Deep flowers for long tongues: reply from L.A. Nilsson

Abstract

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.