Purple martins are pouring in from all directions and falling from the sky like feathered meteorites. The swarm of martins looks like a giant feathered vortex as they spiral down at high speeds. Wave after wave of martins plunge into the small grove of trees in the middle of a city park surrounded by tall buildings, endless traffic, pedestrians, bright lights and noise. The martins have gone to bed for the night in a place that seems so unlikely for a good restful sleep and they are roosting there in the thousands. Now why would martins roost in such an area and in such huge numbers?

We know purple martins are social birds both in the breeding season and afterwards. Martins simply prefer the company of other martins. The colonial lifestyle of martins seems to be related to mating, safety in numbers, nesting territory, foraging for food, and socialization. Having groups of females in one location during the breeding season gives the males ample opportunities to engage in multiple copulations thereby passing on their genes in many offspring. For the females there are plenty of choices for mates. But this breeding factor is not relevant for post nesting behavior when thousands of martins from many different colonies form huge premigratory roosting flocks. There is no longer a need to attract a mate and raise a family.

So you now see after the breeding season, thousands and thousands of largely non-related purple martins from multiple colonies dropping from skies and roosting in a restricted location. Now why would purple martins form this huge premigratory roost? Why not small groups scattered all over the place? Why would thousands of purple martins gather in a small area of a few trees, a reed bed, an island in the middle of a lake, or in a city park, industrial setting or around other human constructed facilities? Is it just for the company? Do they like the lights, sounds, warmth, and openness around manmade attractions? Or do they do it because of their natural enemies?

First, there is an inherent socialization factor deeply embedded in the martins’ DNA. Martins are members of the swallow family and like most swallows martins are simply attracted to one another. They are genetically programmed to flock together. This flocking behavior seems to become much stronger AFTER the breeding season and there is no longer a need to exert territorial rights and attract a mate. A premigratory martin roost is like a giant biological magnet that pulls more and more martins to a central location. It just never stops growing as martins pour in from all directions. The huge roost becomes “ground zero” for all martins in the area and they are drawn there from miles around. So, the old cliché, “birds of a feather flock together”, is clearly evident in martin premigratory roosting behavior.

Humans may have facilitated and unknowingly “encouraged” the formation of giant premigratory martins roosts by creating an “attractive” environment. For those martins choosing manmade structures for roosting, perhaps the lights, sounds, open area and maybe even warmth could be important factors in attracting martins to a central location. Nocturnal migrating birds are attracted to light. Martins are birds of the open skies, so any “cleared” location with available perching structures may be attractive. Such a location is easy to approach and fly to for roosting. And martins roost in huge numbers at a petrochemical refinery plant in Brazil with its hot pipes and loud hissing steam. Maybe the martins like the heat generated in such a place? Plus such locations may not be as “predator prone” as those in the “wild”. Before the advent of city parks, industrial plants, bridges and other manmade structures, martins may have roosted in reed beds, on “islands” in large lakes, on coastal barrier islands and in “isolated pockets” of vegetation located away from heavy forests.

And finally, there is the predation and safety in numbers factor that exerts an enormous impact on highly social species such as purple martins. This is an “ancient” factor that martins developed over thousands and thousands years of evolution and association with their natural predators. This was before the era of bright lights and sounds of the human cities. There is safety in numbers in the natural world. More eyes are watching for danger. This factor may indeed be the key determinant in creating those huge premigratory purple martin roosts even though predation may not be as prevalent around manmade structures. However, martins are still responding to their evolutionary past that is stored in their genes and expressed in their behavior. Nothing terrorizes a prey animal like its most fearful predator.

But there is another more intriguing mathematical angle to the predation and safety in numbers factor and its relationship to those giant martin roosts. It is essentially a lottery where both predator and prey benefit in the long run. A numbers game based on odds. I think of it as a paradoxical system where winning is also losing but on an insignificant scale for mainly the martins and in the big picture, both martins and their predators hit the lottery. I call it “winning the predation lottery”. Sounds like a bunch of double talk, but there is logic to this seeming madness. Here it is…

To illustrate “winning the predation lottery” phenomenon, lets change purple martin premigratory roosting behavior. For the sake of discussion, suppose purple martins formed small premigratory roosts that were distributed over a wide area of many miles. These could be flocks of ten, 30 or maybe 50 birds. There might be 50,000 martins around, but they are disbursed over a vast area in small groups. Now you say to yourself, this makes sense. The martins are less noticed by predators because of smaller numbers (at least that is what you think) and are not getting into squabbles with humans over all that smelly mess the martins create. Plus a natural phenomenon like a storm would not adversely impact large numbers of martins in one location. Surely this is a better way of roosting and keeping out trouble with natural predators, humans and the weather. Or is it?

It really may not be from a predation perspective for the following reasons. Most martin predators are territorial and scattered over a vast area. There are always many fewer predators than their prey. A Cooper’s hawk or sharp-shinned hawk’s territory may cover a square mile or more. So if martins were dispersed many miles in small groups, then each roosting flock could become susceptible to different predator attacks. A Cooper’s hawk here, a sharp-shinned hawk there, a barred owl, a great horned owl and on and on. So in reality martins in general could be exposed to more incidents of predation from a variety of hawks and owls. Small flocks of martins located great distances from each other could be targets of predation by multiple hunters. So premigratory roosting behavior predicated on large numbers of small flocks dispersed over a wide territory may expose martins as a whole to more rather than less predation. Does this make biological sense?

By forming a huge premigratory roost in one small area of a few trees or on a manmade structure, the martins would most likely be exposed to fewer numbers of predators because of the predators’ territorial drives. Only those Accipiter hawks and owls in that territorial domain would probably be involved in active predation. Of course, larger numbers of martins in a smaller location would no doubt attract any predator in a given area and increase those attacks, but again there may be fewer predators at work due to territorial influences. See how martins are winning the predation lottery with those huge premigratory roosts. Maybe it does make biological sense.

But there is more to the numbers game. Lets get statistically specific and “play the odds”. If there are ten martins at a roost and a sharp-shinned hawk attacks, then all things being equal, there is a one chance in ten that one of the martins may be caught. Each martin has a one chance in ten of being killed and that is still not good survival odds over the long run of multiple attacks. The hawk has only ten possible prey items and they are doing everything they can to escape. Not great odds for either the martins or hawk.

Now look how things statistically change for both martins and their predators with a giant premigratory roost in a small area. If there are 50,000 martins flying in to roost, and an Accipiter hawk attacks, then the chances are astronomical that any given INDIVIDUAL martin will be taken, though the odds of the hawk succeeding are much greater because of more opportunities for success. Remember: these huge martin roosts would contain young and old, fit and less fit, and that provides any predator with more possibilities. This is MUCH better for both the martins and the hawk than a one chance in ten that was previously discussed. The INDIVIDUAL martin becomes almost indistinguishable because of the large numbers of martins--one among 50,000. Call it blending in with the masses so you no longer stand out as an individual--self-survival among the hordes. Maybe martins know something about “playing and winning the predation lottery” after all!

So for the martins, it may be more adaptive to concentrate in huge numbers in a small area to significantly reduce the chance that an INDIVIDUAL martin will fall prey to an attacking predator while at the same time providing a predator with better odds of securing food--seemingly paradoxical system that does make biological sense. Plus the territorial drives of the predators may reduce the diversity and numbers of attackers in a restricted location. It is a well-designed system that has evolved over thousands and thousands of years where both prey and predator benefit. The predator gets a meal and the vast majority of prey escape. Call it sacrifice of the few to benefit the many. Though martins are highly colonial, they still have individual survival mechanisms in place and the more martins present, the less chance a predator will catch martin “x”. Martin individual vulnerability becomes almost a non-issue because of the huge number of martins.

Purple martins are similar to other social species that form biologically non-related groups. Though martins are colonial and form breeding flocks in dense concentrations, they still retain their individuality for survival. The huge premigratory martin roosts may be formed based on a combination of factors such as inherent socialization influences (genetically programmed to flock), artificial manmade variables like light, sound, warmth and open space, and perhaps most important the predation component. There is safety in numbers in the natural world and a huge flock of martins roosting in a small confined area provides “protection” for the group as a whole and largely immunizes the individual from predation. It is astronomically better for the individual martins to be a member of a flock of 50,000 than of ten when a predator attacks. Much better odds of survival. This system has been in existence for perhaps thousands of years and provides benefits for both martins and their predators. Both martins and their predators know and play the “game” well and they are all winners in long run.