Illusion of flight: Absence, evidence, and the age of winged insects

Authors: SCHACHAT, SANDRA - Stanford University; Goldstein, Paul; DESALLE, ROB - American Museum Of Natural History; BOBO, DEAN - American Museum Of Natural History; BOYCE, C. KEVIN - Stanford University; PAYNE, JONATHAN - Stanford University; LABANDEIRA, CONRAD - Smithsonian Institute

Submitted to: Biological Journal of the Linnean Society, London
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/3/2022
Publication Date: 12/19/2022
Citation: Schachat, S.R., Goldstein, P.Z., Desalle, R., Bobo, D., Boyce, C., Payne, J., Labandeira, C. 2022. Illusion of flight: Absence, evidence, and the age of winged insects. Biological Journal of the Linnean Society, London. 138(2):143-168. https://doi.org/10.1093/biolinnean/blac137.
DOI: https://doi.org/10.1093/biolinnean/blac137

Interpretive Summary: This paper investigates a problem in our understanding of when winged insects originated. Estimates based on analyses of DNA sequence data suggest that winged insects appeared up to 100 million years before any of their fossils. Using a combination of statistical approaches, first principles, and reanalyses, it is demonstrated that these estimates are not plausible, and possibly an analytical artifact. This conclusion is crucial to our understanding of paleo-terrestrial invertebrate communities and is important because it suggests that time-calibrated molecular studies must be interpreted with caution. This work is of interest to entomologists, paleontoligists, molecular geneticists, and evolutionary biologists.

Technical Abstract: The earliest fossils of winged insects (Pterygota) are mid-Carboniferous (latest Mississippian, 328–324 Mya), but estimates of their age based on fossil-calibrated molecular phylogenetic studies place their origin at 440–370 Mya during the Silurian or Devonian. This discrepancy would require that winged insects evaded fossilization for at least the first ~50 Myr of their history. Here, we examine the plausibility of such a gap in the fossil record, and possible explanations for it, based on comparisons with the fossil records of other arthropod groups, the distribution of first occurrence dates of pterygote families, phylogenetically informed simulations of the fossilization of Palaeozoic insects, and re-analysis of data presented by Misof and colleagues using updated fossil calibrations under a variety of prior probability settings. We do not find support for the mechanisms previously suggested to account for such an extended gap in the pterygote fossil record, including sampling bias, preservation bias, and body size. We suggest that inference of an early origin of Pterygota long prior to their first appearance in the fossil record is probably an analytical artefact of taxon sampling and choice of fossil calibration points, possibly compounded by heterogeneity in rates of sequence evolution or speciation, including radiations or ‘bursts’ during their early history.