The return migration is, so far as it is known, in a steady northerly direction, rather north-west across Bolivia towards Central America. From Yucatan they cross the Gulf to Texas, then slowly travel up the great Mississippi highway and across Canada to their northern breeding grounds. "Its round trip has taken the form of an enormous ellipse with a minor axis of 2000 miles and a major axis stretching 8000 miles from Arctic America to Argentina." The following is Mr Cooke's suggestion of the origin of this great ellipse. Towards the close of the glacial era, when the ice began to recede, the Florida peninsula was submerged and only a small area in the south-east of the States was free from ice. Plover attempting to follow up the retreating ice were confined to an all-land route from Central America through Mexico to the western part of the Mississippi Valley. As the east gradually became uncovered the route would be extended to the north-east, until the area stretching to the Great Lakes was fit for bird-habitation. As the route lengthened and the power of flight developed, there would be a tendency to shorten the line by cutting off some of the great curve (No. 1) through Mexico and Texas, and a short flight across the Gulf (No. 2) would be gradually lengthened, until the present spring route, then also the autumn route (No. 3), was attained. As Canada opened out, the routes in spring and autumn diverged; in autumn the fruits of Labrador were an attraction, but the Chinook winds made the country east of the Rockies more suitable for spring migration; the fall route tended eastward (No. 4), the spring route remained unchanged. When the fall route had worked eastward to the Gulf of St Lawrence (No. 5), shortening took place in the same way from the great westward curve, culminating in an ocean flight, short at first (No. 6) and later extended, the total distance shortened, until the present route was attained (No. 7).