The possibility should be borne in mind, however, that there may be more than one potential topographic feature for birds to follow at some stations. Moreover, it is conceivable that certain species might follow one feature that would lead them in the direction of their ultimate goal, whereas other species, wishing to go in an entirely different direction, might follow another feature that would lead them toward their respective destination. It would seem unlikely, however, that the species composition of the nocturnal flights would change materially from night to night, although there is a strong likelihood that it might do so from week to week and certainly from month to month.

By ama.s.sing such data as records of flight direction along the same coast from points where the local slant of the sh.o.r.eline is materially different, and comparisons of the volume of migration at night along specialized routes favored during the day with the flight densities at progressive distances from the critical terrain feature involved, we shall eventually be able to decide definitely the role topography plays in bird migration. We cannot say on the basis of the present ambiguous evidence that it is not a factor in determining which way birds fly, but, if I had to hazard a guess one way or the other, I would be inclined to discount the likelihood of its proving a major factor.

D. GEOGRAPHICAL FACTORS AND THE CONTINENTAL DENSITY PATTERN

A study of the total nightly or seasonal densities at the various stations brings forth some extremely interesting factors, many of which, however, cannot be fully interpreted at this time. A complete picture of the magnitude of migration at a given station cannot be obtained from the number of birds that pa.s.s the station on only a few nights in one spring. Many years of study may be required before hard and fast principles are justifiable. Nevertheless, certain salient features stand out in the continental density pattern in the spring of 1948. (The general results are summarized in Tables 2-5; the location of the stations is shown in Figure 34.) These features will be discussed now on a geographical basis.

TABLE 2.--Extent of Observations and Seasonal Station Densities at Major Stations in 1948

========================================================================== |Nights of observation| Hours of observation| OBSERVATION STATION |---------------------+---------------------|Season |March|April|May|Total|March|April|May|Total|density ---------------------+-----+-----+---+-----+-----+-----+---+-----+-------- CANADA | | | | | | | | | Pt. Pelee | | | 1 | 1 | | | 6 | 6 | 2,500 | | | | | | | | | MEXICO | | | | | | | | | S. L. P.: Ebano | 1 | | | 1 | 3 | | | 3 | 1,300 Tamps.: Tampico | 3 | 3 | | 6 | 20 | 20 | | 40 | 140,300 Yuc.: Progreso | | 3 | | 3 | | 18 | | 18 | 60,500 | | | | | | | | | UNITED STATES | | | | | | | | | Fla.: Pensacola | | 2 | 2 | 4 | | 8 | 7 | 15 | 1,500 Winter Park | | 5 | 6 | 11 | | 39 |38 | 77 | 21,700 Ga.: Athens | | 2 | | 2 | | 10 | | 10 | 4,000 Thomasville | | 1 | 1 | 2 | | 8 | 8 | 16 | 4,700 Iowa: Ottumwa | | 5 | 5 | 10 | | 16 |28 | 44 | 134,400 Kans.: Lawrence | 2 | 1 | | 3 | 16 | 4 | | 20 | 68,700 Ky.: Louisville | | 3 | 2 | 5 | | 20 |14 | 34 | 49,300 Murray | | 2 | | 2 | | 13 | | 13 | 26,200 La.: Baton Rouge | | 3 | | 3 | | 15 | | 15 | 11,000 Lafayette | | 1 | | 1 | | 5 | | 5 | 2,800 Mansfield | 1 | 5 | 4 | 10 | 2 | 16 |22 | 40 | 22,400 New Orleans | 1 | 1 | | 2 | 5 | 2 | | 7 | 1,900 Oak Grove | | 2 | 2 | 4 | | 16 |15 | 31 | 33,900 Mich.: Albion | | 1 | | 1 | | 3 | | 3 | 1,100 Minn.: Hopkins | | | 1 | 1 | | | 4 | 4 | 2,000 Miss.: Rosedale | | 1 | 1 | 2 | | 6 | 8 | 14 | 12,600 Mo.: Columbia | | 2 | 1 | 3 | | 8 | 6 | 14 | 13,100 Liberty | | 1 | 1 | 2 | | 7 | 7 | 14 | 4,800 Okla.: Stillwater | 1 | 2 | 1 | 4 | 5 | 11 | 3 | 19 | 8,400 S. Car.: Charleston| 1 | 1 | 1 | 3 | 5 | 8 | 9 | 22 | 3,000 Tenn.: Knoxville | | 2 | 2 | 4 | | 18 |14 | 32 | 35,400 Memphis | 2 | 3 | 2 | 7 | 13 | 20 |12 | 45 | 29,700 Tex.: College | | 3 | 1 | 4 | | 19 | 8 | 27 | 32,200 Station Rockport | | 1 | | 1 | | 4 | | 4 | 6,200 --------------------------------------------------------------------------

TABLE 3.--Average Hourly Station Densities in 1948

======================================================== OBSERVATION STATION | March | April | May | Season ------------------------+-------+-------+-------+------- CANADA | | | | Pt. Pelee | | | 400 | 400 | | | | MEXICO | | | | S. L. P.: Ebano | 400 | | | 400 Tamps.: Tampico | 700 | 6,300 | | 3,500 Yuc.: Progreso | | 2,800 | | 2,800 | | | | UNITED STATES | | | | Fla.: Pensacola | | 0+| 200 | 100 Winter Park | | 300 | 200 | 300 Ga.: Athens | | 400 | | 400 Thomasville | | 500 | 100 | 300 Iowa: Ottumwa | | 1,700 | 3,800 | 3,100 Kans.: Lawrence | 4,000 | 1,400 | | 3,400 Ky.: Louisville | | 2,000 | 700 | 1,500 Murray | | 2,000 | | 2,000 La.: Baton Rouge | | 700 | | 700 Lafayette | | 600 | | 600 Mansfield | 0 | 700 | 800 | 600 New Orleans | 60 | 800 | | 300 Oak Grove | | 1,400 | 800 | 1,100 Mich.: Albion | | 400 | | 400 Minn.: Hopkins | | | 500 | 500 Miss.: Rosedale | | 1,100 | 700 | 900 Mo.: Columbia | | 400 | 1,700 | 900 Liberty | | 500 | 200 | 300 Okla.: Stillwater | 500 | 200 | 1,000 | 400 S. Car.: Charleston | 200 | 200 | 0+| 100 Tenn.: Knoxville | | 1,300 | 800 | 1,100 Memphis | 300 | 800 | 900 | 700 Tex.: College Station | | 1,100 | 1,500 | 1,200 Rockport | | 1,600 | | 1,600 --------------------------------------------------------

TABLE 4.--Maximum Hourly Station Densities in 1948

====================================================== OBSERVATION STATION | March | April | May ------------------------+---------+---------+--------- CANADA | | | Pt. Pelee | | | 1,400 | | | MEXICO | | | S. L. P.: Ebano | 600 | | Tamps.: Tampico | 3,100 | 21,200 | Yuc.: Progreso | | 11,900 | | | | UNITED STATES | | | Fla.: Pensacola | | 100 | 700 Winter Park | | 2,300 | 1,000 Ga.: Athens | | 900 | Thomasville | | 1,500 | 200 Iowa: Ottumwa | | 3,800 | 12,500 Kans.: Lawrence | 14,500 | 2,200 | Ky.: Louisville | | 5,000 | 1,400 Murray | | 3,700 | La.: Baton Rouge | | 3,400 | Lafayette | | 1,800 | Mansfield | | 2,100 | 1,600 New Orleans | 200 | 1,100 | Oak Grove | | 2,700 | 2,500 Mich.: Albion | | 700 | Minn.: Hopkins | | | 1,100 Miss.: Rosedale | | 2,200 | 1,400 Mo.: Columbia | | 800 | 3,400 Liberty | | 800 | 800 Okla.: Stillwater | 900 | 700 | 1,400 S. Car.: Charleston | 400 | 600 | 200 Tenn.: Knoxville | | 5,800 | 1,900 Memphis | 1,200 | 3,400 | 2,100 Tex.: College Station | | 3,400 | 3,100 Rockport | | 2,400 | ------------------------------------------------------

TABLE 5.--Maximum Nightly Densities at Stations with More Than One Night of Observation

====================================================== OBSERVATION STATION | March | April | May ------------------------+---------+---------+--------- | | | MEXICO | | | Tamps.: Tampico | 5,500 | 63,600 | Yuc.: Progreso | | 31,600 | | | | UNITED STATES | | | Fla.: Winter Park | | 6,200 | Ga.: Athens | | 2,600 | Thomasville | | 3,900 | Iowa: Ottumwa | | 15,300 | 54,600 Kans.: Lawrence | 51,600 | 5,400 | Ky.: Louisville | | 17,000 | 8,400 Murray | | 16,400 | La.: Baton Rouge | | 6,200 | Mansfield | | 4,900 | 5,200 Oak Grove | | 13,600 | 5,800 Miss.: Rosedale | | 6,800 | 5,800 Mo.: Columbia | | 1,400 | 10,300 Okla.: Stillwater | 2,700 | 1,900 | 3,000 Tenn.: Knoxville | | 15,200 | 9,000 Memphis | 3,600 | 7,900 | 7,000 Tex.: College Station | | 6,200 | 13,200 ------------------------------------------------------

[Ill.u.s.tration: FIG. 34. Stations at which telescopic observations were made in 1948.]

_Gulf Migration: A Review of the Problem_

In view of the controversy in recent years pertaining to migration routes in the region of the Gulf of Mexico (Williams, 1945 and 1947; Lowery, 1945 and 1946), the bearing of the new data on the problem is of especial interest. While recent investigations have lent further support to many of the ideas expressed in my previous papers on the subject, they have suggested alternative explanations in the case of others. In the three years that have elapsed since my last paper dealing with Gulf migration, some confusion seems to have arisen regarding the concepts therein set forth. Therefore, I shall briefly re-state them.

It was my opinion that evidence then available proved conclusively that birds traverse the Gulf frequently and intentionally; that the same evidence suggested trans-Gulf flights of sufficient magnitude to come within the meaning of migration; that great numbers of birds move overland around the eastern and western edges of the Gulf; that it was too early to say whether the coastal or trans-Gulf route was the more important, but that enough birds cross the water from Yucatan to account for transient migration in the extreme lower Mississippi Valley; and, that, in fair weather, most trans-Gulf migrants continue on inland for some distance before coming to land, creating an area of "hiatus" that is usually devoid of transient species. I tried to make it emphatically clear that I realized that many birds come into Texas from Mexico overland, that I did not think the hordes of migrants normally seen on the Texas coast in spring were by any means all trans-Gulf migrants. I stated (1946: 206): "Proving that birds migrate in numbers across the Gulf does not prove that others do not make the journey by the coastal routes. But that is exactly the point. No one has ever pretended that it does." Although some ornithologists seem to have gained the impression that I endorse only the trans-Gulf route, this is far from the truth. I have long held that the migrations overland through eastern Mexico and southern Texas on one hand, and the over-water flights on the other, are each part of the broad movement of transients northward into the United States. There are three avenues of approach by which birds making up the tremendous concentrations on the Texas coast may have reached there: by a continental pathway from a wintering ground in eastern and southern Mexico; by the over-water route from Yucatan and points to the southward; and, finally, by an overland route from Central America via the western edge of the Gulf. As a result of Louisiana State University's four-year study of the avifauna in eastern Mexico, I know that migrants reach Texas from the first source. As a consequence of my studies in Yucatan of nocturnal flight densities and their directional trends, I strongly believe that migrants reach Texas from this second source. As for the third source, I have never expressed an opinion. I am not prepared to do so now, for the reason that today, as three years ago, there is no dependable evidence on which to base a judgment one way or another.

TABLE 6.--Computed Hourly Densities at Tampico, Tamps., in Spring of 1948

========================================================================= | Average hour of observation DATE |-----+------+-------+-------+------+------+------+------+---- | 8:30| 9:30 | 10:30 | 11:30 |12:30 | 1:30 | 2:30 | 3:30 |4:30 -----------|-----+------+-------+-------+------+------+------+------+---- 22-23 March| 600| 700 | 1,000 | 800 | 100 | 100 | 0 | 100 | ..

23-24 March| 0| 400 | 1,200 | 3,100 | 800 | .. | .. | .. | ..

24-25 March| 300| 700 | 800 | 1,600 |1,100 | .. | .. | .. | ..

21-22 April|1,100|7,000 |14,900 |12,900 |8,100 |3,800 |3,500 | 200 | ..

22-23 April| 700|2,900 | 7,500 | .. | .. | .. | .. | .. | ..

23-24 April| 600|4,700 |19,100 |21,200 |5,500 |5,900 |4,000 |2,000 |200 -------------------------------------------------------------------------

_Western Gulf Area_

Among the present flight density data bearing on the above issues, are the six sets of observations from the vicinity of Tampico, Tamaulipas, already referred to. These were secured in the spring of 1948 by a telescope set up on the Gulf beach just north of the Miramar pavilion and only a hundred feet from the surf (see Figure 25, _ante_). The beach here is approximately 400 feet wide and is backed by scrub-covered dunes, which rapidly give way toward the west to a rather dense growth of low shrubs and trees. One might have expected that station densities at Tampico in March would be rather high.

Actually, though they are the second highest recorded for the month, they are not impressive and afford a striking contrast with the record flights there in April (Table 6). Unfortunately, only a few stations were operating in March and thus adequate comparisons are impossible; but the indications are that, in March, migration activity on the western edges of the Gulf is slight. It fails even to approach the volume that may be observed elsewhere at the same time, as for example, in eastern Kansas where, however, the migration is not necessarily correlated with the migration in the lower Gulf area.

Strangely enough, on the night of March 22-23, at Tampico, approximately 85 per cent of the birds were flying from north of an east-west line to south of it, opposite to the normal trend of spring migration. This phenomenon, inexplicable in the present instance, will be discussed below. On the other two nights in March, the directional trend at Tampico was northward with few or no aberrant components.

Observations made approximately thirty-five miles inland from the Gulf, at Ebano, San Luis Potosi, on the night of March 25-26, show lower station densities than the poorest night at Tampico, but since they cover only a three-hour watch, they reveal little or nothing concerning the breadth of the so-called coastal flyway.

April flight densities at Tampico are the highest recorded in the course of this study. The maximum hourly density of 21,200 birds is 46 per cent higher than the maximum hourly density anywhere else. The average hourly density of 6,300 in April is more than twice as great as the next highest average for that month. These figures would seem to satisfy certain hypotheses regarding a coastwise flight of birds around the western edge of the Gulf. Other aspects of the observations made at that time do not satisfy these hypotheses. Texas ornithologists have found that in periods of heavy spring migration, great numbers of birds are invariably precipitated by rainy weather.

On April 23, in the midst of the record-breaking telescopic studies at Tampico, Mr. Robert J. Newman made a daytime census immediately following four hours of rain. He made an intensive search of a small area of brush and low growth back of the beach for traces of North American migrants. In his best hour, only thirteen individual birds out of seventy-five seen were of species that do not breed there. The transient species were the Ruby-throated Hummingbird (1), Scissor-tailed Flycatcher (1), Western Wood Pewee (1), Black-throated Green Warbler (2) Orchard Oriole (7), and Baltimore Oriole (1), all of which winter extensively in southern Mexico. Perhaps, however, the apparent scarcity of transients on this occasion is not surprising in the light of the a.n.a.lysis of flight density in terms of bird density on the ground which I shall develop beyond. My only point here is to demonstrate that rain along the coast does not always produce birds.

As large as the nocturnal flights at Tampico have so far proved to be, they are not commensurate with the idea that nearly all birds follow a narrow coastwise route around the Gulf. To establish the latter idea, one must be prepared to show that the migrant species returning to the United States pa.s.s along two flyways a few miles wide in the immense volume necessary to account for their later abundance on a 1500-mile front extending across eastern North America. One might expect at least ten to twenty fold the number observable at any point in the interior of the United States. In actuality, the highest nightly density of 63,600 birds at Tampico is barely sufficient to account for the highest nightly density of 54,600 at Ottumwa, Iowa, alone.

Of course, there is no way of knowing how closely a ratio of anywhere from ten to one through twenty to one, employed in this comparison, expresses the true situation. It may be too high. It could be too low, particularly considering that preliminary studies of flight density in Florida indicate that the western sh.o.r.es of the Gulf of Mexico must carry the major part of the traffic if migratory flights back to the United States in spring take place only along coastwise routes.

Consideration of the data obtained in Florida in 1948 will serve to emphasize the point.

_Eastern Gulf Area_

At Winter Park, Florida, seventy-seven hours were spent at the telescope in April and May. This was 71 per cent more hours of actual observation than at the next highest station. Nevertheless, the total seasonal density amounted to only 21,700 birds. The average hourly density was only 300 birds, with the maximum for any one hour being 2,300 birds. In contrast, forty-five hours of observation at Tampico, Tamaulipas, in March and April, yielded a total station density of 140,300 birds. At the latter place, on the night of April 23-24, almost as many birds pa.s.sed _in a single hour_ as pa.s.sed Winter Park in all of its seventy-seven hours of observation.

Should future telescopic studies at Florida stations fail to produce densities appreciably higher than did Winter Park in 1948, the currently-held ideas that the Florida Peninsula is a major flyway will be seriously shaken. But one consideration must be kept in mind regarding the present picture. No observations were made at Winter Park in March, when it is conceivable that densities may have been materially higher. We know, for instance, that many of the early migrants to the southern United States are species whose winter homes are in the West Indies. Numbers of Vireonidae and Parulidae (notably the genera _Vireo_, _Parula_, _Protonotaria_, _Mniotilta_, _Seiurus_, _Geothlypis_, _Setophaga_, and certain _Dendroica_ and _Vermivora_) winter extensively in this region and are among the first birds to return to the southern states in the spring. Many of them often reach Louisiana and other states on the Gulf coastal plain by mid-March. In the same connection, it may be mentioned that many of the outstanding instances of birds striking lighthouses in southern Florida occurred in March and early April (Howell, 1932).

_Yucatan Area_

I have long felt that the answers to many of the questions which beset us in our study of Gulf migration are to be found on the open waters of the Gulf of Mexico itself or on the northern tip of the Yucatan Peninsula. Accordingly, in the spring of 1945 I crossed the Gulf by slow freighter for the purpose of determining how many and what kinds of birds might be seen between the mouth of the Mississippi River and the Yucatan Peninsula in fair weather, when it could not be argued that the birds had been blown there by inclement weather. To my own observations I was able to add those of other ornithologists who likewise had been aboard ship in the Gulf.

The summary of results proved that birds of many species cross the Gulf and do so frequently. It failed to demonstrate beyond all doubt that they do so in large numbers. Nor had I expected it to do so. The consensus of Gulf coast ornithologists seemed to be that transient migration in their respective regions is often performed at too high an elevation to be detected unless the birds are forced to earth by bad weather. I saw no reason to antic.i.p.ate that the results would be otherwise over the waters of the Gulf of Mexico.

The application of the telescopic method held promise of supplying definite data on the numbers of trans-Gulf migrants, however high their flight levels. The roll and vibration of the ship had prevented me in 1945 from making telescopic observations at sea. Since no immediate solution to the technical difficulties involved presented itself, I undertook to reach one of the small cays in Alacran Reef, lying seventy-five miles north of Yucatan and in line with the coast of southern Louisiana. Because of transportation difficulties, my plans to place a telescopic station in this strategic location failed.

Consequently, I returned in 1948 by freighter to Progreso, Yucatan, where telescopic counts were made for three nights, one of which was rendered almost valueless by the cloud cover.

[Ill.u.s.tration: FIG. 35. Positions of the cone of observation at Progreso, Yucatan, on the night of April 23-24, 1948, from 8:53 P. M. to 3:53 A. M. Essential features of this map are drawn to scale. The telescope was set up on the end of a one-mile long wharf that extends northward from the sh.o.r.e over the waters of the Gulf of Mexico. The triangular (white) lines represent the projections of the cone of visibility on the earth at the mid-point of each hour of observation. Only briefly, in the first two hours, did the cone lie even in part over the adjacent mainland. Hence, nearly all of the birds seen in the course of the night had actually left the land behind.]

The observation station at Progreso was situated on the northern end of the new wharf which projects northward from the beach to a point one mile over the Gulf. As will be seen from Figure 35, the entire cone of observation lay at nearly all times over the intervening water between the telescope on the end of the wharf and the beach. Therefore, nearly all of the birds seen were actually observed leaving the coast and pa.s.sing out over the open waters of the Gulf. The hourly station densities are shown in Table 7 and Figures 24 and 36. In the seventeen hours of observation on the nights of April 23-24 and April 24-25, a total computed density of 59,200 birds pa.s.sed within one-half mile of each side of Progreso. This is the third highest density recorded in the course of this study. The maximum for one hour was a computed density of 11,900 birds. This is the fourth highest hourly density recorded in 1948.

[Ill.u.s.tration: FIG. 36. Hourly station density curve for night of April 23-24, 1948, at Progreso, Yucatan.]

TABLE 7.--Computed Hourly Densities at Progreso, Yuc., in Spring of 1948

===========+============================================================ | Average hour of observation DATE +-----+------+------+-------+------+------+------+-----+----- |8:30 | 9:30 |10:30 | 11:30 |12:30 | 1:30 | 2:30 |3:30 |4:30 -----------+-----+------+------+-------+------+------+------+-----+----- 23-24 April| 400 |3,000 |5,100 |10,000 |9,000 |2,800 | 900 | 400 |....

24-25 April| 0 | 500 |3,700 |11,900 |7,900 |1,900 |1,100 | 400 | 200 -----------+-----+------+------+-------+------+------+------+-----+-----

It is not my contention that this many birds leave the northern coast of Yucatan every night in spring. Indeed, further studies may show negligible flight densities on some nights and even greater densities on others. As a matter of fact several hours of observation on the night of April 25-26, at Merida, Yucatan, approximately twenty-five miles inland from Progreso, indicated that on this night the density overhead was notably low, a condition possibly accounted for by a north wind of 10 mph blowing at 2,000 feet. I merely submit that on the nights of April 23-24 and 24-25, birds were leaving the coast of Yucatan _at Progreso_ at the rate indicated. But, as I have emphasized in this paper and elsewhere (1946: 205-206), the northern part of the Yucatan Peninsula is notably unmarked by streams or any other physiographic features which birds might follow. The uniformity of the topography for many miles on either side of Progreso, if not indeed for the entire breadth of the Peninsula, makes it probable that Progreso is not a particularly favored spot for observing migration, and that it is not the only point along the northern coast of Yucatan where high flight densities can be recorded. This probability must be considered when comparisons are made between Progreso densities and those at Tampico. The argument could be advanced that the present densities from Tampico do not sufficiently exceed those at Progreso to establish the coastal route as the main avenue of traffic in spring, since there is every reason to suspect topography of exerting some influence to produce a channeling effect in eastern Mexico. Here the coast parallels the directional trend of the migratory movement for more than 600 miles. Likewise the Sierra Madre Oriental of eastern Mexico, situated approximately 100 miles inland (sometimes less), lies roughly parallel to the coast. Because of the slant of the Mexican land ma.s.s, many winter residents in southern Mexico, by short northward movements, would sooner or later filter into the coastal plain. Once birds are shunted into this lowland area, it would seem unlikely that they would again ascend to the top of the Sierra Madre to the west. In this way the great north-south cordillera of mountains may act as a western barrier to the horizontal dispersion of transients bound for eastern North America. Similarly, the Gulf itself may serve as an eastern barrier; for, as long as migrants may progress northward in the seasonal direction of migration and still remain over land, I believe they would do so.

To put the matter in a slightly different way, the idea of a very narrow flight lane is inherent in the idea of coastwise migration.

For, as soon as we begin to visualize flights of great volume over fronts extending back more than fifty miles from the sh.o.r.e line, we are approaching, if indeed we have not already pa.s.sed, the point where the phenomenon is no longer coastwise in essence, but merely overland (as indeed my own unprocessed, telescopic data for 1949 indicate may be the case). In actuality, those who have reported on the migration along the western edge of the Gulf of Mexico have never estimated the width of the main flight at more than fifty miles and have intimated that under some circ.u.mstances it may be as narrow as two miles. No evidence of such restrictions can be discerned in the case of the trans-Gulf flights. If it cannot be said that they may be a.s.sumed to be as wide as the Gulf itself, they at least have the potential breadth of the whole 260-mile northern coast of the Yucatan Peninsula.

On these premises, to be merely equal in total magnitude, the coastwise flights must exhibit, depending on the particular situation, from five to 130 times the concentrations observable among trans-Gulf migrants. This point seems almost too elementary to mention, but I have yet to find anyone who, in comparing the two situations, takes it into consideration.

Judged in this light, the average hourly density of 2,800 birds at Progreso in April would appear to be indicative of many more migrants on the entire potential front than the 6,300 birds representing the average hourly density for the same month at Tampico.

That the Progreso birds were actually beginning a trans-Gulf flight seems inevitable. The Yucatan Peninsula projects 200 miles or more northward into the vast open expanses of the Gulf of Mexico and the Caribbean Sea, with wide stretches of water on either side. The great majority of the birds were observed _after_ they had proceeded beyond the northern edge of this land ma.s.s. Had they later veered either to the east or the west, they would have been obliged to travel several hundred miles before again reaching land, almost as far as the distance straight across the Gulf. Had they turned southward, some individuals should have been detected flying in that direction. As can be seen from Figures 23, 42, and 44, not one bird observed was heading south of east or south of west on either night. No other single piece of evidence so conclusively demonstrates that birds cross the Gulf of Mexico in spring in considerable numbers as do flight density data recorded from Progreso in 1948.

_Northern Gulf Area_

Unfortunately only a few data on flight density are available from critical localities on the northern sh.o.r.es of the Gulf in spring. As the density curves in Figure 30 demonstrate, several sets of observation, including some phenomenal flights, have been recorded at Baton Rouge. This locality, however, lies sixty-four miles from the closest point on the Gulf coast, and the point due southward on the coast is eighty-four miles distant. Since all of the birds seen at Baton Rouge on any one night may have come from the heavily forested area between Baton Rouge and the coast of the Gulf, we cannot use data from Baton Rouge as certainly representative of incoming trans-Gulf flights. Data from repeated observations at stations on the coast itself are needed to judge the degree of trans-Gulf migration northward. On the few nights of observation at such localities (Cameron and Grand Isle, Louisiana, and Pensacola, Florida), flight densities have been zero or negligible. To be sure, negative results have been obtained at stations in the interior of the United States, and flights of low density have been recorded on occasion at stations where the flight densities are otherwise high. Nevertheless, in view of the volume of migration departing from Progreso, Yucatan, it would appear, upon first consideration, that we should at times record on the coast of Louisiana enough birds arriving in a night of continuous observation to yield a high density figure.

Upon further consideration, however, there are factors mitigating against heavy densities of birds in northern flight on the northern coast of the Gulf. In the first place, presuming the main trans-Gulf flight to originate from northern Yucatan, and that there is a directional fanning to the northward, the birds leave on a 260-mile front, and arrive on a front 400 miles or more wide. Consequently, other factors remaining the same, there would be only approximately half the number of birds on the coast of arrival, at a given time and place, as there was on the coast of departure. Secondly, we may now presume on the basis of the telescopic studies at Progreso, that most migrants leaving northern Yucatan do so in the few hours centering about midnight. The varying speeds of the birds making the 580-mile flight across the Gulf distribute them still more spa.r.s.ely on the north coast of the Gulf both in time and in s.p.a.ce. Also we can see only that segment of the flight, which arrives in that part of a twenty-four hour period when the moon is up. This circ.u.mstance further reduces the interceptive potential because the hours after dark, to which the present telescopic studies have been restricted, comprise the period in which the fewest migrants arrive from over the water. To ill.u.s.trate: it is a mathematical certainty that _none_ of the birds leaving Yucatan in the hours of heaviest flight, before 12 P. M., and flying on a straight course at a speed of approximately 33 mph will reach the northern Gulf coast after nightfall; they arrive in the daytime. It will be useful to devise a technique for employing the sun as a background for telescopic observation of birds, thereby making observations possible on a twenty-four hour basis, so as to test these inferences by objective data.