PART II. THE NATURE OF NOCTURNAL MIGRATION 408

Horizontal Distribution of Birds on Narrow Fronts 409

Density as a Function of the Hour of the Night 413

Migration in Relation to Topography 424

Geographical Factors and the Continental Density Pattern 432

Migration and Meteorological Conditions 453

CONCLUSIONS 469

LITERATURE CITED 470

LIST OF FIGURES

Figure Page

1. The field of observation as it appears to the observer 374

2. Determination of diameter of cone at any point 375

3. Temporal change in size of the field of observation 376

4. Migration at Ottumwa, Iowa 377

5. Geographic variation in size of cone of observation 378

6. The problem of sampling migrating birds 380

7. The sampling effect of a square 381

8. Rectangular samples of square areas 382

9. The effect of vertical components in bird flight 383

10. The interceptory potential of slanting lines 384

11. Theoretical possibilities of vertical distribution 388

12. Facsimile of form used to record data in the field 391

13. The identification of co-ordinates 392

14. The apparent pathways of birds seen in one hour 393

15. Standard form for plotting the apparent paths of flight 395

16. Standard sectors for designating flight trends 398

17. The meaning of symbols used in the direction formula 399

18. Form used to compute zenith distance and azimuth of the moon 400

19. Plotting sector boundaries on diagrammatic plots 402

20. Form to compute sector densities 403

21. Determination of the angle [alpha] 404

22. Facsimile of form summarizing sector densities 405

23. Determination of net trend density 406

24. Nightly station density curve at Progreso, Yucatán 407

25. Positions of the cone of observation at Tampico, Tamps 411

26. Average hourly station densities in spring of 1948 414

27. Hourly station densities plotted as a percentage of peak 415

28. Incidence of maximum peak at the various hours of the night in 1948 416

29. Various types of density-time curves 418

30. Density-time curves on various nights at Baton Rouge 422

31. Directional components in the flight at Tampico, Tamps 428

32. Hourly station density curve at Tampico, Tamps 429

33. The nightly net trend of migrations at three stations in 1948 431

34. Stations at which telescopic observations were made in 1948 437

35. Positions of the cone of observation at Progreso, Yucatán 443

36. Hourly station density curve at Progreso, Yucatán 444

37. Sector density representation on two nights at Rosedale, Miss. 451

38. Over-all sector vectors at major stations in spring of 1948 455

39. Over-all net trend of flight directions shown in Figure 38 456

40. Comparison of flight trends and surface weather conditions on April 22-23, 1948 460

41. Winds aloft at 10:00 P. M. on April 22 (CST) 461

42. Comparison of flight trends and surface weather conditions on April 23-24, 1948 462

43. Winds aloft at 10:00 P. M. on April 23 (CST) 463

44. Comparison of flight trends and surface weather conditions on April 24-25, 1948 464

45. Winds aloft at 10:00 P. M. on April 24 (CST) 465

46. Comparison of flight trends and surface weather conditions on May 21-22, 1948 466

47. Winds aloft at 10:00 P. M. on May 21 (CST) 467

INTRODUCTION

The nocturnal migration of birds is a phenomenon that long has intrigued zoologists the world over. Yet, despite this universal interest, most of the fundamental aspects of the problem remain shrouded in uncertainty and conjecture.

Bird migration for the most part, whether it be by day or by night, is an unseen movement. That night migrations occur at all is a conclusion derived from evidence that is more often circumstantial than it is direct. During one day in the field we may discover hundreds of transients, whereas, on the succeeding day, in the same situation, we may find few or none of the same species present. On cloudy nights we hear the call notes of birds, presumably passing overhead in the seasonal direction of migration. And on stormy nights birds strike lighthouses, towers, and other tall obstructions. Facts such as these are indisputable evidences that migration is taking place, but they provide little basis for evaluating the flights in terms of magnitude or direction.

Many of the resulting uncertainties surrounding the nocturnal migration of birds have a quantitative aspect; their resolution hinges on how many birds do one thing and how many do another. If we knew, for instance, how many birds are usually flying between 2 and 3 A. M. and how this number compares with other one-hour intervals in the night, we would be in a position to judge to what extent night flight is sustained from dusk to dawn. If we could measure the number of birds passing selected points of observation, we could find out whether such migration in general proceeds more or less uniformly on a broad front or whether it follows certain favored channels or flyways. This in turn might give us a clearer insight into the nature of the orienting mechanism and the extent to which it depends on visual clues. And, if we had some valid way of estimating the number of birds on the wing under varying weather conditions, we might be able to understand better the nature and development of migration waves so familiar to field ornithologists. These are just random examples suggesting some of the results that may be achieved in a broad field of inquiry that is still virtually untouched--the quantitative study of migratory flights.

This paper is a venture into that field. It seeks to evaluate on a more factual basis the traditional ideas regarding these and similar problems, that have been developed largely from circumstantial criteria. It is primarily, therefore, a study of comparative quantities or volumes of migration--or what may be conveniently called flight densities, if this term be understood to mean simply the number of birds passing through a given space in a given interval of time.

In the present study, the basic data permitting the numerical expression of such migration rates from many localities under many different sets of circumstances were obtained by a simple method. When a small telescope, mounted on a tripod, is focused on the moon, the birds that pass before the moon's disc may be seen and counted, and their apparent pathways recorded in terms of coördinates. In bare outline, this approach to the problem is by no means new. Ornithologists and astronomers alike have recorded the numbers of birds seen against the moon in stated periods of time (Scott, 1881a and 1881b; Chapman, 1888; Libby, 1889; West, 1896; Very, 1897; Winkenwerder, 1902a and 1902b; Stebbins, 1906; Carpenter, 1906). Unfortunately, as interesting as these observations are, they furnish almost no basis for important generalizations. Most of them lack entirely the standardization of method and the continuity that would make meaningful comparisons possible. Of all these men, Winkenwerder appears to have been the only one to follow up an initial one or two nights of observation with anything approaching an organized program, capable of leading to broad conclusions. And even he was content merely to reproduce most of his original data without correlation or comment and without making clear whether he fully grasped the technical difficulties that must be overcome in order to estimate the important flight direction factor accurately.

The present study was begun in 1945, and early results obtained were used briefly in a paper dealing with the trans-Gulf migration of birds (Lowery, 1946). Since that time the volume of field data, as well as the methods by which they can be analyzed, has been greatly expanded. In the spring of 1948, through the cooperation and collaboration of a large number of ornithologists and astronomers, the work was placed on a continent-wide basis. At more than thirty stations (Figure 34, page 437) on the North American continent, from Yucatán to Ontario, and from California to South Carolina, observers trained telescopes simultaneously on the moon and counted the birds they saw passing before its disc.

Most of the stations were in operation for several nights in the full moon periods of March, April, and May, keeping the moon under constant watch from twilight to dawn when conditions permitted. They have provided counts representing more than one thousand hours of observation, at many places in an area of more than a million square miles. But, as impressive as the figures on the record sheets are, they, like the published observations referred to above, have dubious meaning as they stand. Were we to compare them directly, station for station, or hour for hour, we would be almost certain to fall into serious errors. The reasons for this are not simple, and the measures that must be taken to obtain true comparisons are even less so. When I first presented this problem to my colleague, Professor William A. Rense, of the Department of Physics and Astronomy at Louisiana State University, I was told that mathematical means exist for reducing the data and for ascertaining the desired facts. Rense's scholarly insight into the mathematics of the problem resulted in his derivation of formulae that have enabled me to analyze on a comparable basis data obtained from different stations on the same night, and from the same station at different hours and on different nights. Astronomical and technical aspects of the problem are covered by Rense in his paper (1946), but the underlying principles are discussed at somewhat greater length in this paper.