A female cuckoo (left) watching reed warblers in the reeds below ignores a displaying male who has landed beside her.
Wicken Fen, 9 June 2014.
Dragging a rope over the Derbyshire moors during the Easter holidays is bound to attract attention. Mike and I have been searching all day for meadow pipit nests. He holds one end of the rope, I the other, and we walk along slowly, hoping to flush sitting pipits. The rope floats lightly over the ground, like a gentle brush, and doesn’t damage the vegetation or the pipit nests, which are tucked away in little hollows under tufts of grass. Even when we’re sure of the exact position from where the bird flew, it’s often hard to locate the nest as it is hidden so beautifully, and the dark brown eggs are well camouflaged in the shadows. We’ve found 23 nests today, a good day’s work. To each clutch we’ve added a model cuckoo egg, and we’ll return tomorrow to check whether the pipits have accepted them. But our nest searching has been reported to the local police, and now a policeman, in blue uniform and tall helmet, is striding over the moors towards us.
We explain that we’re not collecting eggs, rather we are putting extra eggs into birds’ nests. Not surprisingly, the policeman finds this hard to believe. Our routine of playing cuckoo now seems so normal and sensible to us, that we forget how ridiculous it must seem to someone who doesn’t know why we are doing something so odd. But we describe our experiments, show him our model eggs and research permit, and he is reassured. Having checked how to spell ‘cuckoo’ for his record of this strange encounter, he bids us good luck and sets off back to his village beat.
In Britain, the meadow pipit is the cuckoo’s favourite host in moorland. This race of cuckoo, which is the predominant one on the moors of western and northern England, in Wales and throughout Scotland, lays a brown speckled egg which matches the dark brown eggs of the pipits. Our experiments showed, just as for reed warblers in the fens, that egg mimicry was important for fooling these hosts. The pipits were more likely to reject badly matching model eggs, for example those with a paler, greyish-white background, typical of the cuckoo race that specialises on pied wagtails, or plain blue eggs, typical of cuckoos that specialise on redstarts. We often found the rejected eggs lying in the grass, just a short distance from the nest.
We also put our model cuckoo eggs into pied wagtail nests, another favourite cuckoo host in Britain, which nests in crevices in banks, rocks and walls. The greyish-white speckled eggs of its cuckoo race are a good mimic of the wagtails’ eggs. Once again, we found that the mimicry was important, because the wagtails tended to reject model eggs unlike their own. Experiments with a number of other favourite hosts of the common cuckoo, for each of which there is a specialist race with a mimetic egg, also show that these hosts are particular about egg appearance and will reject badly matching eggs. Clearly then, host rejection has selected for cuckoo specialisation. The result has been the evolution of distinct races of cuckoo, each with an egg type that tends to match the eggs of its particular host species.
There is, however, a glaring exception that proves the rule – prove in the original sense of the word, namely probing or testing an idea. A favourite cuckoo host in woodland, parkland and hedgerows is the dunnock, formerly known as the hedge sparrow. Dunnocks build their nests in bushes, hedges and low trees, usually well concealed in dense vegetation. Cuckoo eggs in dunnock nests have a greyish-white background and reddish-brown spots, quite unlike the beautiful, plain turquoise-blue eggs of the dunnock. As Gilbert White remarked over 200 years ago in The Natural History of Selborne:
you wonder, with good reason, that the hedge sparrows can be induced at all to sit on the egg of the cuckoo without being scandalised at the vast disproportioned size of the suppositious egg; but the brute creation, I suppose, have very little idea of size, colour or number.
No one has yet tracked cuckoos to test whether some individuals really do specialise on dunnocks, but it is likely that this is indeed a separate race of cuckoo, because DNA studies show it is genetically different from the other races that specialise on meadow pipits and reed warblers. The most extensive study of cuckoos in farmland and woodland in lowland Britain is by John Owen, who recorded an incredible 509 parasitised nests in the Felsted district, Essex, between 1912 and 1933. Of these, the dunnock was easily the most frequent victim, accounting for 302 of the parasitised nests, so it seems likely that many of the cuckoos specialised on this host. Furthermore, although the cuckoo egg in dunnock nests is clearly not mimetic – it is the wrong colour and has spots – nevertheless the egg is of a distinct type, intermediate in darkness between those of the pipit and wagtail races of cuckoo.
Why is the dunnock race of cuckoo unique among the European cuckoo races in showing no mimicry of the host eggs? The answer is clear. It’s because dunnocks are the only major host species to show no egg discrimination. Our experiments show that they will accept model eggs of any colour or pattern. We wondered if dunnocks simply had poor colour vision, or found it difficult to discern egg colour in the dense cover where they build their nests. Perhaps, in a dark nest, the dunnock’s own blue eggs and the various model eggs all appeared a similar shade of grey? Therefore we tested them with white or black model eggs, which were clearly different in shade, but they accepted these too. Dunnocks even accepted whole clutches of model eggs unlike their own, so their acceptance of a single model egg was not simply because they regarded it as a harmless lump of resin.
Our results with the dunnock, the odd one out among the cuckoo’s favourite hosts in Europe, are illuminating. They show that a cuckoo race evolves a mimetic egg only when its host is discriminating. If dunnocks ever did begin to reject badly matching eggs, then we can be sure that their cuckoos could respond by evolving egg mimicry. There is a cuckoo race in Finland that specialises on redstart hosts, which have pure blue eggs just like those of the dunnock. This cuckoo race lays a pure blue egg, a perfect match.
We are still left with the problem of why the dunnock race of cuckoo lays a distinctive egg. Many woodland and farmland birds have pale mottled eggs. If dunnock-specialist cuckoos sometimes parasitised other species, which were discriminating, then they may have evolved a generalised match for their secondary hosts.
So far, we have assessed the match between cuckoo and host eggs by human vision. But birds have a different visual system to ours. We have three colour cones in our retinas, attuned to different wavelengths of light: long wavelengths, which we see as red; medium wavelengths, green; and short wavelengths, blue. Birds have a fourth cone type, which enables them to detect still shorter wavelengths, ultraviolet, or UV. So the blue on the crown of a male blue tit, or on the throat of a male bluethroat, will appear even more dazzling to a bird’s eye than it does to our own. Experiments have shown that this UV component is important in mate choice. If UV-blocking cream is applied to the crowns of male blue tits, or to the throats of male bluethroats, then females are more reluctant to pair with them than with males who have been treated in the same way but with a control cream that does not affect UV reflectance. To human eyes, males treated with both types of cream look exactly the same. But birds can tell the difference.
Cassie Stoddard and Martin Stevens, my colleagues at Cambridge, reassessed the match between cuckoo and host eggs for various cuckoo races in Europe. First, they measured the colour match for both background colour and spot colour across the full range of wavelengths of light as perceived by a bird’s eye. They found that cuckoo eggs matched their host eggs not only in the wavelengths visible to humans (blue to red), but also in the UV, visible only to the birds themselves. Then they measured the match in spot sizes between cuckoo and host eggs. These measurements enabled them to quantify the match between cuckoo and host egg, as perceived through a bird’s eye.
They found that the perfection of egg mimicry in the various races of cuckoo was related to how particular their respective hosts were, as revealed by our own experiments with the model eggs. The best matches were for cuckoos with the most discriminating hosts, for example great reed warblers and bramblings. These cuckoo races have evolved wonderful mimicry of their host’s eggs, with background colour, spot colour and spot sizes all copied to perfection. Moderate matches occurred where the hosts were a little less discriminating, for example reed warblers and meadow pipits. And there was no match at all when the host showed no discrimination against odd eggs, as for dunnock hosts.
Claire Spottiswoode, another Cambridge colleague, tested whether cuckoo eggs also evolve thicker egg shells in response to more strongly rejecting hosts. On average, the thickness of a cuckoo’s egg shell is about a tenth of a millimetre, but their shells vary in thickness by eight-hundredths of a millimetre. This sounds trivial, but it translates to over twofold variation in breaking strength. Have thicker shells evolved to resist puncturing by more discriminating hosts? Claire compared the races of cuckoo in Britain by examining eggs in museum collections. She found that cuckoo eggs in the nests of discriminating hosts, such as reed warblers, meadow pipits or pied wagtails, had much thicker shells than those in the nests of undiscriminating dunnocks.
So in response to increasing host egg rejection, races of cuckoos have evolved eggs that are both harder to detect and harder to reject.
The arms-race analogy assumes that this wonderful cuckoo trickery has evolved gradually over the generations as hosts have gradually improved their defences. This kind of never-ending arms race has been called ‘Red Queen’ evolution, after the Red Queen in Lewis Carroll’s Through the Looking Glass. In this tale, the Red Queen grabs Alice by the hand and they run together, faster and faster. To Alice’s surprise, they never seem to move but remain in the same spot. ‘In our country,’ says Alice, ‘you’d generally get to somewhere else if you ran very fast for a long time.’ The Queen replies: ‘A slow sort of country! Here it takes all the running you can do to keep in the same place.’
Are cuckoos and hosts engaged in such an arms race, each evolving tricks and counter-tricks to keep pace with improvements in the rival party?
Here’s a fascinating ‘thought experiment’, one that we could dream of performing. If we could go back across the generations with our model eggs and test ancient populations of reed warblers, before they were ever subjected to cuckoo parasitism, then, according to the arms-race theory, we would expect them to have poorer defences and so to be less likely to reject eggs that differed in appearance from their own. Furthermore, if each party really had been evolving simply to keep pace with improvements in the other party, as in Red Queen evolution, then we might imagine that past cuckoos, with poor egg mimicry, might have done just as well against their past hosts, with poor egg discrimination, as current cuckoos are faring against current hosts. So as hosts improved their defences, cuckoos improved their trickery, with the result that their relative success might remain the same. As a final part of our thought experiment, we would predict that today’s hosts would find the eggs of past cuckoo generations easier to detect than those of current cuckoos, because past cuckoo generations would have poorer egg mimicry.
Amazingly, this exact experiment has been done, not with cuckoos and their hosts, but with another pair of enemies, a bacterial parasite called Pasteuria ramosa and its host, the water flea Daphnia magna. This may seem a far cry from our cuckoo study, but the battle between these tiny creatures leads to a similar evolutionary arms race. Water fleas are small crustaceans, just a millimetre or two in length, that live in ponds. They feed by passing water through their mouths and filtering out tiny food particles. Sometimes they ingest harmful bacteria, such as Pasteuria, which attach to the gut wall and then spread through the water flea’s body cavity, eventually causing it to be sterile. So infection by these bacteria reduces the water flea’s reproductive success, just as acceptance of a cuckoo egg reduces the host’s reproductive success.
Some genetic strains of water flea can resist the bacteria by chemical defence, which prevents them from attaching to the gut wall (a direct analogy of hosts rejecting cuckoo eggs). This then favours the evolution of bacteria that can bypass this defence, for example by mimicking the chemicals of food particles (a direct analogy of cuckoo eggs escaping detection by mimicry of host eggs). Different genetic strains of water fleas then evolve, with new chemical defences that see through this disguise (better detection), leading to new bacterial tricks (better mimicry), and so the arms race continues with further bacterial tricks and water flea defences.
Now for the wonderfully neat study that enabled a test of Red Queen evolution in this system. Daphnia eggs and Pasteuria spores can survive in a dormant state for a very long time, perhaps to enable them to survive through long periods of drought. These accumulate in pond sediments and provide a ‘living fossil record’ of past generations. Ellen Decaestecker and her colleagues from the Catholic University of Leuven, Belgium, sampled sediment cores from a pond which contained these resting stages over a 39-year period, which entails hundreds of generations of water fleas and bacteria. Each depth of sediment represents a ‘snapshot’ in the arms race, a historical record of the water flea and bacteria populations at that time. The researchers reactivated the water flea eggs and bacterial spores from across the different generations by placing them in warm temperatures and a summertime daylight regime. Water fleas could now be exposed to bacteria from the same sediment layer (contemporary bacteria) and from older sediment layers (past bacterial populations). So here is the exact equivalent of our dream experiment: comparing how hosts fare against current and past cuckoo populations.
The results showed that water fleas were less susceptible to infection by past generations of bacteria. Therefore water fleas evolved to beat old bacterial tricks, and current bacteria, in turn, evolved new tricks better able to resist host defences. However, when bacteria were pitted against their contemporary water flea populations from the same generation, there was no change in infection rates across the 39 years. So the relative success of the two parties had remained the same over time. They had indeed been ‘running to stay in the same place’.
Mike and I could only dream of doing such an experiment with our cuckoos and hosts. However, although we couldn’t go back in time to test if old host populations were less discriminating, we could do an equivalent experiment. Some songbirds are unsuitable as hosts of the cuckoo. There are two groups of such birds. The first group comprises species that feed their young on seeds, for example some finches. These are unsuitable as hosts because the young cuckoo needs an invertebrate diet for successful rearing. The second group has a suitable diet but nests in small holes, inaccessible to the laying female cuckoo. This group includes the tits, pied flycatchers, wheatears, starlings, house sparrows and swifts. If discrimination against foreign eggs evolves specifically in response to cuckoo parasitism, then we can predict that these unsuitable hosts will not show any egg rejection, because they have no history of interacting with cuckoos.
So Mike and I spent two summers cycling around the British countryside looking for nests of these unsuitable host species, and placing model eggs in their nests that were different from their own eggs in appearance. Our colleagues from the University of Trondheim in Norway, Arne Moksnes, Eivin Røskaft and Bård Stokke, also played the part of the cuckoo and tested unsuitable hosts with model eggs in Norway. The results of both sets of experiments revealed that these species accepted most, if not all, of the model eggs we gave them.
Some comparisons between the responses of closely related species are particularly revealing. The spotted flycatcher, whose open nest is accessible to female cuckoos, shows strong rejection of eggs unlike its own, whereas the pied flycatcher, which nests in holes and is inaccessible, shows no rejection at all. In the finch family, the two species that feed their young predominantly on invertebrates, and are therefore suitable as hosts, namely the brambling and chaffinch, show strong egg rejection, while the four species that feed their young mostly on seeds (greenfinch, linnet, redpoll, bullfinch) show little or no rejection. These comparisons suggest that rejection of eggs is not simply determined by taxonomy. Rather, it evolves only when a species becomes exploited by cuckoos.
These results show that dunnocks behave just like those species that have never had an arms race with cuckoos. Could dunnocks be recent victims, which have simply not had sufficient time to evolve egg discrimination? At first, it seems that references from old literature would argue against this view. We have already seen that over 200 years ago Gilbert White mentioned the dunnock (then called the hedge sparrow) as a host in The Natural History of Selborne. We can go back further to Shakespeare’s King Lear, written in about 1605, where the Fool warns Lear that his selfish daughters will ruin him if he continues to dote on them, just as:
The hedge sparrow fed the cuckoo so long,
That it had it head bit off by it young.
Shakespeare intended this metaphorically, but there is one curious account where this actually happened. The cuckoo chick normally swallows its meal only when the delivery is complete, so it avoids injuring the foster parent as it bows deep into the chick’s gape. An exception to this usually smooth operation involved a cuckoo nestling clamping its mouth too soon on the head of a hapless dunnock, which suffered fatal injury.
There is an even more ancient reference to dunnocks and cuckoos in Chaucer’s poem of 1382, The Parlement of Foules, where Merlin chastises the cuckoo:
thow mordrer of the heysugge on the braunche that broghte thee forth!
Heysugge is Middle English for ‘hedge sparrow’, and probably refers to the bird we now call the dunnock.
If the dunnock has been parasitised for at least 600 years in Britain, would we expect it to have evolved egg rejection by now? The answer is that it depends on the rate of parasitism by cuckoos, which will determine the selective advantage of egg rejection. For the past 70 years, thousands of bird watchers have completed nest record cards for the scheme administered by the British Trust for Ornithology (BTO). Largely because of their efforts we have a good picture of cuckoo parasitism rates. During the period 1939 to 1982, before the recent crash in the cuckoo population, the parasitism rates over the whole of Britain for the three main hosts were:
5 per cent for reed warblers (out of 6,927 nests recorded)
3 per cent for meadow pipits (out of 5,331 nests)
2 per cent for dunnocks (out of 23,352 nests)
These figures suggest that dunnocks now suffer similar pressure from cuckoo parasitism as reed warblers and meadow pipits, two hosts which have evolved egg rejection. Surely, then, dunnocks should also reject cuckoo eggs.
We can calculate how long it would take a dunnock population to evolve egg rejection, with just 2 per cent parasitism by cuckoos. The answer is several thousand generations. The reason it takes so long is that 98 per cent of nests are not parasitised, so in the vast majority of cases an inability to reject an odd egg is not penalised. Only in 2 per cent of the nests would ‘rejector’ dunnocks gain an advantage. Several thousand generations is just a blink of an eye on an evolutionary timescale, and many traits we see will have taken at least this long to evolve. Nevertheless, there have been profound changes in the countryside during the past few thousand years. Most of Britain was covered in woodland until a few thousand years ago, and dunnocks are not common in this habitat. Perhaps they did not become a favourite victim of the cuckoo until extensive forest clearance occurred 6,500 to 2,500 years ago, creating the woodland edges and hedgerows where dunnock populations are most dense. So it’s possible that dunnocks are, indeed, relatively recent hosts that simply have not yet had time to evolve defences against cuckoos.
What have we learned from playing cuckoo with our model eggs? Our results enable us to reconstruct the likely stages of the egg arms race between cuckoos and hosts:
(1) At the start of the arms race, hosts accept eggs unlike their own (species with no history of cuckoo parasitism, the unsuitable hosts, exhibit no egg rejection). And cuckoos have no egg mimicry (in the absence of host rejection, cuckoos do not produce mimetic eggs, as shown by the cuckoo race that specialises on dunnocks).
(2) Once the cuckoo begins to parasitise a host species, any hosts that reject cuckoo eggs will raise more of their own offspring, compared with those that accept and get lumbered with raising cuckoos instead of their own young. If offspring inherit their parents’ behaviour, the rejection habit will increase in the host population by natural selection. Therefore, in response to cuckoo parasitism, hosts evolve egg rejection (most cuckoo hosts reject eggs unlike their own).
(3) As hosts begin to reject eggs unlike their own, those cuckoo eggs which, by chance, resemble the host eggs more closely will be more likely to escape host detection and produce young cuckoos. If daughter cuckoos inherit their mother’s egg type, better-matching cuckoo eggs will now increase in the cuckoo population by natural selection. Therefore, as hosts evolve egg rejection, cuckoos evolve egg mimicry (cuckoo races show egg mimicry when their hosts reject non-mimetic eggs).
(4) As a host evolves better egg discrimination, its cuckoo race evolves better egg mimicry (where hosts are more discriminating, the cuckoo’s egg mimicry is better).
So both parties have evolved in response to the other, a case of co-evolution.
But egg mimicry by cuckoos and egg rejection by hosts is only part of the egg arms race. Natural selection has another game to play with egg markings, involving signatures and forgeries. Our field experiments have only just begun.