or white, yellowish varieties, indicating an approach to the xanthic series, are sometimes found in gardens. The auricula, which is originally yellow, passes to reddish-brown, to green, and to a sort of violet, but never reaches pure blue; and single petals occasionally give suggestions of both series in distinct parts of their surfaces.
Some surprise may be felt that white does not figure in De Candolle's table. This is because an absolutely white color does not seem to exist in any flower. The fact may be shown by placing some flowers supposed to be of the purest white, like the lily, the white campanula, or the wood anemone, on a leaf of clear white paper. It will be found that the white of the corolla is really washed with yellow, blue, or orange, according to what flower is taken. If the tint does not appear distinct, infusions of the corollas in alcohol will present tones unmistakably yellow or red, etc. White flowers are therefore flowers with tints appertaining to one of De Candolle's series, but albinized, as if they were etiolated. A small number of flowers begin white, and are subsequently colored under the action of light. The Cheiranthus chameleo passes from white to citron-yellow and a slightly violet-red; the Ænothera tetraptera, at first white, becomes rose and then almost red; the petals of the Indian tamarind are white the first day and yellow the second; and the corolla of the Cobea scandens comes out greenish-white and turns to violet the second day. The most remarkable plant in this respect is Hibiscus mutabilis, which Rumph calls the hourly flower, because it starts white, turns flesh-color toward noon, and becomes red at sunset.
In his recent work on Plants and their Cosmic Media, M. Costantin has some remarks concerning the precocity of various races and the tint of their flowers. Hoffmann made observations on this point for several years. He remarked, as the result of eight years' observation, that the common lilac with white flowers blooms on an average six days earlier than the normal form with purple flowers. This might be a curious anomaly with no bearing, but the more we advance in the study of Nature the more we perceive that all phenomena, even the most insignificant, deserve to be examined. Similar results have been observed in varieties of radish (Raphanus raphanistrum) and of saffron (Crocus vernus); in the former the white flowers expand on an average of sixteen days earlier than the yellow ones (twelve years of observation), and in the latter plant the difference is four days.
These changes of tint sometimes appear to depend on the temperature. Thus, the white lilac was obtained by horticulturists under the influence of a temperature of between 30° and 35° C. We can not, however, affirm that spontaneous races with white flow-
