3-D imaging reveals four extraordinary cases of convergent evolution of acoustic communication in crickets and allies (Insecta)

When the same complex trait is exhibited by closely related species, a single evolutionary origin is frequently invoked. The complex stridulatory apparatus present in the forewings of extant crickets, mole crickets, katydids, and prophalangopsids, is currently interpreted as sharing a single common origin due to their similarity and unique function. An alternative hypothesis of convergent evolution in these ensiferan groups has challenged this common view, but remained controversial because of competing interpretations of wing venation. Here we propose another hypothesis for the widely and long debated homology of ensiferan stridulatory apparatus, performing the first 3D reconstruction of hidden structures at the wing bases. This approach allowed defining the homology of each vein from its very origin rather than after its more distal characteristics, which may be subjected to environmental pressure of selection. The stridulatory apparatus involves different veins in these four singing clades. In light of the most recent phylogenetic evidence, this apparatus developed four times in Ensifera, illustrating extraordinary convergent evolutions between closely related clades, by far exceeding the number of evolutionary steps ever proposed for calling ability in this group.

basivenal sclerites, and most of the crossveins that were visible on tomograms although these were not essential for the study; therefore, the edges of wings generally do not appear, unless they are evidenced by a rib.
In order to complete the 3D modeling, each specimen was observed under a binocular  between the two sheets is neutral 6 . The convexity of a vein can vary in its path.
Wing: flight body composed of two layers of epidermal cells forming upper and lower layers. It can be thin, membranous or thick, discolored, transparent, or brightly colored, covered with setae or scales. The wing is traversed by an array of ribs and tracheae 6 .

Phylogeny of the Ensifera
We follow the molecular phylogeny of Ensifera of Song et al. 8  broad basivenal sclerite ScB+RB that also gives birth to ScP. ScP is anteriorly pectinate and strongly convex at its base (contra the situation present in the great majority of the insects).
Between CuP and M+CuA, basal to the fork CuPa -CuPb, a longitudinal structure is observable in the thickness of the wing membrane. It is very thin, looking like a vein but without any relief. This structure does not appear on the tomograms. It is probably a trachea.  Gorochov, 1987 (Fig. 5a) The Triassic male Mesoedischia obliqua Gorochov, 1987 has a concave stridulatory file (see Fig. 5a), emerging from a convex vein that also emits a convex anterior branch at the base of the file. If it was the vein CuPb, it should have been concave and not convex. Thus we propose to interpret it as the vein A1 that has a concave broad posterior branch that bears the file and a convex anterior branch. The concavity of the vein interpreted as the stridulatory file is coherent with this function as the file has to be rubbed by the other tegmen. The branches of CuP are also relatively concave. As Mesoedischia shares a complex apomorphic character with the Tettigoniidae, it probably belongs to the stem group of the Tettigonioidea.

Mesoedischiidae
Mesoedischia differs from the modern Tettigoniidae in the presence of a well-developed CuP 9 with branches CuPa and CuPb (plesiomorphic state of character). CuPb is concave, CuPa is basally fused with M+CuA and re-emits its branches distally. (Fig. 5b, Supplementary Fig. 3)  (Fig. 2) Gu et al. 13 described Liassophyllum caii from the Middle Jurassic of China, on the basis of a series of specimens. These authors considered it belongs to the Haglidae Cyrtophyllitinae. The holotype is a well-preserved tegmen. New photographs, kindly provided by our colleague Jun Jie Gu, clearly show that between CuPa and the file, there is a short vein that has the same base as CuPa, runs towards the file, touches it and vanishes distally in the area between CuPa and the file. This vein typically corresponds to the weak and short CuPb as observed in modern true crickets. Thus, the file is on the vein A1. The situation is quite similar to that of the Grylloidea and the Liassic fossil described above. This tegmen is of great interest because the structure of the distal part of the stridulatory structure is of prophalangopsid type, unlike in the Liassic fossil described above and in the modern Grylloidea. It confirms that some taxa of the fossil record present character states homologous to that of the prophalangopsid-type and other homologous to the grylloid-type of stridulatory apparatus.