- Facultative butterfly-ant interactions - the role of variation in composition of nectar secretions (2004)
- The significance of variation in nectar secretions of facultatively ant-associated lycaenid butterfly larvae was investigated. The strongly myrmecophilous European Polyommatus coridon, and two moderately myrmecophilous species, the Palaearctic P. icarus and the subtropical Zizeeria knysna were used. Both Polyommatus species are closely related, Z. knysna is a far more distant relative. To obtain high numbers of caterpillars a new method for breaking the egg diapause of the univoltine P. coridon was established, resulting in 65% subitaneously hatching larvae. Based on observations of ants tending caterpillars a new “artificial ant” was assembled, allowing further studies on the ants’ stimulatory antennation pattern. Nectar-harvesting with micro-capillaries from caterpillars attended by ants was optimised and allowed determination of individual secretion droplet size based on large samples. Mean droplet size was 3.7nl in P. coridon, 2nl in P. icarus and 1.4nl in Z. knysna, in the latter two species 65-79% smaller than previously reported. Comparative chemical analyses (HPLC) revealed sucrose as main sugar component in nectar of all three species. In P. coridon it was accompanied by glucose and rarely by further sugars, but never by melezitose. In P. icarus and Z. knysna melezitose was the second most-important component, followed by fructose and glucose. Total sugar contents were 43.6±14.8g/l for P. coridon, 74.2g/l for P. icarus and 68.3±22.6g/l for Z. knysna. P. coridon nectar contained up to 14 amino acids. Major component was always leucine (50% of total), further important were tyrosine, proline, arginine, and phenylalanine. P. icarus nectar comprised up to six amino acids, dominated by tyrosine and phenylalanine. Z. knysna nectar contained only alanine and proline. Total amino acid contents were 9.7±3.4g/l for P. coridon, 1.2g/l for P. icarus and 0.3±0.2g/l for Z. knysna. Nectar composition was considerably different from hemolymph composition. Larval food had minor influence on P. coridon nectar composition. Caterpillars fed with semi-synthetic diet secreted more sucrose, with a trend towards higher total sugar content, and produced nectar with a more homogeneous amino acid pattern than larvae reared on natural host plants. Bioassays with ants from three different subfamilies (Lasius niger (Formicinae), Myrmica rubra (Myrmicinae), Tapinoma melanocephalum (Dolichoderinae)) demonstrated a preference for sucrose (standard concentration 0.1mol/l, as in P. coridon nectar) over monosaccharides. Melezitose in nectar concentration (P. icarus) was not preferred to sucrose. Some single amino acids in sucrose solutions were preferred over pure sucrose, e.g. leucine by L. niger, or phenylalanine and tyrosine by M. rubra. In general, raising amino acid concentration did not enhance preferences and even reduced them in some cases. Mixtures of four or eleven amino acids in sucrose and complete nectar analoga were preferred to sucrose. L. niger preferred a balanced mixture over an energetically similar, less balanced mixture. Due to the tremendous variability in gustatory preferences exhibited by the broad range of largely unpredictable ant visitors, lycaenid caterpillars should either decide on a balanced mixture, containing possible amino acid ‘key compounds’ in moderate concentrations, or if this investment does not pay, should secrete sugar-rich nectars. Using the data from the HPLC analyses and the droplet size measurements combined with published studies on secretion rates allowed estimations of the lifetime energetic value of secretions. One P. coridon caterpillar would deliver 5.5-24.4J, one P. icarus 1.9-14.2J and one Z. knysna 0.7-2.8J. Dissection and gravimetric analysis allowed estimation of the caloric equivalent of larval biomass. One early P. coridon third instar (onset of nectar secretion) would yield 17.6J, and prepupae 656J. Z. knysna second instars would yield 0.95J, third instars 3.2J, and prepupae 83.1J. Thus preying on the caterpillars rather than harvesting their secretions would be of greater energetic benefit. Comparisons of this model with further reports show the rather low benefit accruing to ants from tending facultatively myrmecophilous lycaenids, underpinning that manipulation of ants (by means of still unknown chemicals) must also be involved. The model data also suggest that the mutualistic nature of facultative caterpillar-ant associations will not always be granted, and be strongly conditional. Nectar composition data support the view that in myrmecophilous lycaenids secretions rich in amino acids are related to intimate, often obligate ant-associations, whereas facultative and unspecific myrmecophiles rely more on sugars. Degree of myrmecophily seems to be a better predictor of secretion content than taxonomic relatedness. The low investment costs of the caterpillars into nectar secretions well explain the enormous taxonomic and geographical distribution of facultative myrmecophily.