which, through geographical position and isolation, are the liker refuges for wind-borne migrants. Also that the accidental departure from the directions aimed at by the birds is, where wind and barometric systems are so Variable, far more frequent than is usually suspected. Direct routes are doubtless aimed at, but only accomplished under favourable conditions for the whole journey; migration is less infallible than we have been led to think. It is, too, an evolving habit, strengthened by those which survive its perils, now as it was in its early days.
During a long overland journey, winds will probably have less influence, though for rapid passages high flights certainly appear to be not uncommon. There is, however, another aspect of the connection between migration and weather which we have hardly touched, migration synchronal to the change of season. Mr Cooke shows that. in North America the push forward in spring is not in most species so soon as the weather permits; they do not actually move on the spring wave. Many warblers which nest in the Great Slave Lake region in an average temperature of 47°, linger in the Tropics, and reach New Orleans when the temperature is about 65° F. Then they hasten northwards, outstripping the advancing spring, finding in Minnesota a temperature of about 55°, and 52° in Manitoba, and gain another 5° on the season by the time they reach their home.
