The study was a literature and database-based review, not a controlled experiment. The author searched the database underlying the earlier 2001 paper, systematically reviewed journal articles and faunal monographs, and added records identified through Google Scholar, regional butterfly websites and the Ant-Butterfly Interactions web project. Only field observations were included, while laboratory encounters and rearing conditions were excluded. Ants had to be identified at least to genus, either in the original report or from photographs. Records were analyzed as butterfly species paired with ant genus because species-level ant identification is often uncertain and records assigned to different names within the same genus were combined.

The analysis separated obligate from facultative myrmecophiles. Obligately associated species were those whose early stages were almost continually accompanied by ants, particularly when females oviposited only in the presence of host ants; species living as ant-parasitic nest inhabitants were also included. All other species with recorded ant associations were classified as facultative. The review covered seven ant subfamilies and six biogeographical regions, and ant-genus richness was assessed against the number of described species in each genus using data from AntWeb.

Ant genera from three subfamilies dominated the expanded database. The most frequently recorded genera were Crematogaster, with 221 butterfly-species-by-ant-genus pairs, and Camponotus, with 145. They were followed by Lasius, Tapinoma, Formica, Pheidole, Technomyrmex, Iridomyrmex, Plagiolepis and Polyrhachis. Most ant genera added since 2001 were recorded with only one to three lycaenid species, while changes in the taxonomic boundaries of Nylanderia, Iberoformica and Tetramorium also affected the comparison. The representation of genera in the two compilations was strongly correlated, with r = 0.986 and p < 0.0001.

Rarefaction and extrapolation indicated high sample coverage: 0.9890 for 2001 and 0.9899 for 2021. Extrapolation to more than 1,600 record pairs produced an estimate of about 66 ant genera, with a 95% confidence interval of 60 to 73. This suggests that additional records may add genera, particularly from under-explored regions such as South America and tropical Africa, although the overall list is unlikely to change dramatically. The estimate concerns the completeness of the compiled records rather than the total number of ant genera that could ever associate with lycaenid caterpillars.

Ant-genus richness was positively associated with the number of recorded associations, although the relationship showed substantial variation, with r = 0.589 and p < 0.001. Pheidole, Polyrhachis, Tetramorium and Temnothorax were recorded less often than their global species richness might suggest, while Crematogaster was over-represented. Two pronounced examples were Oecophylla, represented by three recognized species but recorded with 27 lycaenid species, and Anoplolepis, represented by nine species and also recorded with 27.

Obligate myrmecophiles represented 210 of the 523 species, or 40.1% of the database. Their recorded host ranges were substantially narrower than the visitor ranges of facultative myrmecophiles, a difference supported by a Welch test with t = 12.19 and p < 0.0001. Apart from exceptional records under unusual circumstances, obligate species were associated with ants from one genus. Ten facultative species had been observed with ants from at least 10 genera: Lampides boeticus with 15, Ogyris oroetes with 13, Plebejus melissa with 12, Icaricia icarioides and Theclinesthes albocincta with 11 each, and Ogyris amaryllis, Jamides celeno, Echinargus isola, Nacaduba berenice and Glaucopsyche lygdamus with 10 each. The review interprets this contrast in relation to the opportunistic nature of facultative encounters and the greater specificity of obligate host relationships.

Recorded ant-partner diversity was not consistently highest in tropical regions. It was highest in the Oriental region and especially the Australian region, while the Neotropical and Afrotropical regions had the lowest values. Crematogaster accounted for almost half of the Afrotropical record pairs and was the most prevalent genus in the Oriental region, with 50 pairs. Iridomyrmex led in the Australian region with 35 pairs, Formica in the Nearctic with 26, and Camponotus in the Neotropical and Palaearctic regions, with 16 and 56 respectively. The author identifies documentation gaps, ecological differences among butterfly clades and the presence of Riodinidae ant associations as possible explanations for the Neotropical pattern, but these explanations were not tested as causal mechanisms.

Important limitations remain. South American lycaenid life histories are incompletely documented, tropical tree canopies are under-studied, and some recent publications do not identify attendant ants beyond a general description. The database therefore combines unequal sampling effort with real variation in association patterns. Although the analysis provides a broad global synthesis of recorded interactions, it does not establish the frequency, intensity or evolutionary causes of most species-level associations. The author notes that, for the majority of lycaenid species, the identity of ant partners and the strength of their relationships remain unknown.

Subject of Research: Global field-record review of ant associations with Lycaenidae butterfly caterpillars

Article Title: The ant associates of Lycaenidae butterfly caterpillars – revisited

Article References: Fiedler, K. (2021). The ant associates of Lycaenidae butterfly caterpillars – revisited. Nota Lepidopterologica, 44, 159-174. https://doi.org/10.3897/nl.44.68993

Image Credits: AI Generated

DOI: 10.3897/nl.44.68993

Keywords: Lycaenidae, butterfly caterpillars, ant associations, myrmecophily, facultative interactions, obligate interactions, Crematogaster, Camponotus, trophic biology, biogeography, associates, butterfly