Phylogeny
The aligned amino acid dataset (PCGs_faa) included 3,467 sites. The aligned nucleotide sequence datasets included 6,934 sites (PCG12_fna) and 9,218 sites (PCG12_fna+2rRNAs), respectively. A total of five phylogenetic trees were generated, including one Bayesian inference tree (maxdiff value = 0.11, minimum effective sizes>58) (Figure S1) and four maximum likelihood trees (Figure S2–S5). As expected, all phylogenetic analyses robustly supported the monophyly of Typhlocybinae with high nodal support values (PP=1, SH‒aLRT=100, UFB=100) (Figure 2). Unlike the ML trees (Figure 2, c, e), which support the subfamilies Cicadellinae and Typhlocybinae as sister group, the other three phylogenetic trees consistently support that the latter as sister to Mileewinae (Mileewini) (Figure 2, a, b, d). However, these alternative phylogenetic arrangements received low to moderate nodal support. In tribal relationships of Typhlocybinae, the sister-group pairs Dikraneurini+Erythroneurini (Figure 3, b) and Alebrini+Empoascini (Figure 3, e) were recovered with moderate to strong nodal support. However, the results differed in the placement of Typhlocybini, with some analyses placing this tribe sister to Alebrini+Empoascini (Figure 2, a, b) and others placing Typhlocybini sister to Dikraneurini+Erythroneurini (Figure 2, c, d, e). The included representatives of Zyginellini (Figure 3, in bold) were never recovered as a monophyletic group but were derived from within Typhlocybini, consistent with the five tribe classification adopted by some previous authors (Ahmed, 1983; Dietrich, 2013a; Zhou et al., 2020).Zyginella , the type genus of Zyginellini, and Limassollawere consistently placed as sister to the rest of Typhlocybini with strong nodal support, but other included genera of Zyginellini (Yangisunda and Paraahimia ) occupied more derived positions within Typhlocybini with low to moderate nodal support (Figure 3). In addition, the relationships of several deep internal nodes within Typhlocybini were unstable (Figure S1–S5).
Statistical tests of alternative tree topologies rejected hypotheses a (sister group relationship of Cicadellinae and Typhlocybinae) and b (sister group relationship of Typhlocybini and Zyginellini) but failed to reject either alternative hypothesis of the relationship of Typhlocybini to the other tribes (Table 2). Thus, the placement of Typhlocybini must be considered equivocal according to our results. The recent analysis of Lu et al. (2021) based on sequence data from 5 genes recovered Typhlocybini as sister to Dikraneurini+Erythroneurini but with only low to moderate nodal support. All analyses consistently supported the monophyly of the four tribes for which more than one exemplar was included: Dikraneurini, Empoascini, Erythroneurini, Typhlocybini (sensu lato, including Zyginellini). Zyginellini was consistently polyphyletic with the four included genera each representing an independent branch (Figure 3). Many branches pertaining to relationships within tribes were consistently well supported in all analyses but some were unresolved, particularly within Typhlocybini; for example, relationships among four major lineages of Typhlocybini were resolved inconsistently across analyses and form a polytomy in the Bayesian consensus tree (Figure 3).
The phylogenetic results, combined with evidence from morphological characters (see below) and genetic distances (Table S3) support the establishment of a new genus, Subtilissimia Yan & Yanggen. nov. , including two new species: Subtilissimia fulva Yan & Yang sp. nov. and Subtilissimia pelliculaYan & Yang sp. nov. , but suggest that the speciesFarynalasinistraYan & Yang, 2017 and Farynala dextra Yan & Yang, 2017, which have the male aedeagus identical but with processes curved in opposite directions (Yan & Yang, 2017, Figs 13–32), should be treated as synonyms (p-distance < 0.7%, except COX3 gene),Farynala sinistra Yan & Yang, 2017 syn. nov. The previously sequenced species identified as “Typhlocyba sp.” (GenBank, KY039138) belongs to the genus Aguriahana Distant, 1918 (COX1 p-distance=1.2%; PP=1, SH‒aLRT=100, UFB=100). Based on previous study (Qin et al., 2015; Yu et al., 2015) and analyses of phylogeny and genetic distance, we also show that the species previously identified as “Empoasca vitis (GenBank, NC_024838)” probably equals Matsumurasca onukii (COX1 p-distance = 0.3%). Other recent studies have shown that the latter species, a major pest of tea, has been widely misidentified in China (Qin et al., 2015). Unfortunately, we were unable to check voucher specimens from previous studies in order to confirm the identifications suggested by our results.
Ancestral charac ter state reconstruction
Eight typhlocybine morphological characters were traced over the BI consensus tree and ML tree (PCG12_fna). Overall, however, these reconstructions indicate that some characters used previously to diagnose tribes of Typhlocybinae exhibit considerable homoplasy (Figure 4; Figure S6). Ocelli were retained by Alebrini and Empoascini and lost in the common ancestor of Dikraneurini, Erythroneurini and Typhylocybini. However, a reversal to the ancestral state occurred in Typhylocybini corresponding to the genus Hiratettix (Figure 4, a). The forewing appendix is present in most non-typhlocybine leafhoppers but absent in all Typhlocybinae except Alebrini; thus Alebrini has traditionally been viewed as the most “primitive” group of Typhlocybinae (Figure 4, b). However, the recovered sister-group relationship between Alebrini and Empoascini suggests that the appendix may either have been lost independently in Empoascini and the remaining tribes or lost in the common ancestor of Typhlocybinae and regained in Alebrini. The bifurcation of hind wing vein CuA vein, shared with the outgroup, was lost independently in Empoascini and Typhlocybini (Figure 4, f). The distal part of the hind wing submarginal vein (Figure 4, e) was lost independently in Erythroneurini and Typhlocybini and hind wing veins RP and MA (Figure 4, g) became confluent and separated multiple times independently within Typhlocybini. The unique condition of the submarginal vein connecting to CuA directly traditionally used to diagnose Zyginellini as a distinct tribe was acquired independently in three different lineages of Typhlocybini (sensu lato) (Figure 4, h). Except for the single derivation of the latter “Zyginellini” character state in the analysis of Lu et al. (2021) our results are consistent with this prior study.