A Route of Evanescence,
With a Revolving Wheel −
A Resonance of Emerald
A Rush of Cochineal −
And every Blossom on the Bush −
Adjusts its tumbled Head −
The Mail from Tunis, probably,
An easy Morning’s Ride –
Writing in 1880, the American poet Emily Dickinson encapsulates the essence of the Ruby-throated Hummingbird – speed, iridescence, and a unique relationship with flowers – in her minimalist poem A Route of Evanescence. In the last two lines, she ponders on the hummingbird’s overall physicality and imagines it flying with ease from some foreign shore, implying that the bird is completely in harmony with nature and serenely confident of its powers. Little could Dickinson have imagined that science would one day support her poetic musings and confirm that hummingbirds did indeed arrive from some distant land. However, this was not from Tunis after ‘an easy morning’s ride’, but from Eurasia following a dispersal that took many millions of years.
Hummingbird origins
Despite their very different lifestyles, the hummingbirds (Trochilidae), true swifts (Apodidae) and treeswifts (Hemiprocnidae) are closely related families that are placed within a single combined group, the Pan-Apodiformes. As we have seen, their nearest relatives are the owlet-nightjars, nightjars, potoos, frogmouths and the Oilbird (Figure 11.2). Evidence that hummingbirds and swifts have a common ancestry has come from a well-preserved fossil, Eocypselus rowei, from the early Eocene Green River Formation of southwestern Wyoming. In 2013, Daniel Ksepka, while working at North Carolina State University, found the 50.6-million-year-old specimen after it had been overlooked for several years, and named it rowei after John Rowe, Chairman of the Field Museum’s Board of Trustees.1 It is an exceptionally well-preserved fossil that includes most of the skeleton, as well as many complete feathers with their melanosomes, tiny cellular organelles that contain pigment. Eocypselus rowei was a small bird, about 12 centimetres long and weighing less than 30 grams, with a swift-like beak, long legs, and a wing structure intermediate between that of hummingbirds and swifts. A feathered head-crest may have been present, as in extant treeswifts, and its non-specialised glossy black wings lacked the necessary modifications for either soaring or hovering. Overall, the fossil’s morphological features indicate that the common ancestor of hummingbirds and swifts was already small-bodied before each family evolved its characteristic flight behaviour. The swift’s very short legs, therefore, must have developed after the lineage’s divergence, possibly to reduce weight and enable a highly aerial lifestyle. Indeed, Common Swifts are one of the fastest-flying birds and can spend up to 10 months continuously airborne, taking ‘power naps’, capturing food, obtaining nest material and even mating on the wing.2
The next fossil in the ascent of hummingbirds, Parargornis messelensis, was recovered from the Messel Pit in Germany, a site that was a steep-sided volcanic lake surrounded by subtropical rainforest during the Eocene. The area is now a UNESCO World Heritage Site, one of only a few that has ever been listed exclusively for its fossil assemblage. Periodically, the vast caldera released clouds of toxic fumes that poisoned scores of creatures in the surrounding area, including Parargornis as it flew across the lake’s noxious waters 47 million years ago. A stem hummingbird, Parargornis had a swift-like beak, short wings and a long tail, with feathers resembling those of the owlet-nightjars. Gerald Mayr, the German palaeontologist who described the fossil in 2003, believes that its beak shows that the Trochilidae evolved from insectivorous ancestors and that its owlet-nightjar-like feathering may well be a primitive trait of early hummingbirds.3 Also, the fossil’s peculiar wing structure – a combination of a short humerus and broad wing – has no counterpart among modern birds and reflects an early stage in the evolution of hovering flight.4
In the late twentieth century, the Russian palaeontologist Alexandr Karhu described two incomplete fossils from 35-million-year-old deposits of the northern Caucasus. Both specimens are now known to belong to the stem lineage of modern hummingbirds, since they share several characteristic features, including a modified ‘elbow’ joint and a humerus head that allows rotation of the wing during hovering flight.5 Several years later, a more convincing hummingbird fossil was found by Mayr after he had searched through the collection drawers of Stuttgart Natural History Museum. Hidden among the many specimens, he noticed two tiny unclassified bird skeletons unearthed from the Lower Oligocene deposits in southern Germany. Remarkably, both fossils possess a combination of features unique to modern hummingbirds: small size, a short humerus adapted for hovering flight, and a long bill. The 30-million-year-old species was named Eurotrochilus inexpectatus, the ‘unexpected European hummingbird’, as it had been assumed that no modern-type hummingbirds ever existed outside the Americas. Despite the similarities to extant hummingbirds, Eurotrochilus inexpectatus still expresses some primitive features, suggesting that it is not a particularly close relative of any living species.6 Three years later, another Eurotrochilus species, this time from the Luberon National Park in France, was reported by Antoine Louchart after he spotted the fossil in a private collection.7 This Eurotrochilus skeleton is the most complete ever found and includes the skull and bill, while the surrounding rock matrix reveals the outlines of its wing and tail feathers. Collectively, these Eurasian fossils show that hummingbirds had a much wider distribution in the past, and raise the interesting questions of how hummingbirds reached the New World and why they became extinct elsewhere.
A recent molecular phylogeny has helped provide some of the answers.8 For more than a decade, Jimmy McGuire, an evolutionary biologist at the University of California, collected DNA samples from most hummingbird taxa, as well as DNA from closely related species, including nightjars, swifts and an owlet-nightjar. Working with colleagues from the USA and Canada, McGuire sequenced six genes, four nuclear and two mitochondrial, from each species and constructed a time-calibrated phylogeny, based solely on nucleotide substitution rates. The results suggested the following scenario. Crown hummingbirds split from the swift lineage at least 48 million years ago, a date that is in general agreement with the fossil record. This divergence probably took place in Europe or Asia, given the presence of early fossil hummingbirds from several sites in Europe, and the phylogenetic diversity of swifts and treeswifts in the region. Twenty million years later, the common ancestor of modern hummingbirds reached South America by dispersing across the Bering Strait to Alaska and North America. A transatlantic route is unlikely, since hummingbirds are metabolically constrained from undertaking long overseas journeys. Why these early hummingbirds left no survivors in Eurasia and North America is unclear, but it may relate to both climatic changes and the arrival of passerine nectar specialists. It is also possible that the species-rich fauna of herbivores in the Old World added to the pressures for the limited availability of energy-rich and nutritious flowers. Once hummingbirds reached South America, around 22.4 million years ago, the founding population dramatically radiated into new ecological niches to produce the nine distinct lineages recognised today: topazes, hermits, mangoes, brilliants, coquettes, the Giant Hummingbird, mountain gems, bees and emeralds.
Speciation was especially fast in the Andes, since, although the mountains represent just 7 per cent of the land area occupied by hummingbirds, they are home to 40 per cent of the species. It seems likely that the Andean orogeny contributed directly to their dramatic diversification, since speciation was greatest when the Andes were rapidly increasing in height. But mountain environments are cold at night, and if hummingbirds stopped feeding they would cool too rapidly. Rather than consume energy trying to keep warm, high-altitude species have evolved the ability to reduce their metabolic rate by as much as 95 per cent and enter a sleep-like state known as torpor. By doing so, species such as Andean hillstars (genus Oreotrochilus) consume up to 50 times less energy, and reduce their core temperature to a level that is barely sufficient to maintain life.
Ten million years ago, a drought-tolerant ancestor of the mountain gem and bee clades recolonised North America, which at that time was still separated from South America by the Central American seaway. The accumulation of species in North America was slow at first, but then rapidly increased owing to multiple invasions of emeralds, coquettes, mangoes and hermits once the Panamanian isthmus had formed. The Caribbean was also invaded many times, including by the bee lineage from North America, which then recolonised South America and produced further species alongside existing lineages.
In the space of just 22 million years, hummingbirds have diversified from a single common ancestor that lived in the lowlands of South America to over 350 extant species that span the Americas, from Alaska to Tierra del Fuego and the Caribbean. And yet, according to McGuire, their speciation rate is only slowing slightly, for although some clades have saturated the available environmental spaces, other clades are still evolving into new species at an extraordinary rate. Indeed, by comparing their extinction and speciation rates, McGuire estimates that the number of hummingbird species could double before reaching equilibrium. It seems, therefore, that the ascent of the hummingbirds is far from complete.
Hummingbirds are specialist nectar-feeders, and their ability to detect sugar-rich food sources enabled their colonisation of novel ecological niches and contributed to their dramatic rates of speciation. But how hummingbirds recognise sugars has, until recently, been unclear, since they do not possess sweet taste receptors. This scientific conundrum surfaced over 10 years ago when geneticists obtained the first complete sequence of a bird’s genome, that of the domestic chicken. To their surprise, unlike that of other vertebrates, avian DNA does not contain a gene that codes for a functioning sweet receptor.
Most vertebrates perceive sweet and savoury tastes by expressing a family of receptor genes, called T1Rs. Savoury or umami flavours are detected by the heterodimer T1R1–T1R3, a receptor that is sensitive to amino acids, while the T1R2–T1R3 heterodimer functions as a sugar receptor. In 2014, Maude Baldwin, a doctoral student at Harvard University, and her colleagues used this knowledge to analyse the genomes of 10 species of birds, from chickens to flycatchers.9 They found that seed- and insect-eating species possess savoury receptors, but not a T1R2 gene needed for sugar detection. Since the lack of the T1R2 gene is widespread among birds, it is likely that their carnivorous ancestors, the therapod dinosaurs, also lacked sweet receptors. According to Baldwin, ancient birds lost their T1R2 gene because there was no need for meat-eaters to detect sugars. But this reasoning poses a problem. While the ability to detect sugars is not necessary for chickens and flycatchers, many species, including hummingbirds, live on nectar, a food source made up almost entirely of simple sugars. Indeed, hummingbirds consume more than their own body weight in nectar each day and can instantly tell the difference between a weak sugar solution and water. So how can nectarivorous species find food if they lack a gene for sweet taste?
To answer this question, Baldwin and her team cloned the taste receptors from three species of bird: the sugar-insensitive domestic chicken, Anna’s Hummingbird and the closest living relative of the hummingbirds, the insectivorous Chimney Swift. After expressing all three receptors in cell lines, the scientists were able to show that the hummingbird’s savoury receptor responds to sugars, unlike those of the chicken and swift. When they looked more closely, they found that at least 19 amino acids had been substituted in the hummingbird’s T1R3 protein and that these changes imparted sugar sensitivity to its savoury receptor. In other words, hummingbirds have evolved the capacity for carbohydrate recognition by converting a savoury receptor into a sugar one – an event that must have occurred after their lineage diverged from its insectivorous ancestors at least 48 million years ago. Future studies on other nectar-feeding families, such as honeyeaters and sunbirds, are awaited to see if evolution has used a single strategy, or a range of different strategies, to solve the problem of sweet detection in birds.
While we cannot be sure how the change in hummingbirds’ taste perception occurred, one can imagine a likely scenario. An ancestral population that lacked sugar taste could have accidentally ingested some nectar while hunting insects among flowers. Any individual with an appropriately mutated T1R3 gene, one that enabled the detection of sugar for the first time, would have been given access to a novel energy source. If fitness were improved, then the mutated gene would increase in frequency from generation to generation. Eventually, after multiple receptor modifications, the nectar-seeking population would have gained a marked evolutionary advantage over their insect-eating ancestors.
The emergence of sugar receptors changed the course of hummingbird evolution and enabled their nectarivorous lifestyle to develop. However, to satisfy their daily energy needs, hummingbirds have to consume an extraordinary amount of nectar, equal to several times their body weight each day. This value is far higher than in any other bird species of the same size, and, as a result, their kidneys have had to evolve the ability to excrete large volumes of dilute urine. Hummingbirds also possess a unique glomerular structure and a dense nephron blood supply that allows a precise control of blood electrolyte levels, despite consuming nectars with widely different sodium and potassium concentrations.
How hummingbirds syphon nectar so quickly has, until recently, been a mystery. It had always been assumed that capillary action filled the two grooves along their tongues, in the same way that sponges and paper towels soak up water, even though such a mechanism would struggle with the volumes required. Now, according to Alejandro Rico-Guevara and Margaret Rubega at the University of Connecticut, it seems that hummingbird tongues act like miniature pumps.10 Using slow-motion videos and transparent artificial flowers, the scientists studied 18 species from seven of the nine clades of hummingbird. In all cases, they found that when the bird’s tongue extends, the grooves on each side are compressed shut by the bill, storing potential energy in their walls. But once the tongue touches the nectar, the grooves spring open, and the released energy sucks up the nectar to fill the tubes in just a few milliseconds. Each time the bird compresses its tongue to release the nectar, the pump is reset for another mouthful, a process that can occur up to 14 times a second.
But hummingbirds must also supplement their diet with occasional insects, because nectars are deficient in proteins that provide essential amino acids. To do so, they have evolved a unique means of flexing their lower mandible that involves bending the jaw in two directions simultaneously. This movement, which widens the gape and enables flying insects to be taken, is associated with a complex deformation of surrounding bone. It seems that despite hummingbirds’ close evolutionary relationship with flowers, their past insectivorous lifestyle continues to have an influence on their form and function.11 Nectar is low not only in protein but also in calcium, an element that is required in significant quantities around the time of egg production. Females of most avian species store calcium in tissue called medullary bone, but hummingbirds possess only small amounts of this substance. Instead, the birds consume all sorts of mineral-rich compounds, including wood ash, slaked lime and sand, which they obtain by hovering over the ground and flicking their long tongues in and out.12
To be able to hover long enough to obtain nectar, hummingbirds have evolved an insect-like flight style. Indeed, there is even a Cuban endemic named the Bee Hummingbird that, at only 5 centimetres long and 2 grams in weight, is the smallest bird in the world. Unlike most flying vertebrates, which can only produce lift when their wings flap downwards, hummingbirds can do so on the upstroke as well. By filming Ruby-throated Hummingbirds in flight, Tyson Hedrick and his team showed that this ability comes from the bird’s relatively small wrist bones, which allow the wings to move through a 140-degree arc.13 Hummingbirds are also able to beat their wings faster than any other species, up to 70–80 beats a second, and, with only slight changes in wing pitch, can fly in any direction, even upside down. To power such energy-demanding flight hummingbirds have evolved the highest metabolic rate of any vertebrate, about 30 times that of humans, and have developed flight muscles with the highest known density of energy-releasing mitochondria.14 Their cardiovascular system is no less astonishing. Hummingbird lungs have an oxygen diffusion capacity that is 10 times greater than similar-sized vertebrates, while their hearts are proportionately twice as large, beating 250 times a minute at rest, rising to 1,200 beats a minute during flight. Furthermore, species that live at high altitudes, such as the Andean Hillstar, have evolved high-affinity haemoglobins to cope with the low oxygen levels (see The Waterfowl’s Story).
Hummingbirds must also process visual information and respond to their environment quickly to avoid collisions, especially when hovering and fighting off intruders. To do so, a highly conserved part of the brain, the nucleus lentiformis mesencephali (nLM), is enlarged and contains neurones that are tuned to detect motion in all directions.15 In contrast, the nLM of other birds and all four-legged vertebrates (where it is known as the nucleus of the optic tract) is proportionally smaller and primarily detects back-to-front motion. This evolutionary adaptation provides the Trochilidae with the fine motor control needed to hover and zoom quickly in every direction possible, at speeds of up to 60 kilometres an hour. Furthermore, hummingbirds have evolved a markedly enlarged hippocampal formation: an area of the brain that is responsible for memory and learning.16 As a result, they can recall the nectar quality and content of flowers, as well as their location and distribution, so that they can forage efficiently without wasting time and energy. Spatial cognition also enabled the development of trap-lining, a feeding strategy in which some species visit the same few flowers over long distances, much as trappers check their lines of traps. In effect, trap-lining allowed the allocation of limited resources between different taxa and contributed to further hummingbird speciation.
Since the highlighted adaptations characterise most hummingbirds, they must have evolved after the lineage’s divergence from swifts and before their arrival in South America 20 million years later. Collectively, they enabled hummingbirds to develop their nectarivorous lifestyle and so kick-started their remarkable divergence and speciation. But flowers do not provide hummingbirds with energy-rich nectar out of kindness. As payment, they require cross-pollination – and many plants have evolved a ‘pollination syndrome’, consisting of a range of ‘pro-bird’ and ‘anti-bee’ inducements. These include the provision of sucrose-rich nectar, since bees prefer fructose and glucose, and brightly coloured red flowers that lack scent, as smell is essential for insects whereas hummingbirds rely on vision. Bird-pollinated (ornithophilous) plants typically have long tubular flowers and an orientation of stamen and stigma to maximise the chances of fertilisation and prevent contamination from the wrong type of flower. Furthermore, ornithophilous species have corollas that lack a landing platform, and their petals are usually angled to prevent access to insects. Over 7,000 plants in 404 genera from 68 families are now dependent on one or more of the 353 species of hummingbird for their pollination. Nevertheless, ornithophily is thought to be a costly strategy for plants, and the condition has only evolved where there are obvious benefits, as in high-altitude ecosystems that lack insect pollinators, in dry environments, and for sparsely distributed species.
The close relationship between plant and bird has led to some remarkable morphological adaptations. The Buff-tailed Sicklebill, for example, sports a bill that arcs a full 90 degrees downwards to enable it to reach nectar located deep within the corollas of Centropogon flowers (Plate 19A). At the same time, the plants have evolved protruding brush-like anthers to ensure that the sicklebill inadvertently collects a dusting of pollen on its forehead while feeding. In other words, the shape of the bird’s bill has coevolved with the form of the plant, as both species benefit from an exclusive nectar–pollen relationship. The Sword-billed Hummingbird, in contrast, has the longest bill of any family member and is the only species with the reach to obtain nectar, and hence pollen, from certain species of passionflower in the genus Passiflora (Plate 19B). Because of its bill shape, the Sword-billed Hummingbird uses its feet to preen and adopts a slightly diagonal stance, with its head pointing upwards to balance.
But which evolved first: passion flowers with 10-centimetre nectar-tubes, or hummingbirds with 11-centimetre bills? By analysing DNA from 43 species of passion flower, German scientists found that the plants with the longest nectar-tubes evolved 10.7 million years ago, shortly after the Sword-billed Hummingbird diverged from its shorter-billed relative, the Great Sapphirewing.17 Subsequently, bird and plant evolved together, since each species would struggle to survive without the other. For this particular plant–bird relationship, the scientists made an even more instructive observation. When the Sword-billed Hummingbird population fell dramatically around 3–4 million years ago because of environmental upheavals, some passionflowers quickly evolved shorter corollas to allow pollination by other bird species, as well as bats, and ensured their survival. Biologists have also shown that some hummingbird flowers can rapidly change to bee pollination as the result of a single mutation that alters their colour from red (preferred by hummingbirds) to violet (liked by bees).18 Evolution, it seems, is not always irreversible, and a few species can escape the perils of overspecialisation. For them, survival of the fittest is, in reality, the survival of the fastest to evolve.
The relationship between plants and insect pollinators has led to a marked speciation in flowers, as individual populations adapt to their primary pollinators and coevolve over time. However, a recent paper by Stefan Abrahamczyk and Susanne Renner from the University of Bonn has revealed that speciation has been less dramatic for ornithophilous plants.19 One explanation might be that hummingbirds cover greater distances than insect pollinators and, by increasing gene flow between plant communities, have reduced the likelihood of the population fragmentation needed for speciation. Furthermore, the driving force for the speciation of ornithophilous plants is lessened by the fact that most hummingbirds, unlike insects, rarely restrict their food source to one taxon and will often pollinate several plant species. While the longstanding evolutionary relationship between hummingbirds and flowers is undeniable, the work by Abrahamczyk and Renner implies that the relationship may be a stagnant one for some species, since ‘without the promise of greater fidelity, plants will only change so much to accommodate their partners.’20
Given that hummingbirds evolved in Eurasia, one might expect to find flowers in the Old World that still exhibit a ‘pollinator syndrome’. In fact, botanists believe they have found such plants. Several species in Asia and Africa appear to have retained morphological features similar to those pollinated by hummingbirds in the Americas, despite growing in areas devoid of hovering avian pollinators. They include the Himalayan Lantern (Agapetes serpens), a beautiful shrub with tiny red hanging lantern-like flowers, and the herbaceous Canarian group of plants from west Africa (family Campanulaceae) that also have red bell-shaped flowers.5 Both plant groups are now pollinated by other avian species, especially sunbirds. However, Canarina canariensis from the Canary Islands has survived by adapting to pollination by non-specialist nectar-feeders, such as the Canary Islands Chiffchaff and the local race of Spectacled Warbler.
Could it be that hummingbirds originally coevolved in parallel with the nectar-laden flowers of tropical Africa long before the emergence of nectarivorous passerines? Perhaps Emily Dickinson was right all along, and the hummingbirds did indeed originate not so far from Tunis.