Harris's sparrow supplies an interesting example of a narrow migration route in the interior of the country ([fig. 9.]) This fine, large finch is known to breed only in the region from Churchill, on the west shore of Hudson Bay, northwest to the shores of Great Bear Lake. Very few actual breeding records of the species are available, but these are sufficient to indicate that the breeding range is in the strip of country characterized by more or less stunted timber just south of the limit of trees. When it begins its fall migration, this species necessarily covers the full width of its breeding area. Then it proceeds almost directly south, or slightly southeasterly, the area covered by the majority of the birds becoming gradually constricted, so that by the time it reaches the United States it is most numerous in a belt about 500 miles wide, extending across North Dakota to central Minnesota. Harris's sparrows are noted on migration with fair regularity east to the western shore of Lake Michigan, and west to the foothills of the Rocky Mountains, but the great bulk of the species moves north and south through a relatively narrow path in the central part of the continent. Present knowledge suggests that the reason for this narrow migration range is the close association that Harris's sparrow maintains with a certain type of habitat' including brushy places, thickets, edges of groves, and weed patches. While these environmental conditions are found in other parts of the country, the region crossed by this sparrow presents almost a continuous succession of habitat of this type. Its winter range extends from southeastern Nebraska and northwestern Missouri, across eastern Kansas and Oklahoma and through a narrow section of central Texas, at places hardly more than 150 miles wide.
The scarlet tanager presents another extreme case of narrowness of migration route ([fig. 10]), its breeding range extending in greatest width from New Brunswick to Saskatchewan, a distance of about 1,900 miles. As the birds move southward in fall their path of migration becomes more and more constricted, until at the time they leave the United States all are included in the 600-mile belt from eastern Texas to the Florida peninsula. Continuing to converge through Honduras and Costa Rica, the boundaries there are not more than 100 miles apart. The species winters in northwestern South America, where it spreads out over most of Colombia, Ecuador, and Peru.
The rose-breasted grosbeak also leaves the United States through the 600-mile stretch from eastern Texas to Apalachicola Bay, but thereafter the lines do not further converge, as this grosbeak enters the northern part of its winter quarters in Central America and South America through a door of about the same width ([fig. 11]).
Although the cases cited represent extremes of convergence, a narrowing of the migratory path is the rule to a greater or less degree for the majority of North American birds. The shape of the continent tends to effect this, and so the width of the migration route in the latitude of the Gulf of Mexico is usually much less than in the breeding territory.
The redstart represents a notable case of a wide migration route, although even in the southern United States this is much narrower than the breeding range ([fig. 12]). These birds, however, cross all parts of the Gulf of Mexico and pass from Florida to Cuba and Haiti by way of the Bahamas, so that here their route has a width of about 2,500 miles.
Figure 11.—Distribution and migration of the rose-breasted grosbeak. Though the width of the breeding range is about 2,500 miles, the migratory lines converge until the boundaries are only about 700 miles apart when the birds leave the United States. For migration paths of other widths see figures 9, 10, and 12.
The flyways
In 1935, as a result of studies of banding data, the author discovered the existence of the four great flyway systems. This discovery, based upon analyses of the several thousand records of the recovery of migratory waterfowl then available, was announced by the Biological Survey (Lincoln, 1935c) and, beginning in 1948, it has served as the basis for administrative action by the Fish and Wildlife Service in the annual hunting regulations.
