
Yellow flowers in their natural state exhibit but little variation of color. They change most readily to white, and less often to red or blue. Under cultivation sudden variations from yellow to white have been observed. A double yellow hollyhock, according to Dar-
