COMMON SWIFT (Apus apus) - Martinet noir
© Frank Jarvis
Summary
Present in Europe for only a few months in summer, Swifts are a common sight over fields and villages often in small flocks screaming their calls. But those calls hold some secrets:
One of the real sounds of summer for me is the wild screaming of flocks of Swifts as they chase each other around the skies. Like Swallows, House Martins and Black Redstarts, this is another species that has adapted it's cliff dwelling nature to human habitation, and can be found nesting most frequently beneath the roofs of tall, usually old buildings, like chateaux, church towers or large farm buildings - hence Switzerland is a great place for Swifts and their sound is a feature of many cities, towns and villages in the lowlands. It is found mostly below 1000m and the largest populations are usually found in urban areas.
However you've go to be quick, they are only here for a few short weeks - they first arrive from their African winter quarters in mid-April and begin to leave again by the end of July.
It is one of the fastest fliers of all our birds, feeding at great heights, taking some of the smallest of insects, and specialising on those which may have been caught in thermals and updraughts. They cover vast distances in the sky during the course of a day - packing up to 300 insects in a small throat pouch and dashing back to their nests to feed the young. Wherever it goes the Swift is always in a hurry to get there.
The secret of their speed lies in their long scimitar-shaped wings. The innermost bone (humerus) is a very short one allowing for a very distinctive twisting movement as they fly. This structure is similar in humming birds and kingfishers who are their closest relatives (but NOT the swallows and martins of similar behaviour - very confusing !).
DISPLAYS
Their aerial displays are wonderful, a flock of up to 10 or 20 individuals will chase screaming crazily at each other - no one is really sure what is going on here, mate searching is the obvious thing. It seems to happen about half an hour after their first feed of the morning, but mostly in an evening before going to roost. (These birds are so effective at flying that they will sleep on the wing). Their screams and, if you are close enough, the noise of their wings, are amazing.
Here is a rather longish file of 4 separate sequences of birds at dawn flying around and around the farm buildings in which they nest, if you listen carefully, in addition to the characteristic screams you can hear their fluttering wings and also the buzzing vibration of the feathers as they pass at high speed.
WHAT’S IN A SCREAM?
Let’s take a closer look at what is going in this very active scene. Here is just one of those screams as a bird approaches its farmhouse home and then turns away again, first you hear the call as it approaches and then the whoosh of wings as it passes the microphone:
Let’s focus in on that call. If you listen carefully the call is actually two sounds: the rising scream, a brief pause then a quick “tweet”. You can see this in the following video - play it 3-4 times until you get the hang of it because this bit is important:
From the above you can see the scream starting at about 4kHz, rising rapidly to 6.2 kHz, then more slowly to about 6.4 kHz. Then there is a small gap, only about 0.1 sec, before the final “tweet” note.
The end of the scream and the rapid tweets
BUT look carefully at that final note in the extract on the right. It appears in the sonogram as a series of vertical lines: it is a series of very short notes delivered incredibly rapidly. This is very significant because depending on how quickly those rapid tweets are delivered we can tell which is the male and which is the female.
THE TWEETS TELL A STORY….
Ansorge (2016) found that the male makes those terminal tweets very rapidly, each one delivered at about 17 ms intervals, whereas the female is about 50% slower at 28ms intervals ( a millisecond is a 1000th of a second, so”slower” is all relative you understand!). You may not be able to hear it (I am envious if you can !), but you can see this in in the next video where the male calls first and the female immediately after, the slower delivery of the terminal tweets showing clearly for the female:
You can read more about how this amazing difference was slowly uncovered in a very interesting series of papers by Bretagnolle (1993) (who identified the two parts of the call); Kaiser (1997); and Ansorge (2016). The latter two authors studied colonies in nest boxes, each with a clear panel at the back and, unable to tell them apart from their plumage, watched the female egg-laying, caught them, and marked them with a paint dot. They then made recordings of their calls with microphones in the nest box. Kaiser (1997) concluded that the male made a higher pitched call than the female. Whereas Ansorge (2016) concluded that the pitch depended upon the level of excitement of the bird, and the the sexual differences were in the trill part.
Reading those papers recently (2022), made me go back to my collection of swift recordings and examine them more closely. My recordings were all made with a parabolic microphone on free-flying birds, so the doppler effect (that funny frequency shift you hear from a train whistle) may have had some influence on the absolute frequencies I recorded, but the sexual differences, as shown above, are very clear.
The nest-box studies mentioned above, plus Lack (1956), all describe the battles that take place if another bird intrudes into the nest (see later for why this happens). The birds scream at the intruder to repel it, and if it actually enters then the fight can turn physical. A feature of these fights, and the aftermath, is that the pair will duet together, duetting probably serves to cement the pair bond and help defend the nest site. You can see from the two calls above that the female very rapidly responds to the male, typical of duetting. I found the birds also duetting on the wing as they flew around the farmhouse nesting colony, and this is probably one of the things going on in those screaming parties:
Duetting in free-flying Common Swifts
In that last sonogram the males do seem to call at a higher pitch than the females as suggested by Kaiser (1997). So using the trills to identify sex, I measured a sample of 33 male screams which ranged from 5.2-7.2 kHz, and a sample of 35 female screams ranging from 4.8-6.9 kHz. There was no significant difference between these two sets of numbers, and this would tend to agree with the conclusions of Ansorge (2016) that the pitch of the scream is more likely related to level of excitement. However due to possible doppler effects in my recordings as mentioned earlier, I say this with some caution.
SORTING OUT WHO IS WHO?
It is not possible for birders to distinguish male and female swifts from their plumage, and most likely it is the same for the birds themselves. So having an audible means of identifying sexes probably helps them avoid some dreadful errors during courtship and breeding. But the birds need pretty good hearing to pick out those millisecond difference when travelling at high speed at altitude. We do know that whilst birds hear on the same frequency range as humans, their temporal acuity (time-scale perception) is far better than ours, so swifts get much more information out of these brief calls than we can (Dooling and Prior 2016). See also the articles on Savi’s Warbler, European Wren, and Common Redstart.
Swifts will pair over several years if both survive their annual migrations. So how do they find each other back on their breeding grounds in Europe? From the tweets we now know how they can tell males from females, but how do you know it is your partner that is calling? The answer may lie in those terminal tweets in each call, there may be tiny variations in frequency as it falls off, or they may vary in number of individual tweets. The individual identity may also lie in the longer part of the call itself. You may have noticed that the scream part of the call is not a pure note, it has a sort of harsh or grating quality to it, and if you look at the spectrograms above you can see they are quite “fuzzy” and wavy. This “fuzzyness” is because the sound undergoes what is called amplitude modulation (“AM” for those who remember radio), I won’t pretend I fully understand this - just think of the signal vibrating. But how would a vibrating call contain information on individual identity?
The problem is that we know much more about sound production than sound detection in birds. Detection studies have been limited to behavioural experiments in the laboratory with captive species like Zebra Finch, Canary or Budgerigar. Dooling and Prior (2016) have shown that Zebra Finches have a temporal perception at least 4 times more sensitive than humans. Let’s see what this may imply for Swifts, here is one call from a male Swift at normal speed:
Single male scream at normal speed
Nothing startling in that one, it sounds pretty much like all the others. However if we slow it down to 25% of the normal speed - i.e. give ourselves a four-fold increase in time sensitivity, we might hear something like this (NB. As the speed is reduced the pitch is also reduced, so these next two will sound much lower):
Single male scream at quarter speed
We can now detect much more detail in the call. A four-fold difference in time sensitivity between humans and birds is the current best estimate (Dooling and Prior 2016), but it maybe greater than this. If birds were 10 times more sensitive than humans, which is not impossible since members of this family that live in deep caves in Asia are able to use echo-location for navigation in the dark, then maybe they hear something like this:
Single male scream at one-tenth speed
In this last piece you can begin to pick out all the granularity in the signal created by timing and amplitude modulation. Who knows what information could be coded into that? When you answer a phone call from a friend you can immediately recognise their voice, even though it comes through a microphone and a tiny earpiece. The subtle clues of pitch, speed, modulation, harmonics etc, all combine to give you enough information to identify the speaker. Birds must do something similar to recognise individuals by voice, to decide if the singer is a friend or foe, even though, to our lousy human ears, those individuals all sound identical. There is still much to learn on this topic.
POST SCRIPT:
Earlier I described the war that can break out when an intruder enters a nest-hole already occupied by a pair. This can happen if there is an unpaired bird looking for a soul-mate. Swifts live for between 8-20 years (Lack 1956) and only breed in their third year. They return in the year after they have hatched, but don’t start to breed. Instead they seem to hang around and watch what goes on, looking for potential nest sites and learning how to take care of their two precious eggs when the time comes. For their first time they may return to the exact nest where they were raised or the near vicinity, assessing things for the future. This is one of the theories put forward for the little understood phenomenon that Swift biologists call “banging”. This is when a bird may fly up to a nest box or next cavity and hit it with an audible “bang”. You can hear this activity here when the “banger” hits the woodwork near the nest cavity, often giving a “cheep” call:
There are various theories as to why this happens, a recent review by Olos (2017) listed 7, amongst which were: yearlings revisiting their old nest; adults looking for unoccupied sites; immature birds imitating adults; aggressive behaviour to claim as nesting spot; and Olos’s own observations which led him to offer predator avoidance as a purpose. Like individual voice identification this is another Swift attribute still to be fully understood.
Swift studies Frank Jarvis
