If it can be assumed that nocturnally migrating birds are approximately uniformly spaced across the sky and that the red lights did not attract birds which would otherwise have missed the tower, it is possible to compute the volume of migration from the sample killed. In regard to the first assumption, both Stone (1906:250-251) and Lowery (1951:409-413) have presented evidence of fairly uniform distribution of nocturnal migrants. We have no information on the second assumption beyond the facts that birds do not strike the high towers on clear nights or lower towers even on stormy nights.

On nights when large numbers of birds struck the 950 foot Topeka tower, only a few struck a 500 foot radio tower, also lighted with red lights, at Lawrence, 24 miles east, under similar weather conditions. Most of the birds found at Topeka were fairly close to the base of the tower, indicating that they struck the tower itself or that they were flying high enough to strike guy wires only fairly close to the tower. The scarcity of birds under the guy wires some distance from the tower at Topeka and at the radio tower at Lawrence causes us to think that most of the birds were flying more than 450 feet above the ground. On this basis, we have computed numbers of migrants passing through a plane one mile long and 500 feet high (2,640,000 square feet), intersecting the assumed path of migration at right angles. Vertically, the theoretical plane begins at 450 feet above ground and has its top edge at 950 feet. The solid (discounting spaces between girders, etc.) cross-sectional area of the tower intersecting this plane was computed by actual measurement to be 1685 square feet. On the night of September 30-October 1, 585 birds were killed. By computation (585/1685 = X/2,640,000), approximately 916,000 birds passed through the mile-long plane that night. On each of the nights of October 5-6 and October 6-7, approximately 230,000 birds passed through this plane. By comparison, Lowery (1951:436) recorded maximum station densities in one night in spring of 63,600 birds at Tampico, Mexico, and 51,600 at Lawrence, Kansas, as determined by moon-watching. Lowery's figures refer to numbers of birds crossing any part of a circle one mile in diameter and are roughly comparable to ours if, as we think, most of the birds at Topeka were flying at altitudes between 450 and 950 feet above the ground.

It must be realized that these figures are only approximations. One variable ignored is the frontal extent (or area, viewed from the front, subject to damage by striking an obstruction) of the birds themselves. Since practically all birds killed showed head or trunk injuries, rather than a high proportion with only broken wings, we chose to disregard frontal extent of the birds in our calculations. If our figures are inaccurate by as much as 50 per cent in either direction, which seems unlikely to us, they still give some idea of the tremendous volume of nocturnal migration under some conditions.

It may be more meaningful to compute numbers of migrants by species. This can be done readily by making appropriate substitutions from Table 1 in the equation given above. For example, on the night of September 30-October 1, approximately 147,000 Nashville Warblers passed through the mile-long plane and on the same night, 100,000 Mourning Warblers and 14,000 Philadelphia Vireos. Neither of the last two species would be judged to be abundant migrants in autumn in eastern Kansas by ordinary field observations; the television tower sample, however, indicates that these as well as other species must often be overlooked when they do stop in Kansas.