The Tree Swallows (Tachycineta bicolor) on Middleton Island in 2017

2017 was the ninth year for the Tree Swallow (TRES) research project out on the small (8x2km) and remote Middleton Island, located in the Gulf of Alaska (59°26′59″N 146°18′26″W). It’s one of the most isolated breeding populations and research sites (that’s thus far assumed to be largely predator free) for a beautiful bird that is considered a ‘model species’ in science. This year the Institute for Seabird Research and Conservation (ISRC) provided information on the TRES breeding season after a single nest box inspection (and cleanout) after the breeding season, of the 20 boxes currently present on the island. No in-depth study on their breeding biology this summer, but basic data on box occupancy and an indication for reproduction in 2017 instead. Some additional info and photos came from ISRC 2017 crew members. On 19 September Scott and Martha Hatch recorded the following nest box contents:

1. nothing
2. nothing
3. nothing
4. complete feather-lined nest with feces
5. nothing
6. complete feather-lined nest (feces?)
7. complete feather-lined nest with feces
8. complete nest with 3 dead, nearly fledged chicks
9. large straw nest base but no feather lining
10. nothing
11. complete nest with 5 dead, nearly fledged chicks
12. nothing
13. nothing
14. complete nest with 1 unhatched egg, also feces
15. nothing
16. nothing
17. nothing
18. nothing
19. nothing
20. nothing

(Complete nests with feces are expected to have resulted in fledged young)

Besides these data, I received four photos and some opportunistically noted observations of the birds around their boxes, in particular box 7 (the one facing the FAA weather camera), from which chicks fledged, which seemed to get visited by a TRES as late as July 17 (also the last recorded sighting of a TRES that I received).

Another year added... (By the way, after analyzing the data again and for accuracy I adjusted the number of occupied boxes for 2013 from 10 to 9 (not counting a few straws as an occupied box anymore)).



Distribition of occupied nest boxes in which eggs were laid in 2017.




The FAA weather camera (http://avcams.faa.gov/) was offline throughout most of the summer, due to the installation of a new camera and further maintenance, but it’s up and running agian now and box # 7 is still visible.


Discussion (& some serious speculation...)
2017 was a relatively good year as far as the number of nesting TRES pairs goes. Total reproduction remains unknown, but was likely relatively low due to chick mortality occurring shortly before fledging. During the previous years of the project chick mortality has been low and dead TRES chicks were a rare sight, at least for me. The thought that came to mind among some was that perhaps a predator had been picking off the adults this summer, cutting off food supply to their young. Has predation become an issue for the largely predator free population? ISRC crew members observed a Peregrine Falcon Falco peregrinus, the first predator that would come to mind, on the island but only in May. Additionally, I suspect that an adult TRES would not be an easy prey for a falcon, while around the same time the island would be full with other, more easy prey types (i.e. young sparrows, snipe?) (But Peregrines will lose many larger prey items to the Bald Eagles Haliaeetus leucocephalus there, perhaps the reason they're not nesting there, and this could increase predation risk for the TRES(??)). But no Peregrine was seen in June and July, when the swallows have young.

Previously, on Middleton chick mortality only had a significant influence on reproduction in 2012. The chicks involved that year were also large and close to fledging. Bad weather (cutting off the adults from their food supply) was the obvious cause that year. When looking at the historic climatic data for 2017, the end of June/first half of July appears to be a good period for TRES (http://rp5.md/Weather_archive_in_Middleton_Island_%28airport%29) Unfortunately data from this on site weather station appears to be incomplete again and therefore unreliable). The only significant change to bad weather seemed to occur around July 14 - 18 (complete days missing here), but it's the type of weather that I would expect to kill TRES chicks. During my stays, however, TRES chicks fledged by the end of June/first half of July, so were these boxes still occupied by the time real bad weather appeared? Perhaps they were, since the last sighting of a TRES made by the 2017 crew was on July 17, of a bird exiting box # 7. On Middleton, the TRES do not appear to visit their boxes after fledging (and box 7 resulted in fledged chicks). And when looking at the historical data of this project, first eggs can be laid as late as June (June 2, in 2012), which could indicate that boxes can still contain chicks during the second half of July. So, with only little data to support this wild speculation, 2017 was also a late breeding season. It would be interesting to compare these thoughts on the TRES breeding season again with whatever happened at Long Lake (the nearest TRES study site in Southeast Alaska) this summer, but unfortunately I don't seem to be able to get in touch with the responsible there.


Update 10-25: here's what Chris Gates (www.treeswallowprojects.com) wrote me:

Hi Tim,

Thanks for Middleton update. It's always nice to hear some of these hardy birds made it to the island and nested successfully. I put some of your information about the blog and links to it on the facebook group Tree Swallows, so others here in the US can learn of your efforts.

I do have a comment regarding the possibility of predation on TRES by peregrines and other falcons. In eastern North America TRES form very large flocks after nesting, which usually stage for the summer months in and around large marshes. By late summer and autumn the flocks begin to migrate south with heavy concentrations along the Atlantic shore. Here, the swallow flocks are followed by falcons, peregrines and especially merlins, which apparently take them with relative ease. I've attached a photo of a merlin "packing a lunch." Although the swallow in the photo is an adult I expect hatch-year swallows are heavily targeted. Also, in one of my favorite books, "Curious Naturalists," Niko Tinbergen describes behaviors of hobbies hunting swallows along the North Sea coast. So I guess if peregrines or merlins happened to summer at Middleton some year they could be potential predators of both tree and bank swallows there.

Regards, and thanks again,

Chris gates

(I've seen Merlins there regularly in spring, but never in summer. I recall installing two artificial stick-nests for the species in 2006, hoping to attrack a pair...) 

A Merlin taking a TRES... (photo supplied by Chris Gates)

Update 2: Alexandra Rose informed me 2017 was the latest breeding season she's seen in 12 years of TRES monitoring at Long Lake...

A big thank you to Scott and Martha Hatch for their participation in the project, as well as to the 2017 crew members supplying some additional information and these photos of the 2017 MDO TRES breeding season!!

The Tree Swallows (Tachycineta bicolor) on Middleton Island in 2016

2016 was the eighth year of the Tree Swallow (TRES) nest box project out on the small (8x2km) and remote Middleton Island, located in the Gulf of Alaska. It provides housing to one of the most isolated TRES breeding populations of North America. I myself had the privilege to monitor them this summer, during a stay on Middleton between May 15 – July 15, covering most of the TRES breeding season.




The first TRES for 2016 was seen shortly before my arrival on May 14 (by employees of Alaska Fish & Game), at the central cluster of boxes near the Military pond. During a first round of nest box checks on May 22, it appeared that three pairs occupied three boxes, all containing nests (one unfinished and two finished)…




… and in box 6 also the first eggs for the season (a single 3-egg clutch). Back calculating (with a rate of one egg laid per day, usually laid in the AM) this would mean a first laying date of May 20. These 3 nests resulted in 3 clutches, all containing 6 eggs. Clutches in box 8 and 11 were initiated on May 25 and 29 respectively.




All of these eggs hatched the following weeks…




… allowing to have banding of 18 circa 12 day old TRES-chicks take place between June 18 – 27...




...under the watchful eyes of their parents, which during the operation found their entrance hole blocked by paper tissue. All 18 chicks fledged by the end of June/first half of July.




But things didn’t go this easy for all TRES on the island. An additional number of lonely males (probably 3 - 4) occupied single or clusters of nest boxes. These birds spent most of their days perched on top of their box in the hopes of attracting a partner for the season. The male on the photo above occupied the northern box cluster. It must have gotten quite desperate, as it was observed displaying towards Bank Swallows on a few occasions during the second half of June. The opening photo of this post shows a male that spent most of its summer on top of box 7 (facing the FAA weather camera). It never found a mate, but none the less brought a 5 cm thick layer of white gull feathers into the box (which I don't record as nest). Other places where a lonely male was seen regularly were box 13 and the southern cluster (boxes 15 - 17).




The late appearance of a second calendar year female on the island, first recorded together with a male around box 14 on June 8 (not known to be occupied before that date), must have caused excitement among the lonely males on the island. The two birds would make up Middleton’s fourth and final nesting pair for 2016. On June 9 this female copulated on the box, which on that day contained some feathers as nest material. On June 16 their nests was finished, but instead of them, a total of 4 very noisy birds (including 3 adults and 1 2cy female, photo above) got recorded, harassing each other around the box. On June 20 the nest contained 3 eggs (first laying date: June 18), but the two birds present appeared to both be adults… When on June 27 3 cold eggs and no birds get recorded in or around box 14, the breeding attempt had apparently failed.




What exactly made the pair from box 14 fail remains unknown, but when on July 8 the unoccupied box 15 (located 2 km from box 14) was opened for maintenance, the decomposing remains of an non-banded second calendar year (born in 2015) female TRES were found (thanks to Chris Gates for confirming the ID). Since this box hadn’t been opened since June 16, a time of death remains hard to determine, but the state of decomposition indicated that ‘weeks’ seems more likely than days, making it reasonable to assume that this was the young female from box 14 (last seen on June 16). Were it the lonely male TRES that harassed her to death?









Bank Swallows Riparia riparia were abundant breeders this summer. In 2005, 2006, 2009 and 2010 I counted the total number of nest holes, resulting in a varying amount per year (112, 48, 65 and 106, respectively). This summer a total of 203 holes was counted! In 1956 Robert Rausch recorded the species for a first time, with only a few pairs.
And many young Bank Swallows fledged this summer, which would often hang out together with the young TRES, making it very difficult to catch a glimpse of the young TRES in a brown mass of young swallows. I believe that among the swallows on this stick above the Military pond there´s at least a few young TRES… (or maybe all are?).




The 17 nest boxes withstood the 7 – 8 years of exposure to the Gulf’s extreme weather conditions quite well. But the scheduled maintenance of the project’s hardware took up an unexpected large amount of time. In the end all boxes got a fresh layer of paint, but two needed to be replaced, one repaired...




…and several had to be provided with a new swallow perch (a bended metal pipe at the top), also a measure to keep eagles and gulls away from the boxes. In the absence of this device, this Glaucous-winged Gull Larus glaucescens managed to put its feet on wet paint…




Three new boxes were added to the project this summer, including this one along the Dump Road, making a total of 20 nest boxes currently present on the island. As a tryout, two of these were constructed from white PVC planks, in order to see whether this material suits the TRES needs, and if it has a longer life span than wood. I thank Scott Hatch (ISRC) for providing the materials for this project!

A new map of the new situation, and the boxes used by the four breeding pairs in 2016’s. Boxes 18 – 20 were placed later in the breeding season and were therefore not available to the birds.

Now it's all set up and ready for a new season…

The Tree Swallows (Tachycineta bicolor) of Middleton Island, Gulf of Alaska

The Tree Swallow (TRES) is a migratory passerine bird that breeds in North America and winters in southern U.S., Mexico, Central America and the Caribbean. They're the species of choice for many North American ornithologists pursuing songbird research. This blog entry presents the results of the first four years of a pilot project involving a TRES population and describes and discusses how these birds established themselves on a remarkable location. It will also look in the future of what possibly is one of the most isolated breeding populations of this species in North America.

Tree Swallows (photo: Chris Gates)

I first found out about the existence of the TRES in the summer of 2005, when I worked abroad, volunteering in the USGS seabird studies out on the small (2 x 7 km) and remote Middleton Island (59°26’N, 146°20’W) located in the Gulf of Alaska. Middleton lies about 110 km from the nearest mainland and 80 km southeast from the nearest island, Montague, which forms the southern barrier of the Prince William Sound. On May 26 of that year I observed three of these birds flying around at the island’s north end. These birds did not stay and the species did not breed on the island during that year, neither did the few birds that were observed by the end of May during a following stay in 2006. I expected that this species was absent from the island as a breeder due to a lack of suitable nest habitat, as their name implies a strong relationship with ‘perennial woody plants of reasonable height’. Middleton, however, is not completely treeless; there are several solitary or clustered windblown spruces scattered around on the island. I doubted that any of these contained cavities for these swallows to nest in. Apart from a lack of nest habitat, the island would probably be an interesting place for them, as there are many fresh or brackish ponds and plenty of flying insects, already satisfying the needs of about 65 - 112 pairs of Bank Swallows (Riparia riparia) annually between 2005 and 2010 (Van Nus, personal records).

Location of Middleton Island in the Gulf of Alaska (source: Google Earth)

Middleton Island is relatively low, flat, and has only a few trees (photo: Lauren Bessey)

Surprisingly, Rausch (1958), who thus far published the only complete breeding bird inventory for the island, mentions the TRES as a breeder in 1956, as he observed a few of these birds nesting in the remains of a building. I would later found out this building was a U.S. Coast Guard station located in the center of the island, which was only operational for a few years during World War II. This building got burned down in 1960, probably eliminating the species from the island as a breeder during the following years.

The former U.S.C.G. buildings where TRES nested in 1956 (photo: Armand Biron).

With an increasing interest in historical records of the breeding species of Middleton Island (I've been working on an avifauna for the island for a while now), I would later find out that during the following decades the TRES made irregular appearances as a breeder, with at least one pair utilizing cavities in a building within an abandoned FAA settlement in the north end of the island where it got recorded nesting occasionally during 1976 - 2002. Unfortunately for these birds this settlement got removed during the summer of 2005, again leaving them without a place to nest on the island.

Part of the former FAA settlement at the north end, removed in 2005
(photo: Tim van Nus)

It occurred to me during a stay in Maine (U.S. east coast), after observing a so called ‘nest box grid’ made up of several dozens of pole-mounted nest boxes placed in an open field, that the TRES on Middleton could possibly be persuaded to nest on the island again. All I had to do was place a few nest boxes! So, I gathered more information about the species, nest box design and - placement. Especially Chris Gates’ website (http://www.treeswallowprojects.com) was of great help, as it describes in detail how to start up a TRES nest box project.

Shortly after arrival on Middleton in early April 2009, I constructed seven nest boxes. The dimensions used were those recommended by the Golondrinas project (an international community of biologists studying swallows (http://golondrinas.cornell.edu)). However, in order to survive Middleton’s stormy conditions year-round, their overall construction had to be a lot more robust than the boxes used elsewhere in the world. Five boxes were mounted on existing wooden poles that hold a non-functional electricity wire in the centre of the island and another two were placed nearby and were mounted on steel pipes. With the absence of ground predators on Middleton I did not have to worry about predator guards.

Box 1, 2 and 3 at an individual distance of about 30 meters
(photo: Tim van Nus).

Instant success! – In 2009 the first TRES (three birds) were seen on May 25. Two of these made up a pair and were highly interested in box 6. Both of these birds were in full adult plumage. A third adult male showed interest in the other boxes, but this bird got accompanied by a second year female only much later, on June 12. These two birds also paired and occupied box 2. During an inspection of box 6 on June 18 a nest was found containing six warm eggs. These resulted in six chicks, which all fledged around July 8. The couple nesting in box 2 laid a total of four eggs, which were observed on June 28, and which resulted in the fledging of 4 chicks around July 21. All these birds seemed to have left the island by the end of July. So, during the first year of this project two TRES pairs laid a total of 10 eggs, which resulted in 10 fledging chicks!

One of the birds from box 6 observing the observer (photo: Tim van Nus)

Box 6 containing six eggs and many feathers of waterfowl
and Black-legged Kittiwakes (photo: Tim van Nus)

2010 – My expectations for the following summer were high; I expected more birds to occupy the boxes, as some of the known breeders could return, as well as some of their offspring and maybe a few more 'floaters'. In order to provide enough nest sites for the near future of the project I added another 10 boxes to the seven that were already there and had survived some severe winter storms seemingly without problems. This made a total of 17 boxes present on the island in 2010 and these are all the project contains until the present day. All boxes from this new series were painted white on the outside for extra durability and were mounted on steel pipes, except for one (box 8), which I mounted on one of two existing slanted wooden poles. These two partly burned poles are the only visible remnant of a U.S. Coast Guard station that once contained Middleton's first TRES nest in recorded history (Rausch 1958). Five boxes (# 10 - 14) were placed along a small water stream in the north end of the island where the former FAA station was located. Another three boxes (#15 - 17) were placed just south of the southern end of the island’s infrastructure, making sure that the birds had the opportunity to nest well away from each other, if they’d prefer. I ended this series by placing a box in front of an FAA weather camera; there are two of these cameras located near the island's airstrip, and both take a photo every 12 minutes to inform pilots about the local weather conditions. The box (9) is facing the southwest camera. These images can be viewed online, meaning that when not visiting the island, the swallows could be viewed from any office (http://akweathercams.faa.gov/sitelist.php). Coincidentally, the first TRES for 2010 got recorded by the camera on May 10, when it perched at the entrance hole of box 9.

 Box 9 facing the FAA weather camera (photo: Tim van Nus)

The first recorded Tree Swallow observation for 2010; the bird is that tiny black spot against the front of the nest box.
Only later did I find out that I'd better placed the box a bit closer to the camera...
(Source: Federal Aviation administration/http://akweathercams.faa.gov/)

Location of the 17 nest boxes on Middleton

After the observation of the first swallow at box 9 on May 10, a second individual was seen the same day. Eventually in 2010 three pairs were present on the island, of which only two nested successfully. These birds, however, decided not to use the box facing the camera. Instead one pair nested in box 6 (same box as used in 2009) and had two eggs on May 28. The second pair laid their first egg on May 30 in box 8. Both nesting attempts resulted in six-egg clutches and six fledglings, which all left their boxes during the first days of July. The presence of a third pair only got recorded twice around box 4 on May 20 and 22. The female of this pair was observed in several severe fights with the established female from box 6 and was not seen afterwards for unknown reasons, but due to the severity of their battles I suspect that she may have been killed by her neighbor. At inspection, their box (# 4) contained only a few straws and feathers, which for me is enough to count them in as the third pair for this year. To summarize: in 2010 three TRES pairs occupied the boxes, of which two produced a clutch of six eggs, resulting in 12 fledged chicks.

After its partner disappeared, this lonely male occupied box 4
for several weeks (photo: Tim van Nus)

Eventual contents of box 4 (photo: Tim van Nus)

 

Banding - In 2010 all of the 12 chicks were given a small uniquely numbered light-weight aluminum band before their fledging. These bands allow for individual recognition of the birds. If these birds would ever be re-caught or found dead, a lot of knowledge would be gained about the dispersal and survival of TRES from Middleton. Christophe de Franceschi and Kyle Elliot, both well experienced bird banders, helped out with the banding of all of the 12 young on June 26.

Christophe (left) and Kyle (right) busy banding the chicks from box 8,
under the careful eyes of the two breeding birds (photo: Tim van Nus).

Banding in progress... (photo: Tim van Nus)

Chicks of box 8, at about 12 days old (photo: Tim van Nus)

In 2011 I did not visit Middleton due to other obligations, but prior to the summer I asked the USGS seabird volunteers to keep their eyes out on the TRES development. Thankfully they did and by the end of the summer I was pleased to receive an overview of the results of their nest box inspections. It appeared that seven boxes contained at least some nest material, of which four boxes contained completed nests on June 10. Three of these nests contained eggs on this day (box 2: five eggs; box 4: five eggs; box 6: two eggs). During a second inspection on June 16, box 2 and 4 still contained five eggs, box 6 contained six eggs and box 7 contained a completed nest but no eggs. On these dates some nest material was also found in box 1 ('few straws'), 9 ('few straws') and 12 ('cup with no bottom'), but these would not result in clutches, neither did box 7. No inspections were made during chick stage, but the boxes were all checked after chick departure and their contents, as observed by them on August 3, gives good information about the TRES productivity for 2011: box 2 contained an empty nest, indicating that five chicks had probably fledged from this box. Box 4 contained an unhatched egg and the comment that three chicks fledged from this box (which leaves the fate of one egg unclear). Box 6 contained two dead chicks and fledging for the other four was assumed.

Based on these data I suspect that in 2011 at least five TRES pairs occupied the boxes, of which three produced eggs, resulting in a total of 16 eggs, and these birds fledged at least 12 out of 14 chicks. Unfortunately, no chicks could be banded in 2011. This year would be the first year that an unhatched egg and dead chicks were found.

The 2012 season: a good start, but a near total failure... - This year the TRES activity got monitored by Mark Baran and Kyle Elliot, who both visited Middleton as part of their seabird research out there. Kyle informed me the first TRES was seen on May 12. The first eggs were laid probably just after Kyle left the island for his mid summer break, but throughout the summer Mark did an excellent job in recording the development in and around the nest boxes and even managed to band the birds this season. 

In 2012, five pairs laid eggs during the first two weeks of June. I assume the earliest laying date was June 2, as five eggs were found in box 8 on June 7 (the largest clutch on that date; they lay one egg a day, I believe). On this date box 1 contained one egg; box 6 contained three eggs, box 9 contained four eggs and box 17 also contained four eggs. During the following inspection on June 15 most nests were found to contain completed clutches: box 1: six eggs; box 6: six eggs; box 8: six eggs; box 17: seven eggs (the first seven-egg clutch recorded for this project!). Box 9, however, only contained three eggs and one broken egg. Mark noted that this box had been abandoned by the birds at this time.

During the inspection of June 22 not much changed, but box 1 lost an egg (broken) and went back to five, box 8 contained the first (two) chicks for the season and one egg disappeared form the seven-egg clutch from box 17, which would later be found broken. A little bit of nest material showed up in box 4, but this would not result in a finished nest and was probably not the work of another pair. On this date Mark managed to capture three of the breeding birds from their nest for banding (from boxes 1, 6 and 8). With still four pairs, 21 eggs and two chicks on the list it would not appear to be a bad season.

Unfortunately, during the following chick stage the situation would turn very bad. Box 1 contained four chicks and one egg on June 30, but these chicks were all found dead on July 7 and the remaining egg would not hatch and was later found broken. By June 30, box 6 contained five chicks and an egg. Of these chicks four were still alive by July 3 and these were banded, but all were found dead in their nest by the end of the season and the remaining egg apparently never hatched. In box 8 five chicks were present on June 29, and these were all banded on July 3. During the final inspection on August 11, however, two of these chicks were found dead, resulting in three fledged chicks that left their box during the second half of July. On June 27, box 17 was found to contain five chicks and one egg (which would later be found broken in the nest). Of these chicks only three were still alive on July 7 and these were banded, but only two managed to fledge during the second half of July; the three dead chicks were removed from the nest box during the final and cleanup round on August 11. Based on this information, in 2012, five pairs occupied the nest boxes and these laid a record number of 29 eggs, but these resulted in the fledging of only five chicks. Besides these breeding pairs, there may have been one or more lonely males occupying nest boxes for a while, but without success.

Contents of Box 17 on July 7, 2012 (photo: Mark Baran)

Contents of box 17 on July 7, 2012: three live chicks and two dead; only two would fledge (photo: Mark Baran)

Results - an overview (2009 - 2012):

Discussion

The first four years of this project have shown that, as throughout most of the North American continent, on Middleton Island the TRES can easily be persuaded to nest in nest boxes. Throughout Middleton’s recorded history, the TRES has shown to have a close relationship with human presence on the island, as the species appears to be totally dependent on human constructions for its nest sites. Without people’s activities, there most likely would not be any TRES nesting there. So, why would I put effort in establishing a TRES population on Middleton if the species would not nest there without the presence of man? Well, I could come up with several reasons:

- History has shown that there are TRES that apparently want to nest on Middleton. I'd like to help them doing so.
- As long as there is human activity on the island, I believe it’s good to do something positive for the bird communities by increasing the island’s breeding bird diversity. I expect the presence of a TRES breeding population to (thus far) be of a minimal influence on the other bird species present out there.

- Personally, after my first close encounter with this species in 2009, I believe they’re one of the most enjoyable species around. I was very pleased to see their playful flight and hear their friendly chattering calls above Middleton’s grassy plains, which otherwise can be a bit boring. I hope other people enjoy this too.

- Due to the island’s isolated location, size and the presence of an onsite weather station, Middleton Island would theoretically be a perfect place to study a TRES population. I am hoping that by having a descent number of these birds nesting there annually, monitoring these could make an interesting contribution to scientific research in general, in order to find out more about the island’s bird populations, as well learning more about the TRES itself.

- And finally, I don’t like bugs very much. The TRES eats loads of bugs, so as far as I’m concerned they’re welcome there.

But what about the TRES itself? Is life good for a TRES on the remote Middleton Island? At first sight, it seems that food appears to be widely available to them and with the absence of ground predators, only an irregular presence of almost only a single aerial predator species during the breeding season (the Peregrine Falcon, Van Nus, personal records) and the absence of competing species around their nest sites, a TRES can certainly have a good time out there. However, although the first two years of this study have shown that these birds can indeed have great breeding success, hatching a good number of eggs and fledging an equally good number of young, the results of this study’s third and especially its fourth year show that this is not always the case.
By now, the main cause of breeding failure seems obvious: with the exception of a some rainy periods during my stays on the island in 2009 and 2010, I suspect that continuing periods of bad weather that later field crews had to endure determined the relatively poor breeding results of the 2011 and 2012 seasons. After the almost disastrous 2012 season, both Mark Baran and Kyle Elliot informed me that bad weather conditions out there during the chick stage most likely were the cause of the poor breeding results in that year (and I can’t recall seeing a single sunny day when checking the FAA weather camera regularly throughout the 2012 summer). It seems obvious that eventually bad weather periods limited the foraging opportunities for the adult birds, resulting in starvation of many of the chicks. The relative large number of eggs that did not hatch in 2012 may also have been a result of the poor feeding conditions during an earlier stage of the breeding cycle. In 2012 the earliest egg laying date (June 2) was 7 days later than in 2010 (May 26), indicating that conditions were probably not very good at the start of the season. This is all not unique and apears to occur regularly among TRES populations elsewhere, i.e. Kyle informed me he noticed the same thing happening with the TRES nesting at the Long Point Bird Observatory in Canada and Alexandra Rose informed me she had a very bad 2012 season as well at Long Lake (almost exactly 100 miles northeast of Middleton), with massive nestling mortality. Unfortunately, only current weather data from the island appears to be available online and I can’t find a climate summery for any of these four summers anywhere, which would have been nice just to show you what the Middleton birds had to deal with during the first four years of this project.
But has, thus far, life been really so bad for them? Throughout the first four years of this study, the number of pairs laying eggs appeared to have gradually increased. Their eventual success on Middleton Island will probably depend on the overall climate conditions in summer. We can only hope for these birds that during future summers the climatic conditions will be better than they were in 2012, but they certainly will not always be.

Even with some very good breeding results and over the years still a good number of young fledging, compared to mainland colonies, the population increase that occurred during the first four years seems to be somewhat slow (I would love to compare this with data from the establishment years of other colonies, but I haven't managed to find this). Though the reason for the slow population growth has not been studied, I speculate that thus far the following scenario may have taken place:
Regardless of the bird’s breeding results in any previous year, the number of birds that nest on the island largely depended on a number of breeding birds from a previous year that managed to return to the island + an additional small number of ‘floaters’ (young, inexperienced birds) that shows up there in spring, usually late May. I do not know much about the annual survival of the breeding birds, but based on the records of the number of TRES that pass by the island in spring (before the establishment of this breeding population) I think that these could well account for the additional number of pairs that were found with each new year. I do not have many records of TRES observations in spring, simply because they were rarely seen, but because of the number of birds observed on these occasions (0-8, mostly 2-4, Van Nus, personal data), as well as the timing of the majority of these birds (last week of May) in relation to the initiation of nesting of the breeding pairs, I suspect that Middleton does not make part of the normal flyway for this species in spring and TRES would only show up there irregularly. With a number of birds arriving relatively early during the later three years of this project (May 10, May 12), I suspect that at least some of the experienced breeding birds are aware of their destination and these were deliberately heading to a Middleton Island nest box.
Whatever happens to the young that are born on the Middleton Island, thus far remains a mystery. As there appear to be no recorded observations for this species on Middleton in the fall, I suspect that Tree Swallows do not migrate above the Gulf of Alaska easily. I suspect it will not be easy for them to make their first fall migration, as they may have to cross several hundreds of kilometers of ocean. This makes me very uncertain about their survival, as well as their return rate to the island during a following year.

With the re-establishment of the TRES on Middleton, however, I am suspecting that with a number of these birds flying around there in spring, this could be a great attraction for floaters to the nest sites that are offered on the island, which otherwise may have been overlooked by these birds. Singing males will attract females, but possibly also other males. But again, these are all speculations, and not the result of a scientific study. The only way to get more insight in the way this remote population works is by continuing the TRES banding activities out there.

At this point the only thing I’m pretty sure about is that slowly more and more Tree Swallows are finding their way to Middleton Island and there are currently more pairs nesting out there than ever before.

The spring and fall migration routes that are used by Middleton's Tree Swallows remain unknown.
Do they take the shortest way? Do they fly hundreds of kilometers over the North Pacific Ocean?
Do the chicks all migrate south and find a watery grave?

The future of this project and priorities for further research 

How many TRES pairs Middleton Island eventually can host remains unknown. The number of Bank Swallows nesting there (65 - 112 pairs (2005 - 2010)) could indicate that swallows may be able to hold large populations out there, but as this is a different species, it will be hard to make a comparison. In fact, in 2009 the first two TRES pairs seemed to be doing very well, whereas during the same season the 65 pairs of Bank Swallows remained largely unproductive... (Van Nus, personal records)
With 17 nest boxes still present and only a maximum of about seven boxes occupied (including the non-breeding and lonely males) it seems that no new boxes need to be added yet. The people who kept their eyes on the TRES during the previous year informed me that the boxes are still in good shape and do not need to be replaced yet. As a side note: as mentioned before, the current boxes have been built by using the sizes recommended by the Golondrinas project. However, except for one. Throughout the first four years of the Middleton Island project, the swallows have occupied and laid eggs in seven of the 17 boxes (1 x in box 1; 2 x in box 2; 1 x in box 4; 4 x in box 6; 2 x in box 8; 1 x in box 9 and 1 x in box 17). According to this information, nest box 6 (shown on a photo above) has been the most popular box thus far, being used in every summer and producing more fledglings than any other box (16 fledglings in total). Box 6 was made about an inch less deep and the internal sizes are about half an inch bigger than the others. I suspect the birds prefer the dimensions of this box more than the others, which seems obvious for a secondary cavity nester, but it could all just have been a coincidence. In any way, these alternative sizes may be recommendable when adding an additional number of nest boxes in the future. 

The presence of the onsite weather station is a bonus for this project. However, it would be very useful when the gathered climatic data would actually be stored somewhere and when this would become available online (at least I'd never managed to find this).

Above all, it has become important that, in order to continue to generate good quality data, in future years a standardized method for monitoring the TRES population development on Middleton Island, in and outside of their nest boxes, gets applied. In order to get to know more about the structure of this population and the project’s future perspectives, during the upcoming years banding of the chicks and (re)capturing and banding of the adult breeding birds should be considered a priority. Middleton may be the ideal study site for a number of reasons, but this has yet to be proven.

Short video footage of the swallows of Middleton Island in 2010.

Acknowledgements

I thank all the people that were involved and that helped me out with the TRES-project thus far. Thank you for either helping building and placing some boxes or helping me conducting nest box checks or ringing chicks of the few pairs in 2009 or 2010 (Veronique Frochot, Christophe de Franceschi, Lauren Bessey, Joel White, Paul Solis and Kyle Elliot). During 2011, Lena Agdere, Mike Johns, Sharon van den Eertwegh, Chris, Lucy (sorry, couldn't find your last names) and Thomas Merkling all did a great job monitoring the boxes. In 2012 Mark Baran did a great job too, monitoring and banding the swallows during his rainy stay and he got thoroughly instructed by Kyle Elliot. I'd like to thank Scott Hatch (USGS) for letting me use the necessary power tools to construct 17 nest boxes from scratch and also for supplying the TRES bands. I thank Chris Gates for supplying me with the required information about starting up a project involving this fascinating species.

Literature

Rausch R. 1958. The occurrence and distribution of birds on Middleton Island, Alaska. The Condor 60: 227-241.