before the middle of July, and fifty days for the trip will bring the earliest migrants to the Gulf States in September. Yet both old birds and young of the year have been seen by the middle of July at Key West, Fla., 500 miles south of the breeding range, on August 10 in Costa Rica, and on August 5 on the northern coast of South America. These dates prove conclusively that these early migrants south of the United States could not have been birds from the northern part of the range, but must have been those of the southern part.
Black-throated blue warblers reach Cuba in the fall at just about the time that other migrants of the species appear in North Carolina. The inference is that the arrivals in Cuba are the birds that nested in the southern Alleghenies, while those appearing in North Carolina are from the latitude of northern New England or beyond. Redstarts and summer warblers arrive on the northern coast of South America so early (August 27 to September 2) as to prove that they represent the southern breeding birds. Indeed, these representatives of the species are seen in South America at just about the time the earliest of the northern breeding birds reach Florida.
Recent investigations have also shown that many species of birds do not follow this "normal" order of migration. The most southern-bred Maryland yellow-throats are almost nonmigratory, residing throughout the year in Florida; those breeding in the middle districts migrate only a short distance, while those of Newfoundland go to the West Indies, passing directly over the winter home of their fellows in the South. The red-winged blackbirds of the middle of the range in northern Texas are almost stationary, but are joined in winter by migrant red-wings from the remote Mackenzie Valley. The palm warblers of the interior of Canada, in the course of their 3,000-mile journey from Great Slave Lake to Cuba, pass through the Gulf States early in October. After the bulk of these have passed, the palm warblers of the northeastern British provinces come slowly down to the Gulf States, and settle there for the winter, content with only a 1,500-mile trip. Some of the blackpoll warblers that pass in spring through Florida proceed northeast 1,000 miles to breed in northern New England, while others, traveling northwest more than 3,000 miles summer in Alaska. Among the Maryland yellow-throats that nest in western Pennsylvania are undoubtedly individuals that during the winter are scattered in the Gulf States, the West Indies, and even Central America. Enough examples have been given to show that no invariable rule, law, or custom exists in regard to the direction or distance of migration. The winter distribution can not be certainly determined from the summer home, nor does it positively indicate that home. Although a certain general tendency is observable, yet each species presents a separate problem, to be solved for the most part only by patient, painstaking observation and by the recognition of subspecies.
Spring migration has its own special features. No such synchronous movement occurs in the spring as has been described as "normal migration" in the fall. With many birds, possibly the majority of land birds, the first individuals of a species to appear in spring at a given locality are supposed to be old birds that nested there the previous year. The supposition is that these birds are followed by those that nested in the region just to the north; and that later, those of still more northern homes pass by; and that the last to appear will be those whose homes are in the most northern part of the breeding range. If, then, for any species, the southern nesting birds lead the van in both fall and spring migration, and the near guard in each case is composed of northern breeding birds, it follows that some time between October and April a transposal of their relative positions occurs; and that the more southern birds pass over the more northern ones, which
