The research, conducted by Petr Kočárek and Ivona Kočárková of the University of Ostrava and Robin Kundrata of Palacký University in the Czech Republic, required the team to overcome a fundamental obstacle in paleontology: the very characters that define the modern classification of Zoraptera, namely the intricate morphology of the male copulatory organs, are almost never visible in fossils. Because molecular genetic methods cannot be applied to insects trapped in amber, where DNA degradation over tens of millions of years is essentially complete, the systematic placement of fossil zorapterans had to be inferred entirely from observable morphology. The team found three male specimens among their newly examined Burmese amber material in which the genitalia were sufficiently exposed to allow homologization with the genital structures of living groups.

Living Zoraptera comprise only 47 described species, small soft-bodied polyneopteran insects that typically inhabit the bark of rotting wood in tropical and subtropical regions. The order displays a striking wing dimorphism, with most collected specimens being wingless and lacking both compound eyes and ocelli, while the rarer winged morphs are darker and pigmented. For more than a century, all extant species were placed in a single genus, Zorotypus, reflecting the superficial uniformity of the group. That conservative arrangement was upended when molecular phylogenetic studies using nuclear and mitochondrial markers revealed two deeply divergent lineages, which were classified as the families Zorotypidae and Spiralizoridae, each further subdivided into two robustly supported subfamilies. Crucially, this molecular framework was correlated with specific anatomical features: the symmetry or asymmetry of the male genitalia, the presence or absence of a basal plate on the copulatory organ, the elongation and coiling pattern of the intromittent organ, and the number of spurs on the hind tibia.

The new study harnessed these diagnostic characters and subjected them to a rigorous evolutionary analysis. Using ancestral character state reconstruction (ASR) under a maximum likelihood criterion, based on the published molecular phylogeny of the order, the researchers polarized key morphological traits to determine which states were ancestral and which were derived. Their analysis indicated that the ancestral Zoraptera probably possessed symmetrical genitalia with a basal plate and an elongated intromittent organ. Asymmetrical genitalia, lacking a basal plate, emerged as a synapomorphy, or shared derived character, of the family Zorotypidae. Within Spiralizoridae, a horizontally coiled intromittent organ was reconstructed as the ancestral condition for the family, while the vertically coiled configuration was identified as a defining synapomorphy of the subfamily Spiralizorinae. The metatibial spurs proved equally informative: the ancestral state was the absence of spurs, with three spurs characterizing Zorotypinae and two spurs characterizing Latinozorinae. Because these spur counts correlate reliably with genital morphology, they can be applied even to fossil females and poorly preserved specimens in which the copulatory organs cannot be seen, dramatically expanding the diagnostic reach of the classification.

Armed with this evolutionary framework, the team reshuffled the Mesozoic fossil record of the order. Twelve fossil species are currently recognized, one from Lower Cretaceous Jordanian amber (Albian stage) and eleven from the Upper Cretaceous amber of northern Myanmar (lower Cenomanian, dated to approximately 99 million years ago by uranium-lead dating of zircons in the volcaniclastic matrix). Previously, these fossils had been scattered among the catch-all genus Zorotypus sensu stricto, the monotypic genus Xenozorotypus, and the fossil subgenus Octozoros, all lumped within a single family. The new work erects two entirely new genera, Cretozoros and Burmazoros, elevates Octozoros from subgenus to full genus, and reinstates Paleospinosus Kaddumi, 2005 as a valid genus previously sunk in synonymy.

The resulting classification splits the Mesozoic fauna between the two modern families in ways that illuminate the deep history of the order. The genera Cretozoros, Paleospinosus, and Octozoros are assigned to Spiralizoridae: Latinozorinae, united by the synapomorphic presence of two metatibial spurs and by symmetrical male genitalia equipped with a basal plate encircled by an elongated intromittent organ. Burmazoros and Xenozorotypus, by contrast, join Zorotypidae: Zorotypinae, distinguished by asymmetrical genitalia lacking a basal plate and by three metatibial spurs. Cretozoros acanthothorax, originally described by Engel and Grimaldi in 2002, is diagnosed by eight-segmented antennae, reduced forewing venation lacking the anterior cubitus vein CuA1, and symmetrical genitalia with distinctive lateral rod-shaped accessory sclerites. Burmazoros denticulatus, based on a well-preserved alate male, shows a nine-segmented antenna, a metafemur bearing a middle spine larger than the basal spine, and reduced wing venation in which the media vein M1+2 is absent on the hindwing, an apomorphy defining the genus.

The revision also untangled a number of taxonomic knots created over years of scattered, sometimes simultaneous, species descriptions. Zorotypus hukawngi Chen and Su, 2019 is synonymized with Cretozoros acanthothorax after detailed morphological comparison revealed no diagnostic differences. Zorotypus hirsutus Mashimo, 2018, described independently and unknowingly alongside Liu and colleagues’ Zorotypus robustus in the same year, is synonymized with Octozoros robustus, which is simultaneously restored from its own prior synonymy with Octozoros cenomanianus. The authors caution that such redundancies are a chronic hazard of amber paleontology, where different preservation angles, compression artifacts, and the thin, easily deformed cuticle of zorapterans can make specimens of the same species appear dissimilar. They recommend comparing multiple specimens and paying close attention to the distribution pattern of spines on the metafemur, which appears to be a stable species-level character, while treating features like cerci shape and antennomere proportions with suspicion due to their susceptibility to deformation.

Wing venation, long dismissed as too variable for phylogenetic inference in this group, emerges from the study with partially rehabilitated status. The researchers confirmed that the absence of the vein CuA1 on the forewing is a genuine apomorphy of the extant genus Latinozoros, and found the same reduction independently in Cretozoros, an instance of homoplasy, or convergent evolution, that must be recognized to avoid misleading groupings. More distinctive still are unique hindwing patterns: Burmazoros lacks the M1+2 vein entirely, while Xenozorotypus uniquely retains M3+4, a vein absent in all other known zorapterans and considered an autapomorphy of that genus. The authors argue that wing venation deserves renewed, systematic analysis in light of the molecular phylogeny, noting that the last serious attempt to base a generic classification on wing veins, by Kukalova-Peck and Peck in 1993, was rejected largely because of the group’s overall morphological uniformity.

Two Mesozoic species stubbornly resist classification even after this overhaul. Zorotypus cretatus and Zorotypus dilaticeps remain incertae sedis, or of uncertain placement, owing to poor preservation, the availability only of apterous or dealated specimens, and the absence of observable male genitalia. Zorotypus dilaticeps is particularly enigmatic: a large species at 3.9 millimeters, its metatibia is armed with six acute spines and seven spine-like setae, an arrangement unique among fossil zorapterans. Because its holotype is a dealate female, no family-level assignment is currently possible. Their fates, the authors note, may change if better-preserved specimens surface from the amber deposits of Kachin State.

Beyond the taxonomic housekeeping, the study carries genuine evolutionary significance. The dominance of eight-segmented antennae among Mesozoic fossils, eight of eleven species, contrasts with the universal nine-segmented condition of all living adults, revealing that the reduction in antennomeres is a derived state that later reverted to the ancestral count in surviving lineages. The team defines a clade comprising Octozoros, Cretozoros, and Paleospinosus as the sister group of the extant genus Latinozoros, and proposes that Xenozorotypus is sister to a clade containing Burmazoros and the modern genera Zorotypus and Usazoros. Ancestral reconstruction of abdominal characters likewise showed that the presence of median projections on both the tenth and eleventh tergites is primitive, with the loss of the projection on T10 arising repeatedly, a further example of homoplasy that the new genus-level diagnoses carefully account for. By anchoring Cretaceous angel insects within the modern molecularly informed system, the study transforms a collection of morphologically uniform amber curiosities into a structured evolutionary narrative, one that documents the diversity of internal lineages flourishing in Cretaceous tropical forests and clarifies how the phantom-like order Zoraptera, ancient enough to predate the fragmentation of Gondwana, navigated the deep time between the age of dinosaurs and the present day.