Migration owes its origin to the potentiality of flight, enabling birds to advantage themselves by extended dispersals, which through heredity become instinctive, regular and periodical. Geological changes, especially the passing away of the glacial epoch, only influenced by opening up new lands for summer colonisation, but climatic conditions prevented these lands from becoming permanent abodes and fostered the habit of periodical migration. Whatever the original home or centre of distribution may have been, the dispersal from it was towards new lands with a retreat towards the food-supply when these lands became untenable. Fluctuating food-supply, love of home, sexual impulses, desire for light, varying temperature, and other factors, all have more or less influence, but the force exerted by any or all depends upon the species operated upon and the locality in which it resides. The present route followed or method of migration is little guide to the history of past migration; during the evolution of present-day migration alterations may have been occasioned by environment and changing conditions. As Seebohm puts it, "The desire to migrate is a hereditary impulse, to which the descendants of migratory birds are subject--a force almost, if not quite, as irresistible as the hereditary impulse to breed in the spring" (44). The route is simply the course followed between the breeding area and winter quarters; it is more or less restricted by the size of the area in which food is to be found; it is usually the most direct way from one food-base to the next, in a general direction from the seasonal bases. Most birds move between north and south, but migrations are regularly followed in other directions by some species. Routes may follow coast-lines, these providing visible landmarks, and also, for many species, plentiful food; islands, capes, estuaries and inlets are landmarks, asylums, food-bases, and sites for congregation and departure for cross-sea passages; at these places migration is often specially noticeable. Overland routes may suggest "broad front" migration, when there are no particular restricting influences and the species have no special need for hurry. Migration at great elevations and at high rates of speed is proved, but the highest and quickest possible is as yet unascertained. It may also, under other conditions, be performed at low elevations and very slowly. It is probable that strong air-currents at a high elevation materially assist rapid and lengthened migration. Force not direction of wind influences birds moving at a low elevation.