Search

Trait matching between the Madagascan orchid Angraecum sesquipedale, with a nectar spur of 33 cm, and a hawkmoth with a tongue almost as long has fascinated biologists since Darwin, who saw only flowers but correctly predicted the moth pollinator. This moth, Xanthopan morgani praedicta, was described from museum specimens in 1903 and documented as the pollinator in photographs and film in 1992 and 2004. However, Madagascar harbours c. 30 species of long-spurred orchids and seven species of long-tongued hawkmoths, and mutualisms between moths and 'Angraecum and other deep tubular flowers' (Darwin, 1862: p. 202) probably involve a network of interacting species. We infer the evolutionary time over which Madagascan sphingids and Angraecum have interacted, based on an orchid phylogeny that includes 62 of 144 Angraecum species on Madagascar and a moth phylogeny with all nine Madagascan Sphinginae. Clock models using either rate- or fossil-based calibrations imply that the Madagascan subspecies X. morgani praedicta and the African morgani diverged 7.4 +/- 2.8 Mya, which overlaps the divergence of An. sesquipedale from its sister, Angraecum sororium, namely 7.5 +/- 5.2 Mya;since both have extremely long spurs, long spurs probably existed before that. The phylogenies moreover show that several long-tongued moths and long-spurred orchids probably coevolved on Madagascar since the mid-Miocene, although field data on the moths' levels of polyphagy and pollen transport are lacking.

The display of scent is crucial for plants in attracting pollinating insects to flowers and ensuring successful pollination and reproduction. The large number of aldoxime volatile species present in the scent of the Madagascan orchid Angraecum sesquipedale has been suggested to play a primary role in attracting the sphingid moth Xanthopan morgani praedicta. By solid phase micro-extraction (SPME) coupled with gas chromatography-mass spectrometry (GC-MS), we monitored the scent release from different flowers of a single orchid, day and night throughout the entire flowering period. In separate experiments, the diurnal release was monitored in 3h intervals and the tissue specific release from the different floral parts was tracked. Numerous novel compounds related to the aldoxime metabolism not previously detected in A. sesquipedale were identified and positioned into a proposed pathway for aldoxime metabolism. From the results, we hypothesize that (E/Z)-phenylacetaldoxime and its derivatives could be important attractants for the pollinating moth X. morgani praedicta. By applying an untargeted Partitioning Around Medoids (PAM) cluster analysis to the metabolite profiles in the scent, the proposed pathways for the formation of aldoximes were substantiated. With this study, we demonstrate the powerful utility of a bioinformatics tool to aid in the elucidation of the routes of formation for volatiles and provide a benchmark and guidelines for future detailed observations of hawkmoth pollination of Angraecum species, and in particular A. sesquipedale, in the wild.

In this review we provide a detailed description of Darwin's prediction of the coevolution of a long‐spurred orchid, Angraecum sesquipedale, and a long‐tongued moth, his correspondence on the subject, the history of the moth and the subsequent literature. On seeing the long spur of A. sesquipedale, Darwin predicted that its pollinator would be a moth with a long proboscis. For more than a century following Darwin's prediction this was assumed to be the case. The pollinator was taken to be Xanthopan morganii praedicta, despite the fact that it had not been observed to visit A. sesquipedale flowers. Direct observations, infra‐red cinematography and photographs published between 1993 and 1997 and a video made in 2004, all of which show X. morganii praedicta visiting A. sesquipedale flowers and removing pollinia, proved that Darwin's prediction was accurate. Recent research suggests that selection pressure exerted by predators on the pollinators, resulted in the evolution of extreme tongue lengths and a special hovering flight. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, 169, 403–432.

The concept of coevolution was first developed by Darwin, who used it to explain how pollinators and food-rewarding flowers involved in specialized mutualisms could, over time, develop long tongues and deep tubes, respectively. He famously predicted that Angraecum sesquipedale, a long-spurred Malagasy orchid, must be pollinated by a hawkmoth with an exceptionally long tongue. Darwin’s idea of a coevolutionary “race” was championed by contemporary naturalists, including Alfred Wallace, and a hawkmoth fitting the expected tongue-length profile was eventually discovered in Madagascar during the early twentieth century. However, strong empirical support for the mechanism behind Darwin’s coevolutionary model has been forthcoming only in the past two decades. It is now established that selection often strongly favors plants with floral tubes that exceed the length of their pollinator’s tongues. There is also evidence that pollinators gain an energetic benefit from having tongues that enable them to consume most or all of the nectar in deep tubular flowers. Alternative explanations for the evolution of long pollinator tongues, such as evasion of predators that use flowers as ambush sites, are considered much less compelling and lack quantitative support. Another important advance in coevolution research has been the development of approaches that explicitly predict a geographical mosaic of coevolution. The expectation that coevolution can lead to geographical diversification and trait covariation among strongly interacting organisms is strongly supported by studies of long-proboscid fly and oil-bee pollination systems in South Africa. Macro- and microevolutionary studies of pollination systems suggest that coevolution can operate alongside other one-sided evolutionary processes, such as shifts, in shaping plant and pollinator traits.

It is normally thought that deep corolla tubes evolve when the plant's successful reproduction is contingent on having a corolla tube longer than the tongue of the flower's pollinators. Combining optimal foraging theory and quantitative genetics in a spatially explicit, individual-based model, we show that flowers with long corolla tubes can alternatively evolve because they promote resource partitioning among nectar feeders and increase the probability of conspecific pollen transfer. When there is competition for resources, long-tongued flower visitors feed preferentially at deep flowers and short-tongued visitors at shallow flowers. Both plant species thus benefit when the depths of their corollas are so different that each flower visitor specializes on one species. Resource competition can promote the evolution of deep corollas despite the presence of significant amounts of noise, such as deviations from optimal foraging behavior due to perceptual errors or temporal fluctuations in the relative abundance of competing pollinator species. Our results can explain the evolution of long corollas in a number of systems that do not conform to the traditional view.

Some of the most spectacular examples of coevolution between flowers and their pollinators are reflected in their morphologies. Tongue length of insects and bill lengths of nectar-feeding birds are some of the most significant characters in pollination studies (Kearns & Inouye 1993). As Darwin noted as early as 1862, the evolution of deep floral tubes or spurs and long tongues of flower visitors can be explained by runaway coevolution (Nilsson 1988). However, such a process has only been shown to be likely in a few cases, e.g. between Malagasy orchids and hawkmoths (Nilsson et al. 1985). The pollinator of the Malagasy orchid Angraecum sesquipedale, with a 30 cm nectar spur, is a sphingid moth with a tongue of a similar size (Darwin 1862, Nilsson 1998). However, most studies of such extreme pollination specialization also report that the interaction is asymmetrical, i.e. the pollinators interact with a guild of plants, whereas the plant often depends on only a few pollinators (Johnson & Steiner 1995).

The pollination process of the extremely long-spurred orchids Angraecum sesquipedale and A. sororium is described and documented here for the first time. The pollinaria and viscidia load was examined in moths captured in central and south Madagascar. Visits to orchids by hawkmoths were rarely observed in the field and were therefore systematically recorded in large flight tents using a night-vision video technique and flashlight photography. Angraecum sesquipedale in Fort Dauphin is pollinated by Xanthopan morgani praedicta and A. sororium on Mt. Angavokely by Coelonia solani. By combining a deep nectar spur of extraordinary length with a protruding labellum functioning as a landing platform, these orchids overcome the moth's stereotypic swing-hovering flight thus enabling full insertion of the long tongue. Angraecum compactum in Foret d'Ambohitantely is pollinated by both the shorter and longer-tongued forms of Panogena lingens which never swing-hover but is also exploited by X. morgani and C. solani with wastage of pollinaria. The duration of tongue insertion, nectar exploitation and tongue withdrawal were analyzed: legitimate and illegitimate visitors differ in their time budget and approach to the flower. Nectar volume, nectar level and sugar concentration of A. sesquipedale and A. sororium were compared with the nectar requirements of the pollinating hawkmoths. The evolution of very long spurs in these orchids is likely to have involved a series of pollinator shifts. The orchids adapted to different hawkmoth species with increasingly long tongues which primarily evolved to avoid predator attacks during visits to less specialized flowers. This “pollinator shift” model modifies the classical “coevolutionary race” model. The relevance of the taxon Angraecum bosseri Senghas is questioned.

This paper is a case study of Angraecum sesquipedale, better known as the Madagascar Star Orchid. Darwin’s prediction of a moth pollinator for this orchid, with a proboscis of up to 11 in., was later vindicated, and taken as further vindication for a theory of evolution itself. Borrowing Huxley’s comparison with the prediction and discovery of Neptune, we look at the relevance of Darwin’s prediction for twentieth century philosophical debates on scientific methodology, specifically highlighting the distinction between the naive falsifiability of Karl Popper and the sophisticated falsifiability of Imre Lakatos.

The core for disagreement about the tongue-flower elongation, coevolution versus non-coevolution, is whether a long hawkmoth tongue is used primarily for feeding or for predator avoidance. Wasserthal[1xWasserthal, L.T. Trends Ecol. Evol. 1998; 13: 459–460Abstract | Full Text | Full Text PDF | PubMedSee all References][1]considers his predation hypothesis for tongue length confirmed, but his three cited papers lack quantitative data on natural predation. Svensson et al.[2xSvensson, M.G.E., Rydell, J., and Tove, J. Trends Ecol. Evol. 1998; 13: 460Abstract | Full Text | Full Text PDF | PubMedSee all References][2]argue in support of his interpretation by anecdotal evidence. They miss the flight–elongation focus. However, I agree with them and with Samways[3xSamways, M.J. Trends Ecol. Evol. 1998; 13: 460Abstract | Full Text | Full Text PDF | PubMedSee all References][3]that major predation on large hawkmoths is expected from flying vertebrates, not arthropods in ambush a la Wasserthal.Nectar-extracting mouthparts have attained impressive extremes in euglossine bees (4 cm), nemestrinid flies (8 cm) and hummingbirds (10 cm), each set of taxa using/pollinating a unique guild of deep-flowered plants[4xSchremmer, F. Verh. D. Zool. Ges. (Suppl.). 1961; 25: 375–401See all References, 5xSee all References]. Nothing suggests that predator avoidance has influenced the elongation (co)evolution. Long-tongued hawkmoths and their deep flowers, although displaying world records (30 cm and 40 cm, respectively), are hardly exceptions in this respect. The various very deep-flowered angiosperms (e.g. Epiphyllum in neotropical Cactaceae) are probably coevolved with hawkmoth tongues.Hawkmoth flight repertoires include `on-the-spot' hovering as well as `swing-hovering' during flower-visits. Also, hummingbirds, long-tongued solitary bees (euglossines, anthophorids) and long-tongued flies (horseflies, nemestrinid and bee flies) hover on-the-spot. No doubt this behaviour favours targeting and insertion precision with the extended mouthparts into narrow nectariferous structures. Extremely long-tongued hawkmoths probably require even more precision control. Added swing-hovering may increase visual and/or olfactory information about the position of the flower. Anti-predator adaptations, on the other hand, are to be expected among other traits.Basal hawkmoth clades contain the long-tongued taxa[6xRothschild, L.W. and Jordan, K. Nov. Zool. 1903; 9: 1–972See all References][6]. But modern phylogenetic analyses (so badly needed!) are likely to show a very long tongue to be a derived, relatively late trait within several lines (e.g. Amphimoea, Coelonia). Furthermore, very long-tongued hawkmoths might not be true generalists but instead might have have guild constancy under natural, undisturbed conditions. Their opportunistic exploitation of pantropical weedy Lantana[7xWasserthal, L.T. Bot. Acta. 1997; 110: 1–17CrossRefSee all References][7], a lepidopteran-adapted species, is hardly surprising.Wasserthal's argument that `obligate synchrony and syntopy' of Xanthopan morganii praedicta and Angraecum sesquipedale would have been required for coevolution is flawed. These interactors are participants of coevolving guilds[8xNilsson, L.A. et al. Biotropica. 1987; 19: 310–318CrossRefSee all References][8]. Various pollinator shifts may well have occurred, as are commonplace in the history of orchids and other angiosperms[9xVogel, S. Bot. Stud. 1954; 1: 1–338See all References, 10xJohnson, S.D., Linder, H.P., and Steiner, K.E. Am. J. Bot. 1997; 85: 402–411CrossRefSee all References]. Coevolution and shifting occur by relatively `better' pollinators. The basis for Darwin's reciprocal-effect model is that a competitive visitor with a non-bottom-reaching tongue better contacts the flower's pollination organs[11xSee all References][11]. The depth of spur and the length of reward-extracting organ become inter-dependent and matching at the population level[12xSteiner, K.E. and Whitehead, V.B. Evolution. 1990; 44: 1701–1707CrossRefSee all References][12]. A consistently bottom-reaching, although competitive, hawkmoth tongue excludes pollinator shifts, as indicated by Xanthopan on Angraecum compactum[7xWasserthal, L.T. Bot. Acta. 1997; 110: 1–17CrossRefSee all References][7]. Some 100 Madagascan angraecoids have demonstrably retained adaptation to medium-tongued hawkmoths. Angraecum sesquipedale, guild-member angiosperms, Xanthopan and major hawkmoth competitors are evidently coevolved.The intention of Wasserthal's study has not been overlooked, nor has the difficulty of obtaining natural interaction data. Our team still recorded frequent pollination of A. sesquipedale in 1983 and 1986. Wasserthal's study of the same populations, species and systems in Madagascar has partly led to more questions than answers, and he has yet to adopt a plausible evolutionary interpretation of his observations.

It started on a quiet day in January 1862 with the arrival of a parcel addressed to Charles Darwin from Robert Bateman. The package contained several orchid specimens needed by Darwin for his research on insect pollination of orchids, including one that took Darwin by surprise. It was the “astounding” Angraecum sesquipedale Thouars, a large orchid with a nectary one foot long! Darwin thought, “what insect could suck it?” (Burkhardt & Smith 1985). And so began a forty-year story that illustrated the power of evolution by natural selection, created a controversy regarding creation, and predicted the existence of a “gigantic moth.”

In 1862, in explaining the exceptionally long nectar spur of Angraecum sesquipedale, the star-of-Bethlehem orchid, Darwin proposed that a coevolutionary 'race' had driven the increase in length of a plant's spur and its pollinator's tongue. He predicted the existence of an exceptionally long-tongued moth. It — Xanthopan morgani ssp. praedicta — was discovered in 1903, with a tongue length of 22 cm. But the 'race' model remained contentious as there are other ways in which a curiously long tongue could evolve. Now, using a species-level phylogeny of the columbine genus, Aquilegia, Justen Whittall and Scott Hodges show that nectar spurs in Aquilegia have indeed evolved in an ever-increasing fashion, but due to a predictable series of adaptations to unrelated pollinators, concentrated during speciation events. Nectar spurs in Aquilegia have evolved in an ever-increasing 'co-evolutionary race' fashion, but due to a predictable series of adaptations to unrelated pollinators, concentrated during speciation events. These findings show how evolutionary history and ecological setting can drive predictable changes in morphology. Directional evolutionary trends have long garnered interest because they suggest that evolution can be predictable. However, the identification of the trends themselves and the underlying processes that may produce them have often been controversial1. In 1862, in explaining the exceptionally long nectar spur of Angraecum sesquipedale, Darwin proposed that a coevolutionary ‘race’ had driven the directional increase in length of a plant’s spur and its pollinator’s tongue2. Thus he predicted the existence of an exceptionally long-tongued moth. Though the discovery of Xanthopan morgani ssp. praedicta in 1903 with a tongue length of 22 cm validated Darwin’s prediction3, his ‘race’ model for the evolution of long-spurred flowers remains contentious4. Spurs may also evolve to exceptional lengths by way of pollinator shifts as plants adapt to a series of unrelated pollinators, each with a greater tongue length5. Here, using a species-level phylogeny of the columbine genus, Aquilegia, we show a significant evolutionary trend for increasing spur length during directional shifts to pollinators with longer tongues. In addition, we find evidence for ‘punctuated’ change in spur length during speciation events6, suggesting that Aquilegia nectar spurs rapidly evolve to fit adaptive peaks predefined by pollinator morphology. These findings show that evolution may proceed in predictable pathways without reversals and that change may be concentrated during speciation.

Though the term ‘coevolution’ was coined less than 50 years ago (Ehrlich and Raven, 1964), over 100 years prior to that Darwin first described the process to account for the matching of tongue and tube lengths of bee species and the clover species they visited (Darwin, 1859). He expounded on this process when describing the exceptionally long spur of Angraecum sesquipedale and predicted a moth pollinator with an equally long tongue (Darwin, 1862). Darwin envisioned a ‘race’ in which long-tongued individuals had a fitness advantage in gaining food from exceptionally long-spurred flowers and long-spurred flowers had a fitness advantage in reproduction due to increased pollination. Without a countervailing force, the tongue length of the pollinator and the spur length of the plant would increase due to the reciprocating fitness advantages. When traits of two interacting species such as tongue and floral tube length match tightly, it is easy to assume that Darwin's race, that is, coevolution, is responsible.

Studies on hawkmoth pollination in a primary forest in central Madagascar indicated that at least five of six flowering, long-spurred angraecoid species exploited the same hawkmoth, Panogena lingens, as a pollen vector. Several other long-tongued hawkmoth species were present, but no evidence was found that they participated in orchid pollination. The pollinaria of Angraecum arachnites, A. compactum, Neobathiea grandidierana, and Jumellea teretifolia were deposited on the basal portion of the proboscis. The latter three species utilized the same dorsal portion of the proboscis, whereas attachment by A. arachnites was ventral. The pollinaria of Aerangis fuscata were deposited on the frons and palps. Hawkmoths frequently carried mixed loads, but at least ethological and mechanical mechanisms seemed to restrict interspecific pollination. The P. lingens-angraecoid orchid relationship may have reached a specialized state before progenitors of other long-tongued Sphingidae colonized or evolved in central Madagascar. THE HAWKMOTH FAUNA OF MADAGASCAR contains some 60 species and is highly endemic (Griveaud 1959, Carcasson 1967). Compared to other Old World faunas, it is especially rich in long-tongued forms. Tongues of at least six species are often longer than 7 cm, and two of these, Xanthopan morgani (Walk.) ssp. praedicta R. & J. and Coelonia solani (Boisd.), have tongues that often exceed 20 cm (Nilsson et al. 1985). In the Sphingidae, a long proboscis represents a plesiomorphic state, and all Malagasy long-tongued species belong to relatively archaic genera (Rothschild & Jordan 1903). If long tongues are traits produced through processes of coevolution or diffuse coevolution (sensu Janzen 1980) in relation to food plants of adult hawkmoths, then a corresponding guild of plants with long nectar spurs or corolla tubes should exist. About 10 percent ( > 70 species) of all Malagasy orchids have spurs 8 cm or longer (cf. Perrier de la Bathie 1939-1941); hawkmoth pollination is also suggested by the floral features of many members of other plant families. Darwin (1862) predicted that the Madagascar Star Orchid Angraecum sesquipedale Thou., with a spur often exceeding 30 cm, was pollinated by an unknown giant hawkmoth. One very long-tongued candidate, discovered 41 years later, was accordingly named X. morgani praedicta (Rothschild & Jordan 1903). Thus a considerable guild of Malagasy plants exists that exploit the fuel demands of long-tongued hawkmoths. We recently described the pollination system of A. arachnites Schltr. (Nilsson et al. 1985). The present paper reports further on relations between Malagasy long-spurred orchids and hawkmoths and focuses on specialized pollination systems, shared and partitioned pollinator utilization, and species-isolating mechanisms between plants at low densities in a tropical primary forest. STUDY SITE AND METHODS The study was carried out 13-30 November 1983 in the Ambohitantely forest, 25 km northwest of Ankazobe on the Ankazobe tampoketsa, an old granitic erosion plain at 1500-1600 m altitude in central Madagascar. The forest consists of patches that remain mostly in ravines and along watercourses (Bastian 1964). These fragments of primary forest display a rich epiphytic orchid flora. We explored about 4 km2 of forest for flowering orchids and collected floral reference material of long-spurred species. Every night when the weather permitted, hawkmoths were captured at the border of the forest with a 2 5 0-W mercury vapor lamp. Total collection time was 91 hr. Some hawkmoths were stored intact, but in the majority, the proboscis was cut at the base and stored in separate tubes according to species. The proboscis of each hawkmoth individual was unrolled along a ruler, and under magnification both sides were examined for pollinaria. Position and quantity of pollen load were recorded. Pollinaria, or parts thereof, that could not immediately be identified to species were moistened, removed from the proboscis, dissected in 70 percent ethanol, and compared with reference material. Voucher specimens of hawkmoths and plants are deposited in UPS. I Received 12 October 1985, revision accepted 18 June 1986. 310 BIOTROPICA 19(4): 310-318 1987 This content downloaded from 157.55.39.104 on Mon, 20 Jun 2016 05:29:56 UTC All use subject to http://about.jstor.org/terms TABLE 1. Long-spurred orchids in flower recorded in Ambohitantely forest, and some of their general characteristics. Digits underlined in columns 2 and 5 represent the most common condition; whereas those underlined in column 6 denote the mean spur length. Approximate level aboveground of flower-proFlowering Frequency in No. of ducing indiperiod in flower flowers Length of spur Species viduals (m) the areaa observed producedb (cm) A. arachnites Schltr. 1-2-5 X-XII Abundant 1-3 10-12.7-15.5 A. compactum Schltr. 0-1-7 X-XII Sparse 1-4 13-14.5-17 N. grandidierana (Reichb.f.) Garay 2 XI 2 IC 14 J. teretifolia Schltr. 2-4-7 X-XI Sparse 1 12-13 A. articulata Schltr. 1-2 XI 2 12-19 7.5-11.4d A. fuscata (Reichb.f.) Schltr. 1-2-3 X-XI Sparse Ie 14 a Estimated from the flowering stage of each species. b In those plant individuals that produced flowers. c We have seen a pressed specimen (leg. J. P. Peyrot, Nov. 1961) in P (Paris) with several flowers. d Proximal and distal flower of one 12-flowered rachis. e Large individuals can produce several flowers (J. Stewart, pers. comm.).

DR. ALFRED RUSSEL WALLACE, in an article entitled “Creation by Law,” contributed to the Quarterly Journal of Science in October, 1867, alluded to a Madagascar orchis (Angraecum sesquipedale) with a nectary varying length from 10 inches to 14 inches, and prophesied that a hawk-moth will be discovered with a tongue of equal length to fertilise it. “That such a moth exists in Madagascar may be safely predicted, and naturalists who visit that island should search for it with as much confidence as astronomers searched for the planet Neptune—and they will be equally successful!” Will someone kindly tell me if this prophecy was fulfilled; if so, when, and the name of the moth?

MR. DARWIN, in his “Fertilisation of Orchids,” speaks of a Madagascar orchid (Angraecum sesquipedale) with nectaries 11½ inches long, and supposes that these plants must be fertilised by the efforts of huge moths, with probosces capable of such expansion, to obtain the last drops of the nectar which is secreted in the lower part of these whip-like nectaries. Can any of your readers tell me whether moths of such a size are known to inhabit Madagascar? They would probably be Sphingidae of some kind, as no other moths would combine sufficient size and length of proboscis.