288 AiNNUAL REPORT SMITHSONIAN INSTITUTION, 1919. 
way derange its physiological activities. I look upon such sugar 
glands as safety valves for the relief of excessive osmotic pressures 
that are dangerous to the internal economy of the plant. And not 
only is this conception applicable to extrafioral nectaries in general, 
but it may serve also, in the case of floral nectaries, to explain their 
origin. Having once arisen as osmotic safety valves, the usefulness 
of the floral nectaries as an aid to cross-pollination would tend 
strongly to bring about their natural selection and perpetuation. 
7. The establishment of a dormant condition hefore the advent of 
freezing loeather and the continuation of this dorm^ancy through 
warm periods in late fall and early winter are protective adapta- 
tions of vital necessity to the native trees and shrubs. 
A little consideration will show how important the principle of 
chilling is to those species of trees and shrubs which are subjected 
each year to several months of freezing weather. If they were so 
constituted as to start into growth as easily in the warm days of late 
fall as they do in the warm da^^^s of early spring, many species would 
come into flower and leaf in those warm autumn spells that we call 
Indian summer, and the stored food that the plant required for 
its normal vigorous growth in the following spring would be 
wasted in a burst of autumn growth, which would be killed by the 
first heavy freezes, and would be followed by a winter of weakness 
and probable death. But when two or three months of chilling are 
necessary before a newly dormant plant will respond to the usual 
effect of warmth, such plants are protected against the dangers of 
growth in Indian summer. It is probable that all our native trees 
and shrubs are thus protected. 
Any member of this audience may make, next fall and winter, a 
simple and instructive experiment with such early spring blooming 
plants asalder, hazelnut, pussy willow, yellow bush jasmine, forsythia, 
Japanese quince, peach, and plum. In mid-autumn bring into your 
living room and set in water freshly cut dormant leafless branches 
of these plants. They will not bloom. At intervals of a few weeks 
during late autumn and winter try the same experiment again. 
You will find that the branches cut at later dates will come into 
bloom under this treatment. They will not do so, however, until 
the expiration of the period of chilling appropriate to the various 
kinds of plants included in the experiment. The required period 
of chilling varies greatly. In the case of some of the cultivated 
shrubs about Washington, especially the yellow bush jasmine 
{JasminuTTi nudiforu^n)^ so brief a period of chilling is required 
that extraordinarily cold weather in late October or early No- 
vember may chill them sufficiently to induce them to bloom if a 
period of warm weather follows in late November. The period of 
