R e s e a r c h 
T h e   B O A C   L a b r a d o r   s i g h t i n g   of   J u n e   2 9,   1 9 5 4   visitor responses
 Possible explanations

4.1. Flocking starlings (Sorta Sol) or other birds

This is among the more interesting explanations proposed over the years for the BOAC case (sometimes known as the Centaurus incident). The shape-changing behaviour of the amorphous black object suggested to the Belgian researcher Wim VAN UTRECHT the intriguing idea that they saw a huge flock of starlings. The same theory was put forward by US researcher Brad SPARKS [13].

Below is part of the text (now removed) of Wim VAN UTRECHT's article on the starling cloud explanation as it appeared in the case examples section of this website.


In the early 1980s I had the opportunity to observe a very similar scene as the one reported by Captain HOWARD. This happened while I was standing at a bus stop in a village close to the city of Antwerp, Belgium. Waiting for the bus to arrive, my attention was caught by a dark, more or less oval-shaped mass, flying to and fro over some distant houses. Although the general aspect was that of a solid object, the dark blob not only changed shape constantly, it also seemed to increase and decrease in size from time to time. It was only after about ten minutes, as the phenomenon suddenly headed in my direction, that it revealed its true nature: a flock of birds. A "starling cloud", as I found out later.

Rereading the description of what was observed on June 29, 1954, we wonder if the captain, crew and passengers of the Centaurus were not victims of a similar illusion. The shape-shifting black object described by Captain HOWARD closely resembles a tight formation of thousands of birds silhouetted against a sunset background. The "inverted pear"-shape (see image BI-BS-04 in our bird picture gallery), that transformed into an "arrow" (the most common bird formation for long-distance flights), then morphed into a more elongated object (see BI-BS-09), growing and stretching "very much like a jellyfish that assumes varying patterns as it swims through the water" (compare with the videos of starling clouds listed under "more info and images" below the thumbnails in the gallery), all this matches the aspect and behaviour of a bird swarm. The "smaller blobs dancing attendance" on the bigger one, could have been smaller flocks that moved in and around the "mother cloud". A group of birds would also account for the objects "getting rapidly smaller" before disappearing. Possibly, at that moment, the formations broke up, either because the birds reached their destination and dispersed to find a roosting place on the ground, or because they sped away in different directions after being attacked by a hawk or other predator.

Browsing the specialised literature on the Internet, we found that the first record of a starling on Labrador dates back to 1917 [1]. This European starling (also known as common starling or sturnus vulgaris) began nesting on the peninsula in the 1930s and 40s [2]. The millions of starlings that can be found in North America these days are all descendants of only 100 European birds that were released in New York City in 1890 and 1891 [3].

More information on starlings was gathered and we learned that, during the day, the animals disperse as far as 15-20 miles from the major roosts in search for food. Some even travel up to 50 miles daily from their roost to reach their feeding grounds [4]. During these excursions, starlings navigate by geographical cues such a seashores and rivers, but when returning to their roosting grounds in the evening, they often use the setting Sun as a point of reference [5]. Interestingly, most of the starlings that nest in Newfoundland and eastern Canada spend the colder seasons in the more southern urban regions, many as far as New York [6]). So it's not only possible that the objects sighted over Labrador were birds returning from their feeding grounds, it is also possible that they were migratory birds returning from their wintering grounds in the south.

The starling hypothesis seems to fit nicely with most of the case details, but there are a few problems as well.

For example one may ask oneself if a flock of small birds can keep pace with a Boeing Stratocruiser flying at 270 miles per hour (434 km/h). The mean air speed of the starling is only between 40 and 50 miles per hour (between 64 km and 80 km/h) [7]. It would therefore seem impossible for a swarm of starlings to remain parallel with the plane for 18 minutes. The explanation could be that, when observed from a moving vehicle, objects that are at a great distance (the estimates given by Captain HOWARD range from 3 to 5 miles) can stay in sight for a long time, even if these distant objects travel at a much lower speed than the vehicle itself. The phenomenon is most noticeable with astronomical bodies. The Moon for example is so distant that it remains roughly in the same position when observed through the side window of a vehicle travelling in a straight line. Because of this, the illusion is created that the Moon is following the vehicle. The same is true for airplanes and bright stars or planets, all of which have been responsible for countless reports of cars being chased by UFOs.

A more difficult problem to solve is the altitude at which the phenomeon was situated. In his interview with Graham FISHER, Captain HOWARD stated that the objects were parallel to his plane and "at the same altitude" (19,000 feet or 5.8 km). However, on the NICAP web site HOWARD is quoted as having said:

"They were moving at about the same speed as we were (230 knots approx) on a parallel course, maybe 3 or 4 miles to the north west of us (we were heading NE). They were below the cloud at this time, at a guess at 8,000 ft. Soon after crossing the coast into Labrador, the cloud layer was left behind and the objects were now clearly in view, seeming to have climbed more nearly to our altitude. At this time the sun was low to the northwest, sky clear, visibility unlimited" [8].

In a 1988 article entitled "How Fast and High Do Birds Fly?", EHRLICH, DOBKIN & WHEYE write: "Most birds fly below 500 feet (150 m). When migrating, however, birds often do climb to relatively great heights, possibly to avoid dehydration in the warmer air near the ground. (...) Generally long-distance migrants seem to start out at about 5,000 feet and then progressively climb to around 20,000 feet" [9]. There appears to be no consensus amongst experts whether the starling is a short or long-distance migrant, but it is generally agreed that starlings rarely fly at altitudes above 1,000 feet. However, various authors point out that, under certain circumstances, the birds do fly at higher altitudes, for instance to locate familiar landmarks, fly over fog or clouds, surmount physical barriers, gain advantage of a following wind, or maintain a better thermoregulatory balance (actually, it is the cold at these altitudes that helps them disperse all the heat being generated by their flight muscles) [10]. Since starlings use directional information from the setting Sun, it is conceivable that they will sometimes climb to higher altitudes as not to let the Sun disappear from their view. In the same line of thought, the "foggy weather" and "low cloud at about 5,000 feet" may also have forced the birds to fly at a higher altitude than normal.

Wind tunnel tests showed that starlings are, however, reluctant to fly at ambient temperatures below zero and can't fly at -18° C because of ice formation in their nares and bills [11]. Unfortunately, no meteorological data are available, other than the captain's own statement that, once above the clouds, there was a "clear sky" with "unlimited visibility". To get an idea of the prevailing temperatures for the given time of year (June 29), time of day (evening) and sighting location (Labrador), we checked the soundings taken by weather balloons launched from Goose Bay between 1978 and 2007. Looking at the data for every June 29 from this 29-year period, we find that balloons released at 00:00 Zulu Time (8 p.m. Local Time) recorded temperatures comprised between -10.1° C and -29.3° C when they reached altitudes between 5.4 and 5.9 km (1,646 to 1,798 feet) [12]. In other words, it is possible, at least theoretically, that the objects spotted over Labrador on June 29, 1954, were starlings.

Still, and although starlings are the unchallenged masters in forming well-defined, smooth-looking shapes in the skies, there are several other species of birds that have their breeding grounds in Labrador and gather in massive swarms before assuming a V or arrow-shaped formation for their long-distance flights. Examples are godwits, curlews, dunlins and snow geese, four species that are known to fly at altitudes of 20,000 feet and more [13].

 Our opinion

The explanation we propose is that the captain, crew and passengers of the Centaurus witnessed a distant bird swarm, possibly starlings, backlit by the setting Sun. This explanation is supported not only by Captain HOWARD's detailed account that describes the phenomenon as "black blobs" behaving like "a jellyfish that assumes varying patterns as it swims through the water", but also by the time of day (just before sunset, when most birds return to their roosting grounds).

The major shortcoming of the bird hypothesis is the altitude at which the objects were situated, and which is rather atypical and requires somewhat exceptional circumstances, at least in the case of a starling swarm.

 Notes & References

[1] (pdf).

[2] (pdf) and



[5] and

[6] and

[7] and



[10] More about the altitude at which birds fly can be found on:
- (pdf)

[11] (pdf).

[12] Atmospheric soundings for most parts of the world can be consulted at


The first reference to this theory (to dismiss it as "highly improbable" on the grounds of speed and altitude) appears to have been by Capt. HOWARD himself in an 800-word entry in his Voyage Report written during the Goose-London leg of the flight (CARNELL, 1954; see Appendix B), and as reported in the papers the next day [14][15]. Another similar speculation was published in the London Daily Mail one week later [16].

Such flocking displays can involve hundreds of thousands of birds and do resemble dark shape-shifting blobs in the sky, a phenomenon widely known as Sorta Sol which is Danish for "black sun" [17]. One reason for the especially well-defined shapes of starling flocks is that the bird density increases towards the periphery of the flock, rather than towards the centre, which is believed to be related to defence against predators [18]. Capt. HOWARD in fact used a related simile: "The large object was continually, slowly changing shape, in the way that a swarm of bees might alter its appearance" (HOWARD, 1982). Wim VAN UTRECHT proposed that the apparent stationing of the objects off the aircraft's wing for an extended time might be an illusion of distance perspective, akin to that whereby the moon appears to follow a moving vehicle.

starling cloud Fig. 4 : A starling cloud flanked by four geese over Tøndermarsken, Denmark. [© Bjarne WINKLER - Photo used with permission] - For more images and links to videos of flocking starlings, see the picture gallery on birds.

However, in this case the minimum relative velocity of the 230 kt (265 mph) plane and a starling (assuming the 40 knot maximum velocity of the starling [19] is parallel to that of the plane) is about 190 kt (220 mph). In 18 minutes the starling falls aft the plane by a distance of 66 statute miles. Capt. HOWARD said the UAPs stayed "parallel" and "paced" the Centaurus "exactly" for 18 minutes. This was checked by the navigation officer against the windshield post. (The only reported deviation described in early sources was that the UAPs appeared to move ahead of the aircraft a little at one point and then fell back to the same abeam position [20]. This could maybe have been yaw in the aircraft axis [21].)

Clearly, to explain a 190 kt airspeed difference in terms of differential tailwinds we require the flock of birds to be inside - and the aircraft outside - a very sharply-defined southwesterly jet with a core windspeed greater than this. There are strong reasons why this scenario is highly unlikely even in principle. The reader is referred to the lengthy discussion in Chapter 4.2 (stratospheric balloons) where extreme jet winds are the overriding requirement. It is sufficient to note here that the wind in the sighting area at the 19,000 ft level (500 mbar) measured by radiosonde ascents between 10:00 p.m. and 11 p.m. EST June 29, 1954 was only 25 kt. [Appendix A (ix)], an order of magnitude too slow. There is no indication of violent wind at 19,000 ft in the sighting area. This fits the fact that due to a continental blocking high over Labrador the mid-latitude jet was not only abnormally weak in the East through June 1954, its average position [traced at the 700 mbar level in Appendix A (x)] was confined in a zonal flow far to the South.

In the absence of a violent jet stream we are reduced to considering parallax. Let us assume that despite the report of "exact" pacing the blob fell back over 18 min by as much as 10°, say, which might seem roughly a constant constant to a lay reader - although it is a large angle navigationally speaking. Perhaps a 10° drift was concealed by yaw, or perhaps the pilot and navigator didn't check the angle as accurately as was suggested. In order for this 66 mile distance to equate to 10° our starlings would have to be nearly 380 miles away. (If we require to respect a bearing accuracy of 1° then obviously the distance increases to 3800 miles!)

But if 10° were good enough, could a flock of starlings be observed at 380 miles? Capt. HOWARD estimated the angular size of the main blob when in its "telephone" phase as equal to that of an ocean liner at 5 miles. If an ocean liner is on the region of 1,000 ft (typical for biggest transatlantic vessels in the then-familiar class of the France and Queen Mary [22]) then an angular width in the region of 2° is implied. In his Colorado Project Sighting Report form HOWARD answered question #36 ("What was the APPARENT size of the object compared with the following familar objects?") with "about the same size as the moon", or 0.5°. Of course this was many years after the fact. And not all ocean liners are as big as the iconic translatlantic superliners. If we make due allowances and split the difference we end up with an angular size in the order of 1° [23]. At 380 miles range that means the main flock would be nearly 7 miles across, and the outliers (spanning 5-10 times the angular width of the main blob according to HOWARD's drawings) would cover about 35-70 miles.

According to all quoted witnesses (including crew, cabin staff and a New York BOAC employee flying as a passenger [24]) the blobs were "black", "dark", "solid", "substantial". Even allowing for optical contrast illusions in silhouette this implies a very substantial density of birds, and given a lower bound on flock diameter of several miles the absolute number of birds would be truly colossal. For example, a spherical flock of only 2 miles (3km) radius having a mean density of 1 starling per m³ would contain some 36 billion (36x109) birds, which exceeds the typical numbers in large flocks by 4 or 5 orders of magnitude. Indeed this is in the order of 10 times the estimated total spring and fall traffic of all migratory birds on all flyways throughout the continental US.

Consider also that although the sighting did occur close to the major East coast migratory route known as the Atlantic Flyway (even though not in a migratory season) the impressive dusk flocking displays always occur at low altitude when the birds are roosting, and are formed by the aggregation of very much smaller groups and individuals that pre-assemble in trees and other perching sites over a wide area.

The flock is actually not so much a blob as a pancake, tending to form in a sheet which is relatively thin in the direction of gravity and extended in directions perpendicular to it [25]. The characteristic morphing-blob appearance occurs because the changing aspect when twisting and banking over the nesting site causes transient increases in the density of birds per unit angular area. The shapes change very rapidly for this reason, and the optical density of the blob thins out in between to the point where it becomes a faint smudge or streak.

This geometry means that if a Sorta Sol flock were to be transported to a migratory route at 19,000 ft and made to fly straight and level alongside an aircraft then it would tend to be viewed along its axis of maximum optical density. But as already mentioned, these flocks are roosting displays and are not known to occur at altitude during point-to-point flight. Passerine birds (among which are included starlings) do not form concentrated flocks at all during their migrations [26] and although millions of birds may be involved the process may last many weeks [27]. For starlings in the US, for example, the mass northern migration back to the breeding grounds occurs from late September until the end of November [28]. (A million birds over this period would pass a given point at an average rate of only about 700 per hour, or 180 birds per 18-minute observation period.) Other birds like geese that do form tight flocks at high altitude in order to migrate do so in small numbers in distinctive and recognisable formations, not in vast swarms.

The mass migrations of all species occur in local climatic Spring and Autumn, and although migratory dates in more northerly latitudes will tend to contract towards the breeding peak - so that Autumn migration in the Canadian High Arctic (10°- 20° or more N of the sighting area) is beginning erratically by late July [29] - late June is near the peak of the breeding season everywhere, when the vast majority of birds are near their nest sites.

It is true that the European starling is one of only two birds (the other being the lapwing [30]) found to have a secondary Summer migration, composed of that year's juveniles which often fly a short distance (a couple of hundred miles) in the direction of the wintering grounds, as though to get a head start [31]. But a small sub-population of summer-migrating starlings trickling South does not in any realistic way help to explain our dense, black UAPs apparently heading North.

Brad SPARKS has suggested [32] that the Sorta Sol theory might be rescued if there were numbers of large high-altitude bird flocks distributed over a wide area of Eastern Canada, an abnormal mass migration possibly due to some freak climatic trigger, so that the witnesses saw different concentrations of birds in succession. If their visual attention was interrupted they might have missed the transitions and assumed they were seeing the same group of "objects". This might account for an illusion of a single object (or group of objects) appearing to "pace" the airliner, even though the individual bird speeds remain low.

No doubt each of the witnesses did take his or her eyes off the objects from time to time during the 18 minutes, but it seems safe to say that a lot of people were looking a lot of the time, with the front seaters probably watching most of the time. Yet there is nothing in any account to suggest that anyone ever thought the object(s) disappeared and reappeared at any time.

Nevertheless, allowing (despite absence of evidence) the conjecture that flocking behaviour comparable to the Sorta Sol might occur during point-to-point flight at 19,000 ft, a relay of different flocks would allow us to reduce the range and reduce the absolute size of the flocks. But only to the same extent that the stationary average bearing can no longer be an illusion of vanishing parallax caused by great distance. So although individual birds in these flocks can be allowed to fly at arbitrarily low average speeds (speed can be zero in the limit of an infinite number of successive flocks!) the average position of the successive flocks must still advance at close to the aircraft speed.

In order to preserve an illusion of continuity how many "hand-overs of the baton" are necessary in this relay race?

Let's keep an allowance of 10° for an undetected or unreported drift per flock before it needs to be replaced by a new one, but cut the distance assumption down by a factor 10 so the flocks are 3.8 miles (6 km) away. A drift of 10° is therefore 0.66 miles. A 1° diameter flock only 3.8 miles away is <400 ft (120 m) across and with a density of only 1 bird/m³ the flock contains nearly a million birds, which is the right order for the largest Sorta Sol flocks of starlings. Assuming that all of the velocity of all of the starlings in the flock stays parallel to the aircraft (probably an unrealistic limit) then at the minimum relative speed of 225 mph it has drifted aft to our limiting 10° in about 10 seconds. This million-bird flock then needs to somehow dissipate and be replaced by a new one displaced angularly forward by 10°, and the same thing has to recur more than 100 times during 18 minutes - each time neatly missed by every eyewitness and each time reassembling a similar pattern of smaller sub-flocks deployed symmetrically fore and aft. The total number of birds in this amazing performance is in the order of a hundred million birds at least.

Brad SPARKS suggests [33] trying to make the theory work with a lighter density than 1 bird/m³, a smaller 0.5° degree UAP diameter, and an "accordion effect" in a continuous concentration of birds rather than successive discrete flocks.

One can visualise this idea as an 80-mile ribbon of birds lying parallel to the aircraft track, having a sub-critical density of birds that becomes a visible blob only at the peak of a compression density wave propagating along the ribbon at the speed of the aircraft. Why this would happen - or why some other, more complex disposition of birds would create the effect of it happening - is not clear to me. But it's obviously true that we can reduce the absolute number of birds by assuming a lower peak density, although this is in tension with the need to maintain a "black", sharp-edged" and "substantial" appearance insisted on in early observer accounts.

An empirical study of starling flock parameters [34] measured bird densities in ten events. An average over the ten values gives 0.27/m³, or about 1/4 of the value assumed above (within an order of magnitude - which is not a large difference in the context). Video evidence shows that the average effect of this typical density of birds in Sorta Sol displays is not by any means "solid" and "black", the flock is only intermittently dark in its most favourable aspect. We should therefore assume that this peak density is the minimum value necessary to maintain a constant black, sharp-edged silhouette for 18 minutes in flocks of comparable size.

Obviously the optical depth and edge-definition of a flock due to a given density of birds will also be proportional to the physical depth and thus to the absolute number of birds. We do not wish the absolute number of birds to drop too low. If the peak group is, say, 50 m spherical diameter (for simplicity) and subtending 0.5° at the above distance of 3.8 miles (6 km) then at the typical density of 0.27/m³ it contains only about 10,000 birds. Since the measured typical density provides only an unsatisfactory minimum opacity and definition in Sorta Sol flocks that regularly exceed this absolute number by a factor of 10 or 100, we should apply at least a factor 10 or 100 correction. Then for the same density a more comfortable 100,000 birds roughly doubles the diameter and range of the 0.5° peak-density group to 100 m at about 7 miles (11 km).

Let us plug this figure into SPARKS' "accordian wave" model. Clearly the just-invisible medium for this travelling density peak has to be made of a very much larger number of birds at a density not far below the peak, thus for an 80-mile (130 km) stream of birds the total must still be in the order of 1,000 times the number of birds in the visible 100 m peak blob, which implies, conservatively, a number in the order of 108 or 109 birds concentrated within a few miles of the Stratocruiser during one 18-minute period on June 29, 1954.

This reduced figure is still between 10% and 100% of the estimated total spring and fall traffic of all migratory birds on all flyways throughout the continental US today [35] and seems entirely incredible. The passenger pigeon, historically the most abundant US bird once comprising 25% - 40% of the entire US bird population, is believed [36] to have formed a migratory river up to 300 mi (500 km) x 1 mi (1.6 km) containing up to 2.5 billion birds, or in the order of 1 million birds per km² of plan area, a density comparable to the 1/m² concentration in Sorta Sol flocks. But apart from the anomalous migratory date, the passenger pigeon was totally extinct in the US by about 1,900. No other bird existed in remotely comparable numbers in 1954.

Add to an extraordinary number of birds the facts: - that Sorta Sol is low-level roosting behaviour that occurs only where much smaller flocks and individuals, returning from feeding grounds, congregate over the nest sites; - that passerines like starlings do not tend to flock densely when migrating; - that starling migration is minimal on the Atlantic Flyway (most East Coast birds are resident, and most of what little migration there is comes from US populations flying south, not from Canada) [37]; - that very few birds migrate at all above about 7,000 ft and that most starlings fly below 1,000 ft; - that migration times across the US are Feb-March and Oct-Nov [38], not during the midsummer breeding season [39]; - and that there is no obvious mechanism by which the density peak in this hypothetical great river of birds would track the motion of the plane at a rather consistent bearing for 18 minutes and we conclude that the Sorta Sol theory has several, collectively fatal, problems [40].

 Notes & references

[13] SPARKS, B., personal communication, 05.08.2009; SPARKS, B., Comprehensive Catalogue of 1,600 Blue Book Unknowns, 2001-2008, Case #938,

[14] "Flying Saucers? Yes, says the Captain", Daily Express, July 1, 1954.

[15] "Sky 'objects' Seen from Plane - Pilot's Report", London Daily Telegraph, July 1, 1954.

[16] MAY, K. R. (Salisbury, Wilts.), letter to the Editor, Daily Mail, July 7, 1954, describing roosting flocks seen from a train in South India in 1945. An Editorial comment dismisses the idea on the grounds of altitude.

[17] Many extraordinarily beautiful videos of Sorta Sol can be found on YouTube, for example: video1 or video 2.

[18] BALLERINI, Michele, et al., "An empirical study of large, naturally occurring starling flocks: a benchmark in collective animal behaviour", Animal Behaviour 76(1), July 2008, 201-215.

[19] HOUGHTON, E., "Detection, Recognition and Identification of Birds on Radar", 1964; cited in: BLACKMER, R. H. et al, "Radar and the Observation of UFOs", in GILLMOR (ed.) & CONDON, Scientific Study of UFOs, Vision, 1970, p. 675.

[20] In his December 1, 1967 Sighting Report written for the University of Colorado UFO Project, Capt. HOWARD draws a diagram of initial and final compass directions showing the LOS advancing fully 20° towards the nose of plane. This large and systematic rotation of angle cannot be accounted for by yaw and is in conflict with early statements that the objects "kept station" on the same (average) bearing. If reliable, of course, a steady 20° LOS overtake would be sufficient to rule out the Sorta Sol theory immediately.

[21] According to Capt. HOWARD the aircraft track (autopilot setting) was 49° True and the instant heading would have been 49 ±5°, i.e. a possible variation of up to 10° (letter from R. H. B. WINDER to Robert J. LOW, August 30, 1967, forwarding Capt. HOWARD's answers to question posed by LOW. Colorado University UFO Project files; Library of the American Philosophical Society, Philadelphia).

[22] According to "Captain James HOWARD later described the main object as having 'the size of the Queen Mary'!" but no source is given for this quotation.

[23] This result can be cross-checked against Capt. HOWARD's drawings (Fig. 1) and the blueprint dimensions of the Boeing 377 from which we can calculate that the angular diameter of the engine nacelle ~36 ft from the cockpit would be approximately 5°. The setting sun is of course 0.5° across. Scaling the angular width of the UAP in Sketches #2 and #3 using these two yardsticks yields values of about 4° and 0.5° respectively, and in #1 the values are about 0.2° and 2° respectively. The overall mean of these six values is ~1.9°, reassuring us that an angular diameter of 1° is reasonable and probably conservative.

[24] "Romford Man Sees Flying Saucers", The Recorder, Romford, Hornchurch & Upminster, Friday July 2, 1954, p. 1.

[25] BALLERINI, Michele, et al., op. cit.

[26] ELPHICK, J., Atlas of Bird Migration, New York, Random House, 1995.

[27] KESSEL, B., "Distribution and Migration of the European Starling in North America", The Condor, Vol. 55, # 2, March-April 1953,

[28] KESSEL, B., op. cit.

[29] GUDMUNDUR, A., et al., "Radar Observations of Arctic Bird Migration at the Northwest Passage, Canada", ARCTIC, Vol. 55, # 1, March 2002, pp. 21-43.

[30] According to the American Birding Association the Northern Lapwing, Vanellus vanellus, is classified as an occasional non-breeding visitor to the northeast US and Canada. The estimated global population of 7 million birds breeds in Eurasia and winters mainly in Asia Minor, India, and Southeast Asia. There are over 50 historical reports of Northern Lapwings from eastern Canada and the northeastern US, but large flights in December 1927 and January 1966 account for most of the records. I find no reference to extraordinary Summer visitations in 1954.

[31] ALERSTAM, Thomas, Bird Migration, Cambridge University Press, 1990, p. 41.

[32] Personal communication, September 8, 2009.

[33] Personal communication, September 9, 2009.

[34] BALLERINI , Michele. et al, op.cit.

[35] ELPHICK, J., Atlas of Bird Migration, New York, Random House, 1995.

[36] and

[37] KESSEL, B., "Distribution and Migration of the European Starling in North America", The Condor, Vol. 55, #2, March-April 1953,

[38] GUARINO, Joseph L., "Bird Movements in Relation to Control", Wildlife Damage Management, Internet Center for Bird Control, Seminars Proceedings, University of Nebraska - Lincoln, 1968,

[39] Limited records indicate an unusually warm and generally settled regime over the area of Quebec on June 29, 1954 (see Appendix A). However on June 24, 1954, Hurricane Alice hit Texas, flooding the Lower Rio Grande Valley with 27 inches (686 mm) of rain. U.S.90 was 30 feet (9.1 m) underwater ( Conceivably, the approach of Alice could have disturbed huge numbers of birds from nesting sites in the south, but I have so far been unable to find evidence.

[40] Wim VAN UTRECHT is now satisfied that the starling theory has been ruled out (Personal communication, May 11, 2009).