given velocity if liberated at any given height above the ground. Thus, if a seed liberated 100 feet from the ground falls that distance in half a minute, and the wind is blowing at the rate of, say, 1,000 feet in half a minute (or nearly 23 miles per hour, a good breeze), the seed will be carried 1,000 feet before it reaches the ground. Its course will be represented by the diagonal AD of the accompanying figure, where AB represents the distance which the seed falls in the given time, and AC the distance according to the same scale travelled by the wind in the same period. [Illustration: FIG. 12.] But most seeds sufficiently light to be capable of extended flights are liberated only a few feet from the ground; they are dependent on upward eddies to raise them if they are to achieve more than a very short migration. That such eddies, both upward and downward, occur on a windy day we all know from experience; and it is they that make or mar the fortune of most wind-borne seeds. Only some local or accidental excess of upward over downward eddies will assist a seed on its journey; and as every upward eddy must be compensated somewhere by a downward eddy, the longer the journey is, the more such eddies tend to neutralize each other. Over the sea--that most formidable barrier to plant migration--eddies do not prevail as they do over rough ground, so that, unless by a series of lucky eddies a seed is whirled up to a considerable elevation before it leaves the shore, the chances of its successful passage across a stretch of water are remote. Discussing the possibility of seeds of Portuguese plants reaching the Azores, lying 800 miles to the westward, H. B. Guppy[4] shows, from observations on the rate of fall of seeds made by several workers, that with a 50 miles per hour horizontal wind the light-plumed seed of the Common Groundsel (_Senecio vulgaris_), for instance, would require to be liberated at a height of 9 miles above the ground if it is to reach the islands: or to express it differently, if liberated at ground-level, the seed would need to be raised 9 miles by upward eddies during its journey, even if corresponding downward eddies were absent--which they certainly never are. It is clear that if even light seeds are to achieve anything more than short journeys, they must depend on exceptional disturbances of the air, such as whirlwinds and tornadoes. It is now time to examine the devices by which many seeds achieve a more or less wide dispersal by means of the wind. Seeds possessing these adaptations may be divided into three classes: (i.) Powder seeds, (ii.) winged seeds, (iii.) plumed seeds.