Massive Fly Family Tree Rewrite: 117 Traits Reorganize Over 200 Chinese Lauxaniid Species

In the humid forests and river valleys of China lives a group of small, often strikingly patterned flies that most people have never heard of, yet that taxonomists have struggled with for more than a century. The genus Homoneura, a member of the acalypterate fly family Lauxaniidae, is one of the most species-rich lineages of its entire family, with more than 750 described species spread across eight subgenera and every major zoogeographical region except the Neotropics. More than 220 of those species occur in China, and virtually all of them — over 80 percent of the Chinese members of the genus — belong to a single subgenus, Homoneura in the strict sense. For decades, these flies were sorted into species groups largely on the basis of wing spots and a handful of external features, a system that many specialists suspected was convenient rather than truly evolutionary. A new study published in the Deutsche Entomologische Zeitschrift has now put that suspicion on firm statistical footing, and in doing so has redrawn the classification of the group from the ground up.

The research, led by taxonomists working at the College of Horticulture and Plant Protection at Henan University of Science and Technology, represents the most comprehensive morphological phylogenetic analysis ever attempted for this subgenus. The team examined 230 species and distilled their observations into a data matrix of 117 carefully defined morphological characters, ranging from the proportions of the head and the color of the ocellar triangle to the shape of the hypandrium, the arrangement of spines on the phallus, and the number and position of dark spots on the wing veins. Thirty-one of those characters come from the head, twelve from the thorax, eight from the legs, twenty-one from the wings, thirteen from the abdomen, and a remarkable thirty-two from the male genitalia — an emphasis that reflects a long-standing principle in dipterology: the intimate machinery of the male terminalia often carries the clearest signal of evolutionary divergence.

Preparing that signal was no small task. The researchers dissected and macerated the apical segments of the abdomen in cold saturated sodium hydroxide for six hours, neutralized the tissue with glacial acetic acid, and examined the cleared genitalia in glycerine under a stereomicroscope. External structures were documented with stacked macro-photography using a Canon EOS 6D and Helicon Focus software, while fine details were rendered as line drawings in Adobe Illustrator. Every character was then numerically coded: 98 binary characters and 25 multistate characters, with ancestral states scored as zero, derived states as one through three, missing observations as question marks, and inapplicable states as dashes. Partial characters and reference images were adapted and revised from an earlier morphological cladistic study of the group published in 2022 by Kong and colleagues, extended here to a far deeper sample of species.

To root the tree, the team used two lauxaniine species, Minettia (Frendelia) longipennis and Pachycerina decemlineata, together with representatives of every other homoneurine genus found in China, including Cestrotus, Dioides, Noonamyia, Phobeticomyia, and Prosopophorella. The analysis itself employed maximum parsimony, the classic cladistic criterion that favors the tree requiring the fewest evolutionary changes, implemented in the program TNT with implied weighting to down-weight highly homoplasious characters. The researchers tested a range of concavity constants, with K values from 2 up to 30, and found that the tree topology stabilized once K reached 20. At K equal to 30, traditional searches with 900 replicates yielded 45 most parsimonious trees, whose strict consensus had a length of 1413 steps, a consistency index of 0.11, and a retention index of 0.73. Branch support was assessed with bootstrap analyses in NONA 2.0 and Bremer support values calculated in TNT, and the character changes were mapped onto the consensus tree using WinClada.

The headline result is at once a triumph and a warning. On the positive side, the analysis produced a clear phylogenetic hypothesis for how the Chinese species of the subgenus should be organized, collapsing the previously recognized 21 species groups into 12. Three entirely new species groups are established — the nigrifacies, pallida, and patella groups — seven existing groups are retained, namely the beckeri, formosae, henanensis, nigra, notostigma, ornatifrons, and trispina groups, and the remaining groups are merged into two consolidated units, the laticosta and quinquenotata groups. Crucially, the new scheme no longer relies solely on wing spotting, the traditional and arguably over-used character complex, but integrates evidence from every part of the body, including a companion identification key built directly on the consensus tree that allows any Chinese species to be assigned to its group using visible traits such as frontal stripes, pruinose mesonotal bands, and the shapes of the crossvein spots.

The warning, however, is equally significant: neither the genus Homoneura as a whole nor the nominotypical subgenus emerged as monophyletic. Species of the subgenera Euhomoneura, Neohomoneura, Minettioides, and Chaetohomoneura appear scattered, mosaic-like, within the clade of typical Homoneura, and one member of Chaetohomoneura even sits at a terminal node deep inside Neohomoneura. The monophyly of three of the four testable subgenera is rejected outright, while that of Minettioides could not be verified because too few taxa were available. This is not the first such finding — the sole prior molecular study of the group, by Shi and colleagues in 2017 using two mitochondrial and two nuclear markers, likewise failed to corroborate subgeneric monophyly, and the 2022 morphological analysis by Kong and co-authors reached a similar conclusion — but the present work independently confirms it with a much larger morphological dataset, lending it substantial weight.

The historical context makes the scale of the problem clear. Species groups within Homoneura were first established in a serious way by Miller in his 1977 monograph on the Nearctic fauna, which revised 50 species into nine groups and four subgenera. Papp added groups from the Palaearctic and Oriental regions, Sasakawa proposed another from the Orient, and Kim organized the Australian species into 15 groups while erecting 20 groups for the global subgenera. Shi and Yang divided the Chinese fauna into 21 species groups in 2014, but admitted that some species defied placement. Only a handful of specialists have ever attempted this classification, in part because type specimens have been lost or destroyed and early descriptions were too brief for confident comparison. In short, the framework was inherited from an era of descriptive taxonomy that predated any serious attempt to test whether the groups reflected common ancestry.

What does the low consistency index of 0.11 mean in practical terms? It signals substantial homoplasy — the same character states arising independently in different lineages, a familiar hazard in groups where wing patterns and coloration are under strong ecological selection. The retention index of 0.73 is more encouraging, indicating that once the overall tree shape is fixed, a good proportion of the similarity among taxa is retained as genuine shared ancestry. The authors are candid that relationships among the species groups themselves were less consistently resolved than the broader pattern, that the placement of some early-diverging lineages lacks strong support, and that the fates of the henanensis group differed somewhat among the equally parsimonious trees. For that reason, they deliberately stop short of formally redefining the genus or its subgenera: the taxon sampling covers only Chinese species, and such a decision, they argue, demands a complete phylogeny built from both expanded morphological matrices and molecular sequence data.

Even so, the study delivers a practical and conceptual reset for one of the most diverse corners of the Diptera. It provides a tested, tree-based key to 12 species groups covering the entire Chinese fauna of the subgenus, a documented set of 117 characters that future workers can reuse and extend, and clear evidence that the traditional spot-on-the-wing approach has been masking real evolutionary structure. The team’s next steps include hunting for synapomorphies of the poorly supported early lineages, incorporating the growing repository of molecular data for lauxaniid species, and widening the geographic scope to other zoogeographical regions so that distributional evidence can help reconstruct the full evolutionary history of the genus. For a group in which taxonomists have repeatedly misclassified subgenera when describing new species, the message of this study is plain: the family tree of Homoneura looks quite different from its labels, and rewriting those labels has now begun in earnest.

Subject of Research: Phylogenetic relationships of the Chinese species groups of the subgenus Homoneura based on morphological characters

Article Title: Phylogeny of the Chinese species groups of the subgenus Homoneura Wulp, 1891 ( Diptera , Lauxaniidae , Homoneurinae ) based on morphological characters

Article References: Yao, Y., Kong, C., Miao, P., Zhao, S., & Li, W. (2024). Phylogeny of the Chinese species groups of the subgenus Homoneura Wulp, 1891 (Diptera, Lauxaniidae, Homoneurinae) based on morphological characters. Deutsche Entomologische Zeitschrift, 71(2), 241-254. https://doi.org/10.3897/dez.71.120389

Image Credits: AI Generated

DOI: 10.3897/dez.71.120389

Keywords: Homoneura, Lauxaniidae, Diptera, phylogenetics, morphology, maximum parsimony, species groups, taxonomy, China, male genitalia, monophyly, Homoneurinae