Phylogeny and classification of jumping plant lice of the subfamily Liviinae (Hemiptera: Psylloidea: Liviidae) based on molecular and morphological data

Abstract

Using molecular (COI, Cytb, H3, wg, 12S, 16S, and 28S) and morphological data (61 characters of adults and immatures), the phylogenetic relationships of the 20 nominal genera of Liviinae were analysed, and the monophyly of the subfamily was tested relative to the other two subfamilies of Liviidae. The analyses of the molecular, morphological, and combined datasets provided similar results with a strong or moderate support in the molecular and combined analyses for the subfamilies of Liviidae and two clades in Liviinae given tribal rank (Liviini and Paurocephalini stat. rev.). Three of the six previously recognized genera were confirmed as monophyletic (Aphorma, Camarotoscena, and Livia) and three were shown to be polyphyletic: Diclidophlebia s.l. (Diclidophlebia s.s., Haplaphalara stat. rev., Klyveria Burckhardt, Serbina and Malenovský, gen. nov., Melanastera Serbina, Malenovský, Queiroz and Burckhardt, gen. nov. and Woldaia stat. rev.), Paurocephala s.l. (Liella Burckhardt, Serbina and Malenovský, gen. nov. and Paurocephala s.s.), and Syntomoza s.l. (Anomoterga stat. rev. and Syntomoza s.s.). The new generic definitions require 38 new and eight revived combinations, including the transfer of two species to other families. All genera of Liviinae are diagnosed and keys are provided for their identification. A checklist of world Liviinae is supplemented.

INTRODUCTION

Jumping plant lice or psyllids are small phytophagous insects comprising slightly over 4000 described species worldwide. They are characterized by narrow host ranges, particularly as immatures (Burckhardt et al. 2014, 2021, Ouvrard et al. 2015). Several species are serious pests of cultivated plants such as Bactericera cockerelli (Šulc, 1909) and Diaphorina citriKuwayama, 1908, which are vectors of plant pathogens responsible for the Zebra Chip disease of potato and tomato, and for HLB of citrus (= huanglongbing, greening), respectively (Burckhardt 1994, Hollis 2004, Bové 2006, Hodkinson 2009, Munyaneza 2012). Other species are used to control invasive weeds. The Australian Boreioglycaspis melaleucae Moore, 1964 was introduced into Florida where it now successfully controls Melaleuca quinquinervia (Myrtaceae) (Center et al. 2012). Often related psyllid species develop on closely related host species and, thereby, create a pattern suggestive of cospeciation. Several studies have addressed this aspect but failed to find evidence for cospeciation between psyllids and their hosts (Burckhardt and Basset 2000, Percy 2003, Percy et al. 2004). At a lower taxonomic level, psyllids tend to switch between related plant species; at a higher taxonomic level, however, saltationary host-switching events explain better the wide but patchy distribution of hosts throughout the angiosperm phylogeny (Ouvrard et al. 2015). Three factors, in particular, complicate the study of psyllid host patterns. (i) Host information is lacking for about a third of the described psyllid species and some published information is not trustworthy (Burckhardt et al. 2014, Ouvrard et al. 2015). (ii) It is estimated that the 4000 described species represent at most half of the globally existing psyllid diversity (Burckhardt and Queiroz 2020). Extensive material of new taxa preserved in some psyllid collections contain, apart from many undescribed species, new genera and higher taxa; also previously unknown host relationships (unpublished BMHN, DMPC, MHNG, and NHMB data [cf. Material and methods, Repositories for explanations of acronyms]). (iii) Despite significant recent progress in understanding the phylogenetic relationships within the superfamily (Percy et al. 2018, Cho et al. 2019, Burckhardt et al. 2021), a number of open questions remain unanswered regarding the monophyly of, and the relationships between, certain taxa. This partly reflects the high degree of morphological homoplasy in psyllids, responsible for the long history of shifting psyllid classifications (Bekker-Migdisova 1973, White and Hodkinson 1985, Burckhardt and Ouvrard 2012, Ouvrard et al. 2013, Drohojowska 2015, Burckhardt et al. 2021), and partly reflects the challenge in obtaining sufficient resolution from molecular analyses (Percy et al. 2018, Cho et al. 2019). Examples for the former are provided by Ciriacreminae Enderlein 1910, erected for genera with a cross-vein or point of contact between forewing veins Rs and M_{1 + 2}, and Carsidarinae Crawford 1911, erected for genera with an anteriorly cleft head. Subsequent circumscriptions of the two subfamilies vary considerably between authors (Heslop-Harrison 1958, Bekker-Migdisova 1973, Hollis 1976, 1987, White and Hodkinson 1985, Li 2011, Burckhardt et al. 2021) as a consequence of different interpretations of the significance of these characters. Homoplasy is widespread among different groups of organisms (Brandley et al. 2009, Al Sayad and Yassin 2019, Ward and Boudinot 2021), and the extent of homoplasy in morphological characters influences the ease with which a morphology-based phylogeny can be inferred (Murphy et al. 2021).

Here, we examine the phylogenetic relationships within the Liviinae, one of three subfamilies of Liviidae. Currently the subfamily comprises 145 species (Ouvrard 2023). Some of these are minor pests of commercially valuable tropical timber species, while others are considered as biological agents for controlling invasive plants, such as Diclidophlebia lucensBurckhardt et al., 2005 and D. smithi developing on Miconia calvescens, a devastating weed and one of the most serious threats to rainforest ecosystems of the Pacific islands (Burckhardt et al. 2005, Morais et al. 2013).

The subfamily Liviinae was erected by Löw (1879) for the single genus LiviaLatreille, 1802. This concept was adopted by most subsequent authors, including Bekker-Migdisova (1973), who treated the taxon as a family following Vondráček (1957). An exception is Crawford (1914) who added the genera Aphalara Foerster, 1848, AphalaroidaCrawford, 1914, and Rhinocola Foerster, 1848 to the subfamily, an act that was not followed by other authors. Bekker-Migdisova (1973) suggested that the Liviidae is phylogenetically related to the Aphalaridae ‘and in particular to Paurocephalini and Rhinocolini’, two of the five tribes of Paurocephalinae. White and Hodkinson (1985) included the monotypic Liviinae in the Aphalaridae along with seven other subfamilies, among which is the Paurocephalinae. In the cladogram of White and Hodkinson (1985: fig. 185), Livia is the sister-taxon of Strophingia Enderlein, 1914 [Strophingiinae (=Euphyllurinae)]. Burckhardt (1987) synonymized the Aphalaridae and Spondyliaspididae of White and Hodkinson (1985) with Psyllidae, and Burckhardt (2005) subdivided this large artificial assemblage into three putatively monophyletic lineages, among which is the ‘liviid assemblage’, a taxon given family status by Burckhardt and Ouvrard (2012). Within this family, they recognized the two subfamilies: Euphyllurinae and Liviinae. The latter included Livia and five genera (AphormaHodkinson, 1974, CamarotoscenaHaupt, 1935, DiclidophlebiaCrawford, 1920, PaurocephalaCrawford, 1913, and SyntomozaEnderlein, 1921) referred to Paurocephalinae by Burckhardt and Mifsud (2003) who defined the genera based on a cladistic analysis of morphological characters. Their concept of Aphorma and Paurocephala followed that of Burckhardt and Bänzinger (1995) and Mifsud and Burckhardt (2002), respectively. Two species referred by Loginova (1975) to Camarotoscena were transferred to Syntomoza together with species of AnomotergaKlyver, 1932 and HomalocephalaYang and Li, 1986 which they synonymized with Syntomoza. Burckhardt and Mifsud (2003) further synonymized the monotypic genera AconopsyllaTuthill and Taylor, 1955, ParaphalaroidaLoginova, 1972, SinuonemopsyllaLi and Yang, 1991, and WoldaiaBrown and Hodkinson, 1988, as well as the poorly defined HaplaphalaraUichanco, 1921 with the equally poorly defined Diclidophlebia. They admitted that the genus in this broad definition ‘is diverse with respect to morphology (head, forewing, hindlegs)’ in contrast to the homogeneous morphology of immatures. Navasero and Calilung (1998) transferred Paurocephala kleinhofiaeUichanco, 1921 to Anomoterga on the basis of the presence of tubercles on the abdominal tergites 5–7, and Navasero (2010) erected the genus PaurotergaNavasero, 2010 for this and two other species from the Philippines. He suggested also that the genera Anomoterga and Homalocephala may be synonyms. MarpsyllaNavasero and Calilung, 2001, another genus erected by Navasero and Calilung (2001) for three Paurocephala species from the Philippines, was defined mostly by the short antennae with less than 10 segments and the up-turned female proctiger. Marpsylla and Pauroterga were synonymized with Paurocephala by Burckhardt and Ouvrard (2012).

In the molecular analyses of Percy et al. (2018), the monophyly of the Liviinae of Burckhardt and Ouvrard (2012) was strongly supported, whereas Euphyllurinae was polyphyletic. In that study (Percy et al. 2018), Diaphorina Löw, 1880 and KatacephalaCrawford, 1914 grouped together with Calophyidae, Psyllidae, and Triozidae, whereas Euphyllura Foerster, 1848, PsyllopsisLöw, 1879, and Strophingia formed a strongly supported monophyletic group related to Liviinae and Neophyllura Loginova, 1973. Depending on the dataset analysed, the three clades together constituted a weakly supported monophylum (all-nucleotide mitogenome dataset) or a paraphylum (conserved-codon mitogenome dataset, combined mitochondrial–nuclear dataset). In the most recent psyllid classification, the three groups are treated as subfamilies Euphyllurinae, Liviinae, and Neophyllurinae of the family Liviidae (Burckhardt et al. 2021).

The monophyly of the six genera of Liviinae currently recognized has not been tested with molecular data but the analyses of Percy et al. (2018) suggest that Diclidophlebia sensuBurckhardt and Mifsud (2003) is polyphyletic. Here, we (i) analyse the phylogenetic relationships within the Liviinae using molecular and morphological data from a larger number of taxa than those used in the previous studies, (ii) compare the results from molecular, morphological and combined data, (iii) define and diagnose tribes and genera, and (iv) discuss some aspects of the resulting biogeographic and host patterns.

MATERIAL AND METHODS

Repositories

Specimens used in this study are dry or slide mounted, or preserved in 70% or 96% ethanol. They are deposited in the following collections and institutions: BMNH—Natural History Museum, London, UK; BPBM—Bernice Pauahi Bishop Museum, Honolulu, USA; DMPC—Diana M. Percy, private collection; MHNG—Muséum d’histoire naturelle, Geneva, Switzerland; MMBC—Moravian Museum, Brno, Czech Republic; and NHMB—Naturhistorisches Museum, Basel, Switzerland.

Taxon sampling

For the analyses, species were selected with the aim to represent the generic and morphological diversity within Liviinae. As an outgroup for the phylogenetic analyses, we used Trioza urticae (Linnaeus, 1758) (Triozidae), a member of the putative sister-clade (‘PTCD’ clade) of Liviidae in the best maximum likelihood tree generated using the all-taxa/all-nucleotide mitogenome dataset of Percy et al. (2018). Outgroup and other sequences not generated during this study were obtained from GenBank. In the molecular analyses, we included 45 species of Liviidae (Euphyllurinae eight spp., Neophyllurinae one sp., Liviinae 36 spp.) of which 13 type species of the 20 previously described nominal genera. Not included were Anomoterga homali (Yang and Li 1986) (type species of Homalocephala), Aphorma jiuzhaiensis (Li 2011) (type species of LeprostictopsyllaLi, 2011), Diclidophlebia oceanica (Crawford 1919) (type species of Diclidophlebia), Haplaphalara sterculiae (Froggatt 1901) (type species of Aconopsylla), Livia limbata (Waga 1842) (type species of NeoliviaHedicke, 1920), Paurocephala baltazarae (Navasero and Calilung 2001) (type species of Marpsylla), and Paurocephala zamboangensis (Navasero 2010) (type species of Pauroterga) for which no fresh or recently collected material was available. Voucher specimens were mounted on permanent slides in Canada balsam or are kept in 70% ethanol. They are deposited in BMNH, DMPC, MMBC, and NHMB. Complete voucher data are given in the Supporting Information, Table S1.

For the morphological analyses, we used the same taxa as in the molecular analyses and additional 11 species of Liviinae, including the type species of all 20 nominal liviine genera. The material used for the morphological analyses and illustrations (Figs 1–9) is detailed in the Supporting Information, Table S2.

DNA extraction and alignment

DNA extractions from single individuals were performed using a ‘DNeasy Blood and Tissue Kit’ and an ‘EZNA@Tissue DNA Kit’, according to the manufacturers’ protocols with minor modifications. Four mitochondrial (COI, Cytb, 12S, and 16S rRNA) and three nuclear (H3, wg, and 28S rRNA) gene regions were amplified using PCRBio HiFi Polymerase. These genes were successfully used for phylogenetic reconstructions at species to family levels in many groups of insects, including psyllids (Thao et al. 2000, Cryan and Svenson 2010, Cryan and Urban 2012, Hall et al. 2016, Percy 2017, Tkoč et al. 2017, Martoni et al. 2018, Cho et al. 2019, Kaspřák et al. 2019).

The primer sequences, amplicon sizes, and conditions for polymerase chain reaction (PCR) used for the amplification are summarized in Table 1. The dry-mounted material from museum collections had been stored for up to 28 years before DNA extraction (the oldest samples from which we successfully amplified PCR products were collected in 1990). To enhance the yield of DNA, we worked with small elution volumes (e.g. 10–20 μL) following the instructions of Suchan et al. (2016) for specimens with a low DNA concentration, small specimens, and dry-mounted specimens from museum collections. To increase DNA concentration, we also employed nested PCR amplification providing large gene fragments (this was crucial for 12S and 16S rRNA) and genes present at low abundance. Nested PCR involved two sequential amplification reactions, each of which used a different primer set (Table 1). The first set of primers (outer primers) aimed to increase the amount of non-specific PCR templates by means of a high number of cycles, a wide annealing temperature range, and an increased annealing time in the first PCR round. Amplicons resulting from the first PCR round were then used as a template for the second PCR round employing a second set of primers (inner primers), or a combination of outer and inner primers. The second PCR round targeted a smaller region of the amplicon and thus lowered the risk of amplification of non-target sequences from the first PCR round. To increase the specificity and yield in the second PCR round, a lower number of cycles, higher annealing temperature, and lower annealing time were employed to amplify only the target template segment.

Amplified products were purified using ExoSAP-IT Express (Applied Biosystems) and sequenced in both directions with a BigDye Terminator v.3.1 Sequence Kit (Applied Biosystems). Purified PCR products were sequenced on an AML-part Prism 3130 Genetic Analyzer (Applied Biosystems). Sequence fragments were assembled using SEQUENCHER v.4.8 (http://www.genecodes.com). The protein-coding gene sequences (COI, Cytb, H3, and wg) were aligned manually with the aid of MEGA X (Kumar et al. 2018). The non-protein coding gene sequences (12S, 16S, and 28S rRNA) are rich in indels and, therefore, were aligned using MAFFT v.7 via the online web-server (https://mafft.cbrc.jp/alignment/server/index.html) (Katoh et al. 2019). The G-INS-I strategy was applied for the 16S and 28S rRNA gene regions, and the Q-INS-I strategy was chosen for the 12S rRNA gene region. GBlock v.0.91b, with the options of a least stringent selection, was employed to remove ambiguously aligned blocks in the non-protein coding gene regions (Talavera and Castresana 2007). As a result, 45.8% of nucleotides were removed from the original sequences of 16S rRNA, 22.5% from 28S rRNA, and 50.7% from 12S rRNA gene regions. All gene sequences were combined using SEAVIEW v.3.2 (Gouy et al. 2010) (Supporting Information, File S1) and were submitted to GenBank (Supporting Information, Table S1).

Molecular phylogenetic analyses

Prior to the phylogenetic analyses of the molecular data, the best partitioning scheme was identified using PartitionFinder2 (Lanfear et al. 2016), and the concatenated dataset was divided into 10 partitions by gene regions and codon position, allowing each partition to have its own evolutionary rate. In total, 46 sequences of 3169 bp length were used in the analyses. Setting of codon position for protein coding gene fragments was performed with MESQUITE v.3.61 (Maddison and Maddison 2019).

A maximum likelihood (ML) analysis was carried out using IQ-TREE v.1.6.12 (Chernomor et al. 2016, Trifinopoulos et al. 2016, Crotty et al. 2019). A suitable substitution model for each partition was calculated by IQ-TREE. In addition to the ML standard partition analysis (thereafter, ML-part), we ran a mixture (heterotachous) model analysis (thereafter, ML-mix) (Fig. 10) using IQ-TREE command line allowing an independent calculation of substitution models for any sequence fragment of the dataset executed with the model option –m GTR+H4. For both ML analyses, nodal support was evaluated using a standard non-parametric bootstrap with 1000 replicates.

For the Bayesian inference (BI) analysis, the best substitution model for each partition was estimated using jModelTest v.2.1.10 (Darriba et al. 2012) and a comparison of scores from Akaike information criterion (AIC) and Bayesian information criterion (BIC) (Supporting Information, Table S3). The analysis was run using MrBayes v.3.2.7a (Ronquist et al. 2012) on the CIPRES platform (Miller et al. 2010) conducting two independent runs each with four Metropolis-coupled Markov chains for 15 million generations, sampling every 1000th generation; priors for tree topology and speciation rates were set to uniform, gamma distribution category count set to 4, and probability distribution on branch lengths set to exponential. A Dirichlet prior was used to cover substitution rates of the evolutionary models, as well as stationary nucleotide frequencies. Priors of all other parameters were kept at the default values. Nodal support was assessed by posterior probabilities after 30% of the samples were discarded as burn-in. The ML and BI trees were visualized using iTOL v.6 (Letunic and Bork 2021).

Morphological phylogenetic analyses

A total of 61 morphological characters (42 binary, 19 multistate) of adults and immatures were selected for the analyses and encoded (Table 2). The character matrix was assembled with WINCLADA v.1.00.08 (Nixon 2002) (Supporting Information, File S2). Two parsimony analyses (unweighted and weighted) were performed with TNT v.1.5 (Goloboff and Catalano 2016) using the ‘Traditional search’ method with the following settings: random seed = 0, replications = 1000, swapping algorithm = TBR with 1000 trees to save per replication. The matrix was analysed with equal and implied weights, the latter with the weight function K = 10.0000. For constructing consensus trees the option ‘Strict (=Nelsen)’ was chosen. The character state changes are shown for each node of the strict consensus tree (Fig. 11), using the ‘Unambig changes only’ option of WINCLADA.

Combined phylogenetic analyses

Phylogenetic analyses of a combined molecular and morphological dataset were performed for those taxa (45 species) for which both molecular and morphological data were available. The combined matrix was analysed with IQ-TREE v.1.6.12 and MrBayes v.3.2.7a on the CIPRES platform (see above for references). The ML standard partition and BI analyses were run under the same substitution models and parameters defined in the section ‘Molecular phylogenetic analyses’ (Supporting Information, Table S3). The morphological characters in both ML and BI analyses of the combined dataset were parameterized under the standard discrete model proposed by Lewis (2001).

Conventions

When discussing the nodal support in the molecular analyses, the terms ‘strong’, ‘moderate’, and ‘weak’ are used. For bootstrap values the respective ranges are >90%, 70–90%, and 50–69%, and for the posterior probabilities (PP) >0.95, 0.80–0.95, and 0.50–0.79. With reference to the morphological data, we refer to homoplasy and homoplastic characters in a blanket sense to cover all apomorphic character states that appear more than once on a particular tree topology. The phylogenetic information contained in such homoplastic characters may still be valid (indicated by consistency and retention indices; Speed and Arbuckle 2017) when shown to be shared within some clades (e.g. symplesiomorphies or synapomorphies), and, therefore, their status as homologies remains unfalsified (Begun 2007). Clearly, the extent of homoplasy revealed in our morphological data could be subject to additional analyses to further evaluate the phylogenetic significance of individual homoplastic characters (e.g. Sansom and Wills 2017, Al Sayad and Yassin 2019, Tomaszewska et al. 2021), but at this stage this is not possible given the many nodes in our molecular phylogenies that remain poorly supported.

The morphological terminology mostly follows Halbert and Burckhardt (2020), and that of the mesosternum follows Drohojowska (2015) (see also Fig. 5).

The biogeographical realms are defined as follows: Afrotropics (Afr): sub-Saharan Africa, Madagascar, and Arabian Peninsula;—Australasia (Aus): Australia, New Guinea, New Zealand, and neighbouring islands;—Indomalaya (Ind): Indian subcontinent, South-East Asia, and southern China;—Nearctic (Nea): North America north of Mexico;—Neotropics (Neo): South and Central America and Mexico;—Oceania (Oce): Polynesia (except New Zealand), Micronesia, and the Fijian Islands;—Palaearctic (Pal): temperate Eurasia and North Africa.

Plant names and family concepts correspond to POWO (2023).

RESULTS

Taxonomic impact

On the basis of the phylogenetic analyses, we propose here a new tribal and generic classification that differs from that of Burckhardt and Mifsud (2003) and Burckhardt et al. (2021). In the following, we will use the revised generic names as defined below in the section ‘Taxonomic account’. When referring to other definitions, this will be specified by ‘sensu’ and the respective literature source. Here, we recognize the two tribes Liviini and Paurocephalini stat. rev. and the 12 following genera: Anomoterga stat. rev., Aphorma, Camarotoscena, Diclidophlebia s.s., Haplaphalara stat. rev., Klyveria Burckhardt, Serbina and Malenovský gen. nov., Liella Burckhardt, Serbina and Malenovský gen. nov., Livia, Melanastera Serbina, Malenovský, Queiroz and Burckhardt gen. nov., Paurocephala s.s., Syntomoza s.s., and Woldaia stat. rev.. We provide a checklist of all available species names of Liviinae (Supporting Information, File S3) including the resulting new and revived combinations.

Phylogeny based on molecular data

Both ML analyses (ML-mix, Fig. 10, and ML-part, Supporting Information, File S4A) and Bayesian inference (BI, Supporting Information, File S4B) strongly support the monophyly of the Euphyllurinae, Liviinae, Liviini, and Paurocephalini, and strongly (ML-part, BI) or weakly (ML-mix) corroborate a sister-group relationship of Liviinae and Neophyllurinae.

Within Liviini, in all analyses, Syntomoza sensuBurckhardt and Mifsud (2003) is polyphyletic. In ML-mix, the grouping Syntomoza magna (Kuwayama, 1908) + (Aphorma + Livia) is weakly supported; in ML-part and BI, Livia and Syntomoza magna are moderately supported as sister-taxa. Anomoterga constitutes a clade with Camarotoscena (ML-mix) or with Aphorma and Camarotoscena (ML-part, BI). In all analyses, the monophyly of Camarotoscena and Livia is strongly supported.

Within the Paurocephalini, Diclidophlebia sensuBurckhardt and Mifsud (2003) and Paurocephala sensuBurckhardt and Mifsud (2003) are polyphyletic in all analyses, while the monophyly of Klyveria, Liella, and Paurocephala s.s. and the sister-group relationship of Klyveria + Woldaia are strongly supported. The monophyly of Melanastera is moderately (ML-part, BI) or weakly (ML-mix) supported. The grouping of the remaining species of Diclidophlebia sensuBurckhardt and Mifsud (2003), i.e. Diclidophlebia s.s. and Haplaphalara, differs considerably between the analyses, except for Haplaphalara irvingiae (Burckhardt et al. 2006), comb. nov. and Haplaphalara Laos, which are always strongly supported as sister-species. This species pair represents the sister-group to all remaining Paurocephalini in ML-mix, while it forms a strongly supported clade together with Haplaphalara dahli (Rübsaamen, 1905), comb. rev. in ML-part and with H. dahli and H. Madagascar sp. 1 in BI.

Phylogeny based on morphological data

The cladistic analysis using equal weights resulted in 539 most-parsimonious trees of 197 steps with a consistency index of 0.45 and a retention index of 0.84. The strict consensus tree is shown in Fig. 11. An analysis with implied weights (K = 10.0000) resulted in the same tree topology without improved resolution. It is not discussed further here.

In the morphological analysis (Fig. 11), the monophyly of neither Euphyllurinae nor Liviinae is supported, and Neophyllura arbuti (Schwarz, 1904) (Neophyllurinae) is nested within a clade of Euphyllurinae. The Liviini and Paurocephalini are monophyletic, each supported by two and three synapomorphies, respectively.

Within the Liviini, Syntomoza sensuBurckhardt and Mifsud (2003) is polyphyletic. One or more synapomorphies support the monophyly of each of the genera Anomoterga, Aphorma, and Livia, as well as the sister-group relationships of Aphorma and Livia, and the two together with Syntomoza s.s.. The morphological support of Camarotoscena is weak with two homoplastic characters that we consider to be non-unique synapomorphies (character 18: state 2 and character 56: state 1, which are both unique in Liviini).

Within the Paurocephalini, Diclidophlebia sensuBurckhardt and Mifsud (2003) and Paurocephala sensuBurckhardt and Mifsud (2003) are both monophyletic. Whereas the former is supported by one synapomorphy (and five homoplasies), the latter is grouped by two homoplasies (character 16: state 1 and character 32: state 1), which we consider weak evidence for monophyly. The monophyly of the genera Diclidophlebia s.s., Klyveria, Melanastera, and Paurocephala s.s. is supported by at least one synapomorphy, as is the sister-group relationship of Klyveria and the monotypic Woldaia. Support for Liella is weak with only one homoplastic character that we consider to constitute a non-unique synapomorphy (character 21: state 0).

Phylogeny based on the combined dataset

The results of the ML analysis (ML-part, Supporting Information, File S4C) and BI (Supporting Information, File S4D) of the combined molecular and morphological dataset are similar in the strongly or moderately supported monophyly of the Euphyllurinae, Liviinae, Liviini, and Paurocephalini but differ in the position of the Neophyllurinae: unsupported sister-group relationship with Liviinae (ML-part) and strongly supported sister-group relationship with Euphyllurinae + Liviinae (BI).

Within Liviini, Syntomoza sensuBurckhardt and Mifsud (2003) is polyphyletic in both analyses. The sister-group relationship of Syntomoza magna with Aphorma + Livia is moderately (ML-mix) or strongly (BI) supported. Anomoterga + Camarotoscena together are monophyletic with moderate (ML-mix) or strong (BI) support, but the former is paraphyletic in both analyses.

Within the Paurocephalini, the monophyly of Klyveria and Paurocephala s.s. and the sister-group relationship of Klyveria + Woldaia are strongly supported in both analyses. The monophyly of Liella is strongly supported in BI and moderately in ML-mix, as is the sister-group relationship of Haplaphalara irvingiae + H. Laos. Melanastera is monophyletic with weak support in BI, but polyphyletic in ML-mix. The relationships of the other species are not sufficiently resolved.

Taxonomic account
Liviinae Löw, 1879,
Liviinae Löw 1879: 606.

Diagnosis:

Adult. Head lacking genal processes, sometimes with anterior lobes and/or with ventral part separated as distinct sclerites called the subgenae. Metacoxa bearing well-developed meracanthus. Metafemur with three ventral sense organs in medial or submedial position. Metatibia lacking a genual spine. Metabasitarsus without sclerotized spurs. Lateral edge of first visible abdominal tergite usually bearing a patch of small tubercles or spinules (absent in Livia) that differ from surrounding microsculpture, sometimes forming a finger-like process. Male proctiger one-segmented.—Fifth instar immature. Tarsus with claws; tarsal arolium bearing unguitractor.

Comments:

The monophyly of the subfamily is strongly supported in the molecular and combined analyses (Fig. 10; Supporting Information, File S4) but not in the morphological analysis (Fig. 11). The subfamily consists of the two tribes Liviini and Paurocephalini.

Liviini Löw, 1879,
Liviinae Löw 1879: 606.

=Camarotosceninae Li 2011: 381; synonymized with Liviinae by Burckhardt and Ouvrard 2012: 16.

Diagnosis:

Adult. Preocular sclerite developed as small sclerite between vertex and eye (Fig. 3B). Antenna always 10-segmented. Apex of distal portion of aedeagus not subdivided from stem.—Immature. Antenna seven or eight segments; flagellum without sectasetae. Tarsal arolium fan-shaped, sessile (Fig. 9K).

Comments:

The monophyly of Liviini is strongly supported in the molecular and combined (BI) analyses and moderately in the combined (ML-part) analysis (Fig. 10; Supporting Information, File S4). In the morphological analysis, the tribe is supported by two synapomorphies (and six homoplasies) (Fig. 11). It includes five genera.

AnomotergaKlyver, 1932, stat. rev.

AnomotergaKlyver 1932: 93. Type species: Anomoterga tahuataKlyver, 1932, by original designation and monotypy.

=HomalocephalaYang and Li 1986: 54. Type species: Homalocephala homaliYang and Li, 1986, by original designation and monotypy. Syn. nov.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1A); in dorsal view about as wide as thorax, moderately transverse. Vertex rhomboidal to almost subrectangular; covered in areolate-rugose or imbricate microsculpture, sometimes much reduced on disc; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons small trapezoidal; median ocellus clearly visible in perpendicular view to vertex; subgenae not differentiated into separate sclerites (Fig. 3C); compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, moderate-sized, flattened ventrally, hardly visible in lateral view as it is hidden by genae. Antenna slightly longer than head width; flagellum with simple setae; segment 3 longest, as long as or longer than segments 4–6 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender; dorsal outline, in lateral view, strongly curved (Fig. 1A). Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites subrectangular, divided by perpendicular suture into subequal epimeron and episternum. Metapostnotum medially flattened, or with blunt tubercle or longitudinal ridge (Fig. 4A). Mesosternum narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture hardly visible; basisternum indistinct; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse (Fig. 5A). Pro- and mesotibiae cylindrical. Metacoxa with blunt or subacute horn-shaped meracanthus (Fig. 6A). Metafemur with the three ventral sense organs in the middle; apex with few stout long setae. Metatibia longer than metafemur, slightly widened apically; bearing 9–11 slightly irregularly spaced apical sclerotized spurs and two to five peg-like setae adjacent to inner spurs. Both metatarsal segments relatively short, subequal in length. Forewing oblong-oval or subtrapezoidal; costal and anal margins subparallel or widening towards apex, 2.0–2.4 times as long as wide, membranous or subcoriaceous; vein C + Sc weakly, evenly or irregularly convex, slender, distinctly delimited from cell; costal break developed, sometimes indistinct, close to apex of vein R₁; pterostigma wide, entirely membranous; nodal line developed; veins R and M + Cu subequal; vein Rs almost straight or curved towards costal margin; vein M much longer than its branches; vein Cu_1a almost straight or curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules fine or coarse, spaced or dense, present in all cells. Hindwing slightly shorter than forewing; with one to three costal setae proximal to costal break and two clearly separated groups distal to costal break, of three to five proximal and three to seven distal setae; vein R + M + Cu bifurcating into R and M + Cu. Abdominal base with a weakly sclerotized area on either side covered in spines. Aedeagus with simple proximal portion bearing many weak folds subapically; apex of distal portion not differentiated from stem. Female subgenital plate lacking apical process.

Last instar immature. Antenna five to seven segments; lacking sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs with massive peg-like setae. Dorsal body surface lacking minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in ventral position; sometimes with additional pore fields developed.

Comments:

Supported by one synapomorphy in the morphological analysis (Fig. 11) but paraphyletic in the combined analyses, and paraphyletic (ML-mix) or polyphyletic (ML-part, BI) in the molecular analyses (Fig. 10; Supporting Information, File S4). Burckhardt and Mifsud (2003) synonymized Anomoterga and Homalocephala with Syntomoza but did not include Syntomoza magna, the type species of Syntomoza, in their analysis. Morphology suggests a close relationship of the type species of Anomoterga and Homalocephala with two species originally described in Camarotoscena and transferred by Burckhardt and Mifsud (2003) to Syntomoza. The four species together, however, are only distantly related to, and not congeneric with, S. magna. Similar results are obtained by the molecular and combined analyses, though the type species of Homalocephala was not included. We conclude that Anomoterga and Syntomoza are distinct genera, the latter being monotypic, and Homalocephala is a synonym of Anomoterga. Included available species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3. The following new or revived combinations are here proposed: Anomoterga africana (Loginova 1975: 55), comb. nov. (from Camarotoscena; Syntomoza, Burckhardt and Mifsud 2003: 16); A. homali (Yang and Li 1986: 54), comb. nov. (Homalocephala; Syntomoza, Burckhardt and Mifsud 2003 16); A. hsenpinensisFang and Yang 1986: 137, comb. rev. (Syntomoza, Burckhardt and Mifsud 2003: 16); A. scolopiae (Yang 1984: 23), comb. nov. (Syntomoza); A. tahuataKlyver 1932: 94, comb. rev. (from Syntomoza, Burckhardt and Mifsud 2003: 16); A. unicolor (Loginova in Loginova and Parfentiev 1958: 99), comb. nov. (Camarotoscena; Syntomoza, Burckhardt and Mifsud 2003: 17).

AphormaHodkinson, 1974

AphormaHeslop-Harrison 1949: 783, 800; Heslop-Harrison 1952: 965; nomen nudum, no type species designated.

AphormaHodkinson 1974: 76. Type species: Aphalara bagnalliLaing, 1929, by original designation.

=LeprostictopsyllaLi 2011: 376. Type species: Leprostictopsylla jiuzhaiensisLi, 2011, by original designation and monotypy. Synonymized by Burckhardt et al. 2018: 14.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1B); in dorsal view slightly narrower than thorax, moderately transverse. Vertex rhomboidal; covered in areolate-rugose microsculpture; separated from genae by transverse carina (Fig. 2D); coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons parallel-sided to narrowly triangular; median ocellus not visible in perpendicular view to vertex; subgenae differentiated into separate sclerites (Fig. 3D); compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, moderate-sized, rounded ventrally, hardly visible in lateral view as it is hidden by genae. Antenna about as long as head width; flagellum with simple setae; segment 3 longest, shorter than segments 4–5 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking a wreath of spines. Thorax moderately slender; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, almost straight, subrectagular; propleurites subrectangular, divided by perpendicular suture into subequal epimeron and episternum. Metapostnotum medially with blunt tubercle. Mesosternum narrower than head, subrectangular, less than twice as wide as long laterally; anterior margin weakly concave; pleurosternal suture hardly visible; basisternum large, oval to rhomboidal; katepisternum large antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with long or short subacute horn-shaped meracanthus. Metafemur with the three ventral sense organs in the middle; apex lacking stout long setae. Metatibia longer than metafemur, widened apically; bearing seven or eight slightly irregularly spaced apical sclerotized spurs and a few thorn-like setae postero-apically (Fig. 7B). Both metatarsal segments relatively short, subequal in length. Forewing oblong-oval or subtrapezoidal; costal and anal margins subparallel or weakly widening towards apex, 1.7–2.3 times as long as wide, coriaceous; vein C + Sc weakly strongly convex, slender, distinctly delimited from cell; costal break absent; pterostigma reduced (Fig. 8A); nodal line not visible; veins R as long as or shorter than M + Cu; vein Rs weakly sinuous; vein M much longer than its branches; vein Cu_1a strongly curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface covered in flattened tubercles forming cellular pattern (Fig. 8D). Hindwing about as long as forewing; with one to three costal setae proximal to costal break and two indistinct groups distal to costal break, three to four dense setae proximally, and three to four spaced setae distally; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a sclerotized area on either side covered in spines (Fig. 8L, M). Aedeagus with simple proximal portion bearing many weak folds subapically; apex of distal portion not differentiated from stem. Female subgenital plate lacking apical process.

Last instar immature. Antenna seven or eight segments; lacking sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface lacking minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; with additional pore fields developed.

Comments:

Represented by only one species in the molecular analyses (Fig. 10; Supporting Information, File S4); supported by one synapomorphy in the morphological analysis (Fig. 11). The synonymy of Leprostictopsylla with Aphorma by Burckhardt et al. (2018) is confirmed here. Included available species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3.

CamarotoscenaHaupt, 1935

CamarotoscenaHaupt 1935: 228. Type species: Rhinocola speciosaFlor, 1861, by original designation.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1C); in dorsal view, about as wide as thorax, moderately transverse. Vertex rhomboidal; covered in areolate-rugose microsculpture; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons elongate, parallel-sided to narrowly triangular (Fig. 3E); median ocellus not visible in perpendicular view to vertex; subgenae not differentiated into separate sclerites (Fig. 3E); compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, moderate-sized, rounded ventrally, hardly visible in lateral view as it is hidden by genae. Antenna slightly longer than head width; flagellum with simple setae; segment 3 longest, as long as segments 4–5 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites subrectangular, divided by perpendicular suture into subequal epimeron and episternum. Metapostnotum with blunt tubercle. Mesosternum (Fig. 5B) narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture hardly visible; basisternum triangular; katepisternum small antero-laterally, not bent dorsad laterally; angle of arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with blunt, moderately long, horn-shaped meracanthus. Metafemur with the three ventral sense organs in the middle; apex with few stout long setae (Fig. 6O). Metatibia longer than metafemur, distinctly widened apically; bearing 9–12 slightly irregularly spaced apical sclerotized spurs and two to four peg-like setae adjacent to inner spurs (Fig. 7C). Both metatarsal segments relatively short, subequal in length. Forewing slightly subtrapezoidal; costal and anal margins subparallel or widening towards apex, 2.2–2.3 times as long as wide, membranous; vein C + Sc weakly, evenly convex, slender, distinctly delimited from cell; costal break developed, close to apex of vein R₁; pterostigma wide, entirely membranous; nodal line developed; veins R and M + Cu subequal; vein Rs relatively straight; vein M about as long as M_{1 + 2}; vein Cu_1a almost straight or weakly curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} oblique to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules fine or coarse, spaced or dense, present in all cells. Hindwing slightly shorter than forewing; with one or two costal setae proximal to costal break and two clearly separated groups of two to four proximal and two to four distal setae, distal to costal break; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a sclerotized area on either side covered in spines. Aedeagus with simple proximal portion bearing many weak folds subapically; apex of distal portion not differentiated from stem. Female subgenital plate lacking apical process.

Last instar immature. Antenna seven segments; lacking sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface lacking minute clavate setae. Precaudal abdominal tergites bearing densely spaced simple setae and sectasetae. Anus in ventral position; with additional pore fields developed.

Comments:

Monophyletic with strong molecular support, morphologically supported by two homoplastic characters that we consider non-unique synapomorphies (Figs 10, 11; Supporting Information, File S4). The circumscription of the genus by Burckhardt and Mifsud (2003) is supported here. Included available species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3.

LiviaLatreille, 1802

LiviaLatreille 1802: 266. Type species: Psylla juncorumLatreille, 1798, by monotypy.

=DiraphiaIlliger 1803: 284. Type species: Chermes junciSchrank, 1789, by monotypy. Synonymized with Livia by Latreille 1807: 170.

=DiraphiaWaga 1842: 275, nec Illiger 1803. Type species: Diraphia limbataWaga, 1842, by monotypy.

=NeoliviaHedicke 1920: 71; replacement name for Diraphia Waga nec Illiger. Synonymized with Livia by Loginova 1974: 862.

=VailakiellaBliven 1955: 13. Type species: Vailakiella eosBliven, 1955, by original designation and monotypy. Synonymized with Livia by Hodkinson and Bird 2000: 4.

Diagnosis:

Adult. Head, in lateral view, directed in longitudinal axis of body (Fig. 1D); in dorsal view slightly narrower than thorax, about as long as wide (Fig. 2A). Vertex elongate, forming a lobe on either side of midline, continuing to ventral face; covered in irregularly imbricate, partly reduced microsculpture; clearly separated from genae by transverse suture (Fig. 3F); coronal suture fully developed; genae flattened, not enlarged into processes; frons narrowly triangular (Fig. 3F); median ocellus not visible in perpendicular view to vertex; subgenae differentiated into separate sclerites (Fig. 3F); compound eyes, in dorsal view, elongate, strongly adpressed to head. Clypeus pear-shaped, moderately large, flattened ventrally, hardly visible in lateral view as it is hidden by genae. Antenna about as long as head width; flagellum with simple setae; segment 2 longest, segment 3 shorter than segments 4–6 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender; dorsal outline, in lateral view, flattened. Pronotum in dorsal view, almost straight, subrectagular; propleurites subrectangular, divided by perpendicular suture into subequal epimeron and episternum. Metapostnotum (Fig. 4B) forming backwards directed tooth. Mesosternum (Fig. 5D) relatively long; pleurosternal suture and basisternum well developed; mesosternum narrower than head, subrectangular less than twice as wide as long laterally; anterior margin with medial angle or notch; pleurosternal suture well visible; basisternum large, oval to rhomboidal; katepisternum large antero-laterally, not bent dorsad laterally; angle between arms of precoxale acute of right. Pro- and mesotibiae cylindrical. Metacoxa with long, subacute, horn-shaped meracanthus. Metafemur with three ventral sense organs positioned at midlength; apex lacking stout long setae. Metatibia distinctly widened apically; with three to eight strongly sclerotized apical spurs that are relatively evenly spaced usually arranged to form a posteriorly open crown, laterally on short processes; with several irregularly spaced thorn-like setae postero-apically and many stout spines (Fig. 7D). Both metatarsal segments relatively short, subequal in length. Forewing oblong-oval to ovoid; widest in the middle or in apical third, 1.6–2.3 times as long as wide, rugose, opaque; vein C + Sc relatively evenly convex, slender, distinctly delimited from cell; costal break absent; pterostigma absent or small; nodal line not developed; vein R shorter than M + Cu; vein Rs relatively straight or slightly sinuous, turned towards costal margin apically; vein M longer than its branches; vein Cu_1a strongly curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} relatively perpendicular to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules coarse, often restricted to wing margins. Hindwing slightly shorter than forewing; with two to three costal setae proximal to costal break and two indistinct groups of four to five dense proximal and three to four spaced distal setae, distal to costal break; vein R + M + Cu bifurcating into R and M + Cu. Abdominal base lacking sclerotized area on either side covered in spines. Aedeagus with simple proximal portion bearing many weak folds subapically; apex of distal portion not differentiated from stem. Female subgenital plate lacking apical process.

Last instar immature. Antenna seven to eight segments; lacking sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface lacking minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in ventral or subterminal position; with or without additional pore fields developed.

Comments:

Monophyletic with strong molecular support (Fig. 10; Supporting Information, File S4) and morphologically supported by four synapomorphies (Fig. 11). The generic concept of Hodkinson and Bird (2000) is supported here. Included available species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3.

Syntomoza Enderlein, 1912

SyntomozaEnderlein 1921: 117. Type species: Euphyllura magnaKuwayama, 1908, by original designation and monotypy.

Diagnosis:

Adult. Head, in lateral view, deflexed 45° from longitudinal axis of body (Fig. 1E); in dorsal view much narrower than thorax, moderately transverse. Vertex rhomboidal, covered in areolate-rugose microsculpture, passing smoothly into genae anteriorly; coronal suture weakly developed at base; genae weakly produced ventrally but not enlarged into processes; frons forming narrow longitudinal band which is narrowed in the middle; frontal ocellus barely visible from above; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, small, rounded ventrally, hardly visible in lateral view as it is hidden by genae. Antenna slightly shorter than head width; flagellum with simple setae; segment 3 longest; segments 4, 6, 8, and 9 bearing each a subapical rhinarium surrounded by a wreath of spines. Thorax massive; dorsal outline, in lateral view, strongly curved. Pronotum, in dorsal view, almost straight, subrectagular; propleurites subrectangular, divided by oblique suture into larger epimeron and smaller episternum. Metapostnotum with flattened blunt tubercle. Mesosternum (Fig. 5F) wider than head, forming transverse band, more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture hardly visible; basisternum small, oval; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa (Fig. 6B) with short, knob-like meracanthus. Metafemur with the three ventral sense organs in the middle; apex without stout long setae. Metatibia longer than metafemur, slightly widened apically; bearing 9–10 slightly irregularly spaced apical sclerotized spurs and one peg-like seta adjacent to inner spurs. Both metatarsal segments relatively short, subequal in length. Forewing subtrapezoidal, with relatively straight, subparallel costal and anal margins, about 1.8 times as long as wide, coriaceous; vein C + Sc almost straight, slightly widened but indistinctly delimited from cell; costal break well-developed, in distance from apex of vein R₁, pterostigma indistinct; nodal line not developed; veins R and M + Cu subequal; vein Rs weakly sinuous; vein M much longer than its branches; vein Cu_1a concave relative to anal wing margin; veins Rs, M_{1 + 2}, M_{3 + 4} and Cu_1a apically slightly curved towards costal margin; anal break adjacent to apex of vein Cu_1b; surface spinules present in all cells, densely spaced. Hindwing slightly shorter than forewing; with six to nine costal setae proximal to costal break and two groups distal to costal break, four to five dense setae proximally and three to four spaced setae distally; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a weakly sclerotized area on either side covered in spines. Aedeagus with simple proximal portion bearing many weak folds subapically; apex of distal portion not differentiated from stem. Female subgenital plate lacking apical process.

Last instar immature. Antenna eight segments; lacking sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface lacking minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; no additional pore fields developed.

Comments:

The molecular and morphological analyses (Figs 10, 11; Supporting Information, File S4) suggest that Syntomoza sensuBurckhardt and Mifsud (2003) is polyphyletic (cf. comment under Anomoterga). In the restricted definition the genus is monotypic. The distribution and host plants are summarized in Table 3 and Supporting Information, File S3.

Paurocephalini Vondráček, 1963, stat. rev.

Paurocephalinae Vondráček 1963: 277. Synonymized with Liviinae by Burckhardt and Ouvrard 2012: 16.

=Diclidophlebiini Bekker-Migdisova 1973: 100. Synonymized with Paurocephalinae by Burckhardt and Mifsud 2003: 8.

=Sinuonemopsyllinae Li 2011: 373. Synonymized with Liviinae by Burckhardt and Ouvrard 2012: 16. Syn. nov.

Diagnosis:

Preocular sclerite not developed; subgenae not differentiated into separate sclerites. Antenna 10 segments, rarely eight or nine segments. Apex of distal portion of aedeagus differentiated from stem.—Immature. Antenna 3–10 segments; flagellum usually with sectasetae. Tarsal arolium small, petiolate; petiole broad, often with lateral appendages (Fig. 9L).

Comments:

The monophyly of Paurocephalini is strongly supported in the molecular analyses (Fig. 10; Supporting Information, File S4). In the morphological analyses, the tribe is supported by three synapomorphies (and two homoplasies) (Fig. 11). It includes seven genera.

DiclidophlebiaCrawford, 1919

HeteroneuraCrawford 1919: 152; nec Fallén 1810: 7, 25. Type species: Heteroneura oceanicaCrawford, 1919, by original designation and monotypy.

DiclidophlebiaCrawford 1920: 355; replacement name for Heteroneura Crawford nec Fallén.

=GyrozaEnderlein 1921: 122; replacement name for Heteroneura Crawford nec Fallén. Objective junior synonym of Diclidophlebia.

=ParaphalaroidaLoginova 1972: 851. Type species: Paurocephala fremontiaeKlyver, 1930, by original designation. Synonymized by Burckhardt and Mifsud 2003: 12.

=SinuonemopsyllaLi and Yang 1991: 11. Type species: Sinuonemopsylla excetrodendriLi and Yang, 1991, by original description and monotypy. Synonymized by Burckhardt and Mifsud 2003: 12.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1F); in dorsal view, moderately transverse, slightly or much narrower than thorax. Vertex rhomboidal to almost subrectangular; covered in imbricate microsculpture, sometimes much reduced on disc or smooth; passing smoothly into genae anteriorly; coronal suture ranging from fully developed to completely absent; genae weakly produced ventrally but not enlarged into processes; frons widely trapezoidal; median ocellus visible in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, large, rounded ventrally, well-visible in lateral view or slightly hidden by genae. Antenna 10-segmented, distinctly longer than head width; flagellum with simple setae; segment 3 longest, shorter than segments 4–5 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender to massive; dorsal outline, in lateral view, weakly to strongly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites narrowly subrectangular, divided by perpendicular suture into larger epimeron and smaller episternum. Metapostnotum with laterally compressed tooth, but small in D. leptonychiae. Mesosternum usually narrower, rarely wider (e.g. D. eastopi) than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture not visible; basisternum indistinct; katepisternum small antero-laterally, not bent dorsad laterally; trochantins in obtuse angle to each other. Pro- and mesotibiae cylindrical or robust and flattened (D. eastopi, D. leptonychiae, D. oceanica). Metacoxa with blunt or subacute horn-shaped meracanthus. Metafemur with the three ventral sense organs in medial or submedial position (Fig. 6F–H); apex with a group of stout long setae. Metatibia longer than metafemur, slightly widened apically; bearing 4–11 irregularly spaced to distinctly grouped apical sclerotized spurs that may be on raised processes, without unsclerotized bristle-like setae (Fig. 7E, F). Both metatarsal segments relatively short, subequal in length. Forewing oval to oblong-oval; widest in the middle, 1.8–2.5 times as long as wide, membranous or subcoriaceous; vein C + Sc almost straight to strongly convex, widened, indistinctly delimited from cell; costal break developed, close apex of vein R₁; pterostigma wide, coriaceous basally, membranous apically (Fig. 8B); nodal line sometimes developed; veins R and M + Cu usually subequal, sometimes vein R slightly shorter than M + Cu; vein Rs almost straight, strongly curved or convoluted; vein M shorter or longer than M_{1 + 2}; veins M_{1 + 2}, M_{3 + 4} and Cu_1a variable; anal break adjacent to apex of vein Cu_1b; surface spinules fine, spaced or dense, present in all cells. Hindwing slightly shorter than forewing; with two to four costal setae proximal to costal break and 6–18 irregularly spaced setae distal to costal break (Fig. 8I); vein R + M + Cu bifurcating into R and M + Cu. Abdominal base with a sclerotized area on either side covered in spines (Fig. 8N). Aedeagus with simple proximal portion subapically bearing many weak folds or subdivided; apex of distal portion differentiated from stem (Fig. 9C). Female subgenital plate lacking long apical process.

Last instar immature. Antenna nine segments; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface bearing minute clavate setae. Precaudal abdominal tergites usually lacking densely spaced simple setae or sectasetae. Anus in terminal position; usually with additional pore fields developed.

Comments:

Polyphyletic in the molecular analyses (Fig. 10; Supporting Information, File S4), although the genus as delimited here is recovered in the morphological analysis supported by one synapomorphy (Fig. 11). This concept of Diclidophlebia is much more restricted compared to that of Burckhardt and Mifsud (2003) who included species now referred to Diclidophlebia s.s., Haplaphalara, Klyveria, Melanastera, and Woldaia. Included species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3.

Haplaphalara Uichanco, 1912, stat. rev.

HaplaphalaraUichanco 1921: 260. Type species: Aphalara dahliRübsaamen, 1905, by original designation.

=AconopsyllaTuthill and Taylor 1955: 247. Type species: Psylla sterculiaeFroggatt 1901, by original designation. Syn. nov.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1G); in dorsal view about as wide as or slightly narrower than thorax, moderately transverse. Vertex rhomboidal; covered in areolate-rugose or imbricate microsculpture, sometimes much reduced on disc; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons widely trapezoidal; median ocellus visible or hidden in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, medium- to large-sized, flattened ventrally, not or hardly visible in lateral view as it is hidden by genae. Antenna 10-segmented, distinctly longer than head width; flagellum with simple setae; segment 3 longest, longer than segment 4 and shorter than 4–6 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites subrectangular, divided by oblique suture into larger epimeron and smaller episternum. Metapostnotum with blunt tubercle or shallow longitudinal ridge. Mesosternum narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture not or hardly visible; basisternum indistinct, but large, oval in H. irvingiae and H. Laos; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with blunt or subacute horn-shaped meracanthus. Metafemur with the three ventral sense organs in medial or submedial position (Fig. 6I, J); apex with a group of stout long setae. Metatibia longer than metafemur, hardly to distinctly widened apically; bearing 6–10 evenly spaced to distinctly grouped apical sclerotized spurs that may be on raised processes, without unsclerotized bristle-like setae (Fig. 7G). Both metatarsal segments relatively short, subequal in length. Forewing oblong-oval, widest in the middle or in apical third; 2.2–2.4 times as long as wide, membranous; vein C + Sc weakly, evenly convex, slender, distinctly delimited from cell; costal break developed, close apex of vein R₁; pterostigma wide, entirely membranous; nodal line developed in some species; vein R as long as or shorter than M + Cu; vein Rs weakly or moderately convex relative to costal margin; vein M longer than M_{1 + 2}; vein Cu_1a curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules fine or coarse, spaced or dense, present in all cells or sometimes almost completely lacking. Hindwing slightly shorter than forewing; with one or two costal setae proximal to costal break and two distinct groups one to four proximal and two to five distal setae, distal to costal break; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base usually with a sclerotized area on either side covered in spines. Aedeagus with proximal portion simple or subdivided subapically; apex of distal portion differentiated from stem, though sometimes indistinctly (Fig. 9E). Female subgenital plate with or without apical process.

Last instar immature. Antenna nine segments; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface bearing minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; with additional pore fields developed.

Comments:

In all analyses, Haplaphalara as redefined here is an artificial, paraphyletic assemblage (Figs 10, 11; Supporting Information, File S4). It is morphologically similar to Diclidophlebia and Melanastera from which it differs in the ungrouped apical metatibial spurs and the apex of the metatibia lacking unsclerotized antero-apical bristle-like setae. Except for the strong support in the molecular analyses for H. irvingiae + H. Laos as sister-taxa, there are no consistently supported groupings warranting generic rank within this assemblage. We, therefore, keep it as an artificial, paraphyletic genus awaiting further analyses. Included species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3. The following new or revived combinations are proposed here: Haplaphalara adelaidae (Braza and Calilung 1981: 344), comb. rev. (from Paurocephala; Haplaphalara, Navasero and Calilung 1998: 14; Diclidophlebia, Burckhardt and Mifsud 2003: 13); H. dahli (Rübsaamen 1905: 23), comb. rev. (Aphalara; Haplaphalara, Uichanco 1921: 261; Strophingia, Crawford 1925: 40: 40; Diclidophlebia, Burckhardt and Mifsud 2003: 13); H. dombeyae (Burckhardt et al., 2006: 369), comb. nov. (Diclidophlebia); H. grewiae (Kandasamy 1986: 61), comb. nov. (Paurocephala; Diclidophlebia, Burckhardt and Mifsud 2003: 14); H. irvingiae (Burckhardt et al., 2006: 376), comb. nov. (Diclidophlebia); H. maculata (Crawford 1919: 151), comb. rev. (Paurocephala; Haplaphalara, Loginova 1972: 841; Diclidophlebia, Burckhardt and Mifsud 2003: 14); H. menoni (Mathur 1975: 50), comb. rev. (Paurocephala; Haplaphalara, Hollis 1984: 28; Diclidophlebia, Burckhardt and Mifsud 2003: 14); H. parinari (Burckhardt et al., 2006: 382), comb. nov. (Diclidophlebia); H. sterculiae (Froggatt 1901: 255), comb. nov. (Psylla; Aconopsylla, Tuthill and Taylor 1955: 247; Diclidophlebia, Burckhardt and Mifsud 2003: 14); H. trimaculata (Mathur 1975: 69), comb. rev. (Paurocephala; Haplaphalara, Hollis 1984: 28; Diclidophlebia, Burckhardt and Mifsud 2003: 14); H. tuberculata (Burckhardt et al., 2006: 385), comb. nov. (Diclidophlebia); H. wagneri (Burckhardt et al., 2006: 386), comb. nov. (Diclidophlebia).

Klyveria Burckhardt, Serbina and Malenovský gen. nov.

Zoobank registration:

urn:lsid:zoobank.org:act:568C5A77-9010-4F66-B83E-575317BA95A5.

Type species:

Paurocephala crassiflagellataBurckhardt, 1996, by present designation.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1H); in dorsal view slightly wider than thorax, very short and wide. Vertex rhomboidal, moderately to strongly transverse (Fig. 2B); covered in finely granular microsculpture, sometimes much reduced on disc; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons small trapezoidal (Fig. 3G); median ocellus visible in perpendicular view to vertex; compound eyes, in dorsal view, subspherical, collared. Clypeus pear-shaped, large, flattened ventrally, hardly visible in lateral view as it is hidden by the genae. Antenna 10-segmented, distinctly longer than head width; flagellum with simple setae; segment 3 thickened, longest, about as long as segments 4–5 together; segments 3, 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Thorax moderately slender; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites narrowly subrectangular, divided by perpendicular suture into subequal epimeron and episternum. Metapostnotum with blunt tubercle. Mesosternum (Fig. 5C) narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave with median hump; pleurosternal suture not visible; basisternum triangular; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale right or acute. Pro- and mesotibiae cylindrical. Metacoxa with subacute horn-shaped meracanthus. Metafemur with the three ventral sense organs in medial position (Fig. 6K); apex with a group of stout, long setae. Metatibia longer than metafemur, hardly widened apically; bearing 12 or 13 densely spaced apical, weakly sclerotized spurs, without posterior peg-likw or thorn-like setae (Fig. 7I). Both metatarsal segments relatively short, subequal in length. Forewing oval or oblong-oval, widest in apical third; 2.2–2.3 times as long as wide, membranous; vein C + Sc weakly, evenly convex, slender, distinctly delimited from cell; costal break developed, close to apex of vein R₁; pterostigma narrow or wide, entirely membranous; nodal line developed; veins R and M + Cu subequal; vein Rs weakly convex relative to costal margin or sinuous; vein M longer than M_{1 + 2}; vein Cu_1a weakly or strongly curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules fine, spaced, present in all cells. Hindwing slightly shorter than forewing; with two or three costal setae proximal and five to eight setae distal to costal break; vein R + M + Cu bifurcating into R and M + Cu. Abdominal base with a sclerotized area on either side covered in spines. Male proctiger simple (Fig. 9A). Aedeagus with simple proximal portion bearing many weak folds subapically; distal portion moderately to strongly expanded basally, apex subdivided from stem (Fig. 9G). Female subgenital plate bearing apical process.

Last instar immature. Antenna 10 segments; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Tarsal arolium small, pedicel with lateral wings (Fig. 9L). Dorsal body surface bearing minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; with additional pore fields developed.

Etymology:

Dedicated to the US entomologist Frederick Detlev Klyver for his exemplary contributions to psyllid taxonomy.

Comments:

Monophyly strongly supported in all analyses (Figs 10, 11; Supporting Information, File S4). Klyveria currently includes two described species: K. crassiflagellata (Burckhardt 1996: 79), comb. nov. (from Paurocephala; Diclidophlebia, Burckhardt and Mifsud 2003: 13) and K. setinervis (Burckhardt 1996: 78), comb. nov. (from Paurocephala; Diclidophlebia, Burckhardt and Mifsud 2003: 14). Their distribution and host plants are summarized in Table 3 and Supporting Information, File S3.

Liella Burckhardt, Serbina and Malenovský gen. nov.

Zoobank registration:

urn:lsid:zoobank.org:act:F4331451-EC37-4E7D-BC2B-EAD32D6F198A.

Type species:

Paurocephala urenaeRussell, 1946, by present designation.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–90° from longitudinal axis of body (Fig. 1I); in dorsal view slightly narrower than thorax, moderately transverse. Vertex rhomboidal; covered in imbricate microsculpture; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons widely trapezoidal; median ocellus visible or hidden in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, medium-sized, rounded ventrally, hardly to well visible in lateral view. Antenna 10-segmented, distinctly shorter to longer than head width; flagellum with simple setae; segment 3 longest, longer than segment 4 and shorter than segments 4–6 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking marginal spines. Pronotum, in dorsal view, almost straight, subrectagular, weakly curved posteriad laterally; propleurites narrowly subrectangular, divided by perpendicular suture into larger epimeron and smaller episternum. Metapostnotum with subacute tubercle. Mesosternum narrower than head, forming transverse band about three times as wide as long laterally; anterior margin weakly concave with median hump; pleurosternal suture not visible; basisternum large, oval to rhomboidal; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale acute or right. Metacoxa with short, tubular, apically blunt meracanthus (Fig. 6C). Metafemur with the three ventral sense organs in submedial position (Fig. 6L); apex with a group of stout long setae. Metatibia as long as or longer than metafemur, hardly widened apically, without spur-like setae along length of metatibia; bearing seven or eight evenly spaced, weakly sclerotized apical spurs, without posterior peg-like or thorn-like setae (Fig. 7H). Both metatarsal segments relatively short; subequal in length or basal segment distinctly longer than apical one. Forewing oblong-oval, widest in the middle or in apical third; 2.2–2.7 times as long as wide, membranous; vein C + Sc weakly convex, slender, distinctly delimited from cell; costal break developed, close to apex of vein R₁; pterostigma narrow or wide, entirely membranous; nodal line developed; vein R as long as or shorter than M + Cu; vein Rs weakly or strongly convex relative to costal margin; vein M longer than M_{1 + 2}; vein Cu_1a curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular or oblique to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules (Fig. 8F) fine or coarse, spaced or dense, present in all cells, or sometimes partially lacking. Hindwing slightly shorter than forewing; with one to four costal setae proximal to costal break and two distinct groups distal to costal break, with two or three setae proximally and three to five setae distally (Fig. 8J); vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a sclerotized area on either side covered in spines. Male proctiger with elongate or small rounded lateral plates posteriorly (Fig. 9B). Aedeagus (Fig. 9H) with simple, subapically smooth proximal portion; apex of distal portion differentiated from stem. Female subgenital plate bearing apical process.

Last instar immature. Antenna three segments; usually bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface usually lacking minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in ventral position; lacking additional pore fields developed.

Etymology:

Dedicated to the Chinese entomologist Li Fasheng for his fundamental contribution to the taxonomy of Chinese psyllids.

Comments:

Monophyly strongly supported in the molecular analyses (Fig. 10; Supporting Information, File S4), and weakly in the morphological analysis by one homoplastic character that we consider a non-unique synapomorphy (Fig. 11). Included species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3. The following new combinations are proposed here: Liella abutili (Mifsud and Burckhardt 2002: 1943), comb. nov. (from Paurocephala); L. boxi (Mifsud and Burckhardt 2002: 1949), comb. nov. (Paurocephala); L. gossypii (Russell 1943: 115), comb. nov. (Paurocephala); L. hollisi (Mifsud and Burckhardt 2002: 1956), comb. nov. (Paurocephala); L. insolita (Mifsud and Burckhardt 2002: 1956), comb. nov. (Paurocephala); L. lanceomedia (Brown and Hodkinson 1988: 37), comb. nov. (Paurocephala; Diclidophlebia, Burckhardt and Mifsud 2003: 14); L. lienhardi (Mifsud and Burckhardt 2002: 1959), comb. nov. (Paurocephala); L. lucida (Mifsud and Burckhardt 2002: 1960), comb. nov. (Paurocephala); L. medleri (Mifsud and Burckhardt 2002: 1964), comb. nov. (Paurocephala); L. paucivena (Brown and Hodkinson 1988: 39), comb. nov. (Paurocephala; Diclidophlebia, Burckhardt and Mifsud 2003: 14); L. sinuata (Mifsud and Burckhardt 2002: 1974), comb. nov. (Paurocephala); L. urenae (Russell 1946: 94), comb. nov. (Paurocephala).

Melanastera Serbina, Malenovský, Queiroz and Burckhardt gen. nov.

Zoobank registration:

urn:lsid:zoobank.org:act:4201F6CF-FBF3-4BBD-BA1E-B9EBF39F9137.

Type species:

Diclidophlebia lucensBurckhardt et al., 2005, by present designation.

Diagnosis:

Adult. Head in lateral view, deflexed 45–80° from longitudinal axis of body (Fig. 1J); in dorsal view, slightly narrower than thorax, moderately transverse. Vertex rhomboidal; covered in imbricate microsculpture; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes (Fig. 2C); frons widely trapezoidal; median ocellus visible or hidden in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, medium- to large-sized, flattened ventrally, hardly visible in lateral view as it is hidden by genae. Antenna 10 segments, slightly to distinctly longer than head width; flagellum with simple setae; segment 3 longest, longer than segment 4 and shorter than segments 4–6 together; segments 4, 6, 8, and 9 bearing each a subapical rhinarium usually surrounded by a wreath of spines (partly lacking in Melanastera Venezuela). Thorax moderately slender; dorsal outline in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites narrowly subrectangular, divided by perpendicular suture into larger epimeron and smaller episternum. Metapostnotum (Fig. 4C) with laterally compressed tooth. Mesosternum narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave with median hump; pleurosternal suture well visible; basisternum triangular; katepisternum large antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with blunt or subacute horn-shaped meracanthus. Metafemur with the three ventral sense organs in medial position (Fig. 6M); apex with a group of stout long setae. Metatibia longer than metafemur, slightly widened apically; bearing 4–11 grouped apical sclerotized spurs separated by one to seven unsclerotized bristle-like setae anteriorly or antero-laterally (rarely lacking), without posterior peg-like or thorn-like setae (Fig. 7J, K). Both metatarsal segments relatively short, subequal in length. Forewing oval or oblong-oval, widest in the middle or in apical third; 1.9–2.7 times as long as wide, membranous; vein C + Sc weakly or strongly convex, slender, distinctly delimited from cell; costal break developed, close to apex of vein R₁; pterostigma narrow to wide, entirely membranous (Fig. 8C); nodal line developed; vein R as long as or shorter than M + Cu; vein Rs weakly or strongly convex relative to costal margin; vein M longer than M_{1 + 2}; vein Cu_1a curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular or oblique to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules (Fig. 8G) fine or coarse, spaced or dense, present in all cells or sometimes partially reduced. Hindwing slightly shorter than forewing; with two to four costal setae proximal to costal break and two distinct groups distal to costal break, with three to six dense setae proximally and three to five spaced setae distally; vein R + M + Cu bifurcating into R and M + Cu. Abdominal base with a sclerotized area on either side covered in spines. Aedeagus with proximal portion bearing many weak folds subapically or strongly subdivided subapically; apex of distal portion differentiated from stem (Fig. 9I). Female subgenital plate lacking apical process.—Last instar immature. Antenna 10-segmented; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Tarsal arolium small (Fig. 9N). Dorsal body surface bearing minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; with additional pore fields developed.

Etymology:

Melanastera is a pun using the first syllables of some host families: Mel-astomataceae, An-nonaceae, and Astera-ceae, and is a tribute to the British musician David Bowie in reference to his final studio album ‘Blackstar’ (from Ancient Greek μέλας = black and Latin aster = star).

Comments:

The monophyly of Melanastera is strongly supported in all molecular analyses if the single taxon, M. Venezuela is excluded. With the inclusion of the latter taxon the genus is not (ML-mix, Fig. 11), or only weakly (ML-part; Supporting Information, File S4A) or moderately (BI; Supporting Information, File S4B) supported. In the morphological analysis, the monophyly of Melanastera (including M. Venezuela) is supported by one synapomorphy (and five homoplasies) (Fig. 11). Included species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3. The following new combinations are proposed here: Melanastera fava (Brown and Hodkinson 1988: 44), comb. nov. (from Haplaphalara; Diclidophlebia, Burckhardt and Mifsud 2003: 14); M. heterotrichi (Caldwell and Martorell 1952: 605), comb. nov. (Paurocephala; Haplaphalara, Brown and Hodkinson 1988: 40; Diclidophlebia, Burckhardt and Mifsud 2003: 14); M. longitarsata (Brown and Hodkinson 1988: 42), comb. nov. (Haplaphalara; Diclidophlebia, Burckhardt and Mifsud 2003: 14); M. lucens (Burckhardt, Hanson and Madrigal 2005: 742); comb. nov. (Diclidophlebia); M. maculipennis (Brown and Hodkinson 1988: 47), comb. nov. (Haplaphalara; Diclidophlebia, Burckhardt and Mifsud 2003: 14); M. paucipunctata (Brown and Hodkinson 1988: 40), comb. nov. (Haplaphalara; Diclidophlebia, Burckhardt and Mifsud 2003: 14); M. pilosa (Burckhardt et al., 2006: 384), comb. nov. (Diclidophlebia); M. smithi (Burckhardt et al., 2006: 242), comb. nov. (Diclidophlebia); M. tuxtlaensis (Conconi 1972: 51), comb. nov. (Paurocephala; Haplaphalara, Brown and Hodkinson 1988: 40; Diclidophlebia, Burckhardt and Mifsud 2003: 14).

PaurocephalaCrawford, 1913

PaurocephalaCrawford 1913: 293. Type species: Paurocephala psyllopteraCrawford, 1913, by original designation.

=Paurocephala subgenus ThoracocornaKlimaszewski 1970: 427. Type species: Paurocephala chonchaiensisBoselli, 1929, by original designation. Synonymized with Paurocephala by Loginova 1972: 842.

=MarpsyllaNavasero and Calilung 2001: 126. Type species: Marpsylla baltazaraeNavasero and Calilung, 2001, by original designation. Synonymized with Paurocephala by Burckhardt and Ouvrard 2012: 16.

=PaurotergaNavasero 2010: 20. Type species: Pauroterga zamboangensisNavasero, 2010, by original designation. Synonymized with Paurocephala by Burckhardt and Ouvrard 2012: 16.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–80° from longitudinal axis of body (Fig. 1K); in dorsal view, about as wide as or slightly narrower than thorax, moderately to strongly transverse. Vertex rhomboidal; covered in imbricate microsculpture; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons elliptical (Fig. 3H); median ocellus visible or hidden in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, medium-sized to large, rounded ventrally, visible in lateral view. Antenna 8–10 segments, shorter to much longer than head width; flagellum usually with simple setae, rarely with a bifid seta associated the rhinaria; in 10-segmented antenna, segment 3 longest, longer than segments 4–5 together and shorter than segments 4–7 together, and segments 4, 6, 8, and 9 bearing each a subapical rhinarium lacking a wreath of spines. Thorax moderately slender to relatively robust; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites subrectangular, divided by perpendicular suture into larger epimeron and smaller episternum. Metapostnotum (Fig. 4D) with conical horn. Mesosternum (Fig. 5E) narrower than head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture not visible; basisternum large, oval to rhomboida; katepisternum large antero-laterally, bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with slender, tubular, apically blunt meracanthus (Fig. 6D). Metafemur with the three ventral sense organs in submedial position (Fig. 6N); apex with a group of stout long setae. Metatibia shorter or longer than metafemur, hardly widened apically, with weakly sclerotized spur-like setae along length of metatibia similar to those that are at apex; bearing seven or eight evenly spaced, apical, weakly sclerotized spurs, without posterior peg-like or thorn-like setae (Fig. 7L). Basal metatarsal segments as long as or longer than apical segment. Forewing oblong-oval, widest in the middle or in apical third; 2.1–2.5 times as long as wide, membranous; vein C + Sc weakly convex, slender, distinctly delimited from cell; costal break developed, close to apex of vein R₁; pterostigma narrow or wide, entirely membranous; nodal line developed; vein R usually as long as but rarely shorter than M + Cu; vein Rs weakly or strongly convex relative to costal margin or sinuous; vein M longer than M_{1 + 2}; vein Cu_1a curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} perpendicular or oblique to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules fine or coarse, spaced or dense, present in all cells, or sometimes partially lacking. Hindwing slightly shorter than forewing; with one to three costal setae proximal to costal break and two distinct groups distal to costal break, with one or two setae proximally and two to five setae distally; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a sclerotized area on either side covered in spines, sometimes forming finger-like process (Fig. 8O, P). Aedeagus with proximal portion strongly subdivided subapically; apex of distal portion differentiated from stem. Female subgenital plate bearing apical process.

Last instar immature. Antenna three to nine segments; segmentation of flagellum sometimes incomplete; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface usually lacking clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in ventral position; lacking additional pore fields.

Comments:

The monophyly of Paurocephala is strongly supported in all molecular analyses (Fig. 10; Supporting Information, File S4) and by four synapomorphies in the morphological analysis (Fig. 11). Included species, distribution, and host plants are summarized in Table 3 and Supporting Information, File S3.

Woldaia Brown and Hodkinson, stat. rev.

WoldaiaBrown and Hodkinson 1988: 49. Type species: Woldaia nebulosaBrown and Hodkinson, 1988, by original designation and monotypy.

Diagnosis:

Adult. Head, in lateral view, deflexed 45–70° from longitudinal axis of body (Fig. 1L); in dorsal view distinctly narrower than thorax moderately transverse. Vertex subrectangular; covered in imbricate microsculpture; passing smoothly into genae anteriorly; coronal suture fully developed; genae weakly produced ventrally but not enlarged into processes; frons small trapezoidal; median ocellus clearly visible in perpendicular view to vertex; compound eyes, in dorsal view, hemispherical, adpressed to head. Clypeus pear-shaped, large, flattened ventrally, visible in lateral view. Antenna 10-segmented, distinctly longer than head width; segments 3–5 with indistinct bifid setae; segment 3 longest, shorter than segments 4–6 together; segments 4, 6–9 bearing each a subapical rhinarium lacking marginal spines. Thorax relatively robust; dorsal outline, in lateral view, weakly curved. Pronotum, in dorsal view, weakly curved posteriad laterally; propleurites narrowly subrectangular, divided by perpendicular suture into larger epimeron and smaller episternum. Metapostnotum laterally compressed. Mesosternum as wide as head, forming transverse band more than three times as wide as long laterally; anterior margin weakly concave; pleurosternal suture not visible; basisternum indistinct; katepisternum small antero-laterally, not bent dorsad laterally; angle between arms of precoxale obtuse. Pro- and mesotibiae cylindrical. Metacoxa with blunt horn-shaped meracanthus. Metafemur with the three ventral sense organs in medial position; apex with a group of stout long setae. Metatibia as long as metafemur, slightly widened apically; bearing 9 or 10 irregularly spaced, apical, weakly sclerotized spurs that may be on raised processes, without posterior peg-like or thorn-like setae. Both metatarsal segments relatively short, subequal in length. Forewing ovate, widest in basal third; 2.3 times as long as wide, membranous; vein C + Sc evenly convex, broad, indistinctly delimited from cell; costal break lacking; pterostigma wide, entirely membranous; nodal line absent; veins R and M + Cu subequal; vein Rs weakly sinuous; vein M longer than M_{1 + 2}; vein Cu_1a weakly curved towards anal margin; veins M_{1 + 2} and M_{3 + 4} oblique to wing margin apically; anal break adjacent to apex of vein Cu_1b; surface spinules present in all cells. Hindwing slightly shorter than forewing; with one costal seta proximal to costal break and seven ungrouped setae distal to costal break; vein R + M + Cu indistinctly trifurcating, base of vein Cu indistinct. Abdominal base with a sclerotized area on either side covered in spines. Aedeagus with proximal portion strongly subdivided subapically; apex of distal portion differentiated from stem. Female subgenital plate bearing apical process.

Last instar immature. Antenna 10 segments; bearing sectasetae or lanceolate setae on antennal flagellum. Mid- and hindlegs without massive peg-like setae. Dorsal body surface bearing minute clavate setae. Precaudal abdominal tergites lacking densely spaced simple setae or sectasetae. Anus in terminal position; with additional pore fields developed.

Comments:

Woldaia is monotypic, currently including only W. nebulosaBrown and Hodkinson 1988: 49, comb. rev. (from Diclidophlebia, Burckhardt and Mifsud 2003: 14). Its distribution and host plants are summarized in Table 3 and Supporting Information, File S3.

Species excluded from Liviidae

Cockerell (1915) described a fragment of a forewing from the Oligocene of the Isle of Wight as †Necropsylla anglicaCockerell, 1915. Bekker-Migdisova (1985) transferred the species provisionally to Camarotoscena based on the long, only weakly curved vein Cu_1a and the straight vein Cu_1b. These characters are not diagnostic for Camarotoscena but occur also in †Lapidopsylla Klimaszewski, 1993 (Aphalaridae, Aphalarinae, †Paleopsylloidini), erected for two species from the Eocene/Oligocene of the Isle of Wight. †Necropsylla anglica resembles the type species of †Lapidopsylla, †L. thornessbaya Klimaszewski, 1993, with which it appears congeneric. We suggest the following new combination: †Lapidopsylla anglica (Cockerell 1915: 487), comb. nov. from Necropsylla; Camarotoscena, Bekker-Migdisova 1985: 81.

Syntomoza lebeziaHodkinson, 1986 was described from a single specimen from Belize. Hodkinson (1986) placed this species provisionally in Syntomoza as it shares with S. magna the lobed male proctiger and the bulbous apex of the distal portion of the aedeagus, but differs in the head, metatarsi, and terminalia. As the species is morphologically quite distinct, Hodkinson (1986) suggested that it ‘will probably deserve generic status in its own right’. Burckhardt and Mifsud (2003) concluded that the species ‘should be excluded from Syntomoza and the Paurocephalinae’ and suggested some resemblance in the head to MetapsyllaKuwayama, 1908. However, S. lebezia shares with Katacephala (Psyllidae: Katacephalinae) the metatibia lacking a genual spine and bearing a posteriorly open crown of apical spurs, the presence of two basimetatarsal spurs, the posteriorly lobed male proctiger, and the host association with Myrtaceae (unpublished BMNH data). For this reason, we transfer Syntomoza lebezia Hodkinson formally to Katacephala, as Katacephala lebezia (Hodkinson 1986: 149), comb. nov.

The Afrotropical species P. bicarinataPettey, 1924 and P. hottentottiPettey, 1933 are currently formally included in Paurocephala. They possess basimetatarsal spurs and are, therefore, not members of the Liviinae (Burckhardt and Mifsud 2003).

DISCUSSION

Phylogeny of Liviinae and generic concepts

The trees resulting from the molecular, morphological, and combined analyses share a similar topology without major contradictions (Figs 10, 11; Supporting Information, File S4). The monophyly of the subfamilies Euphyllurinae and Liviinae, of the tribes Liviini and Paurocephalini, and of the genera Camarotoscena, Klyveria, Liella, Livia, and Paurocephala is strongly or moderately supported in the molecular and combined analyses. In the morphological analysis, the monophyly of neither the Euphyllurinae nor the Liviinae is supported, and that of the genera Camarotoscena and Liella is only weakly supported by homoplastic characters that we consider to be non-unique synapomorphies. Support from two or more synapomorphies in the morphological analysis is found for the two tribes and the genera Klyveria, Livia, and Paurocephala. Melanastera, which is not (ML-mix), strongly or weakly (BI) or moderately (ML-part) supported with molecular and combined data, is supported morphologically with one synapomorphy. Aphorma is also supported by one synapomorphy morphologically, but represented only by a single species in the molecular and combined analyses. Syntomoza sensuBurckhardt and Mifsud (2003) in the broad definition (including Anomoterga and Homalocephala) is not recovered in any of the analyses. In the morphological analysis, the type species of Anomoterga and Homalocephala group together in a clade supported by one synapomorphy, along with an African and a Palaearctic species originally described in Camarotoscena (Loginova 1975) and later transferred to Syntomoza (Burckhardt and Mifsud 2003). In the molecular and combined analyses, only two species of Anomoterga are included which form a paraphylum with Camarotoscena (molecular ML-mix; combined) or a polyphylum with Aphorma and Camarotoscena (molecular BI, ML-part). In the molecular and combined analyses, the type species of Syntomoza constitutes the moderately supported sister-group of Livia (molecular BI, ML-part) or the weakly, moderately, or strongly supported sister-group of Aphorma + Livia (molecular ML-mix; combined). The latter relationship is also supported morphologically by one synapomorphy. For this reason, we recognize the monotypic Syntomoza and Anomoterga (with six species, cf. Supporting Information, File S3) as distinct genera and Homalocephala as a synonym of the latter.

In the morphological analysis, Paurocephalini splits into two clades, that correspond to Diclidophlebia sensuBurckhardt and Mifsud (2003) (supported by one synapomorphy) and Paurocephala sensuBurckhardt and Mifsud (2003)(grouped by two homoplasies). The latter, also weakly supported by combined data, contains Liella and Paurocephala, which differs from the molecular analyses where Liella and Paurocephala are distant and nested within Diclidophlebia s.l., though with little to no support. As none of the analyses strongly supports a particular relationship of Liella and Paurocephala within the tribe, but both are strongly supported in the molecular analyses, and weakly or strongly supported in the combined analyses, and morphologically diagnosable (see key), we treat them as distinct genera. The morphological analysis strongly suggests that Marpsylla and Pauroterga are congeneric with Paurocephala, confirming their synonymy.

Within Diclidophlebia sensuBurckhardt and Mifsud (2003), all analyses support the sister-group relationship of Klyveria and the monotypic Woldaia (strong in the molecular and combined analyses, and with one morphological synapomorphy). Apart from ML-part (combined), Melanastera is recovered in all analyses. The support is strong or moderate (combined and molecular BI), weak (molecular ML-part) (BI), or absent (molecular ML-mix), and with one morphological synapomorphy. The relationships of these two clades (Klyveria + Woldaia and Melanastera) vary strongly between the analyses. Klyveria, Melanastera, and Woldaia are morphologically well diagnosable, and we treat them as distinct genera. The phylogenetic relationships of the remainder species of Diclidophlebia sensuBurckhardt and Mifsud (2003) differ strongly between the analyses. Morphologically, a clade of seven species (supported by one synapomorphy) can be diagnosed by the broad vein C + Sc of the forewing, the pterostigma that is separated in a coriaceous basal and a membranous apical portion, and the grouped apical metatibial spurs. We name this clade Diclidophlebia s.s. (including the synonyms Paraphalaroida and Sinuonemopsylla) and the rest of the species we refer to Haplaphalara (with the synonym Aconopsylla), which constitutes an artificial paraphyletic assemblage.

Euphyllurinae, Liviinae, Liviini, Paurocephalini, Aphorma, Camarotoscena, Klyveria, Liella, Livia, Melanastera, and Paurocephala are recovered in most or all analyses and are morphologically well diagnosable (see key), making them useful units for a classification. The monophyly of the Liviinae and its two included tribes were also strongly supported in the molecular analyses of Percy et al. (2018) that included 11 species of Liviinae.

Morphological homoplasy in Liviinae

In the morphological tree, many clades are partially or exclusively supported by character states that have evolved or are lost independently in different clades on the strict consensus tree (homoplasies in Fig. 11). For example, a sister-group relationship of Liella and Paurocephala in the morphological analysis is partly based on weakly sclerotized metatibial spurs (character 32: state 1), a character state present also in Klyveria and Woldaia. Homoplasy and the absence of more convincing morphological synapomorphies have often misled previous authors when classifying Liviinae. Burckhardt (1996) described Klyveria crassiflagellata and K. setinervis in Paurocephala, following Brown and Hodkinson (1988) who diagnosed Paurocephala by the weakly sclerotized apical metatibial spurs, the thickened distal portion of the aedeagus, and the lack of lateral spurs on the metabasitarsi. Burckhardt (1996) noted that K. setinervis and K. crassiflagellata differ quite fundamentally from the other Paurocephala species in the very short head, the thickened antennal flagellum, and the presence of a subapical rhinarium on antennal segment 3 and concluded that the two Neotropical species might not be congeneric with the Old World Paurocephala species. Later, Burckhardt and Mifsud (2003) transferred K. crassiflagellata and K. setinervis to Diclidophlebia, defined in their cladistic analysis by two unique putative synapomorphies: the wide, rhomboidal frons and the complex structure of the distal portion of the aedeagus. Our analysis of the morphological dataset suggests that the resemblance of the frons and the aedeagus between K. crassiflagellata, K. setinervis, and Diclidophlebia s.s. is only superficial. We place both species into a genus, more closely related to another Neotropical species with a highly aberrant morphology, Woldaia nebulosa. This result is strongly supported also by the phylogenetic analysis of the molecular data.

Liella lanceomedia has seen a similarly shifting classification: this Neotropical species was described in Paurocephala (Brown and Hodkinson 1988) and later transferred to Diclidophlebia (Burckhardt and Mifsud 2003). Brown and Hodkinson (1988) suggested that the closest relatives of L. lanceomedia might be two Afrotropical species previously included in Paurocephala (L. gossypii and L. urenae), although L. lanceomedia differs from the latter in lacking stout internal spurs on the parameres and in details of the forewing venation. Our present morphological and molecular phylogenetic analyses confirm the suggestion of Brown and Hodkinson (1988). We place L. lanceomedia into a newly described genus Liella, together with members of the Paurocephala gossypii-group sensuMifsud and Burckhardt (2002). Although this grouping is strongly supported by molecular data, the morphological support for Liella remains weak.

Within the current study, phylogenetically informative, although partly homoplastic, morphological characters for Liviinae were mainly found on the head (e.g. presence of preocular and subgenal sclerites, and shape of frons), thorax (e.g. development of pleurosternal suture, shapes of mesosternum, basisternum, and metapostnotum) and hindlegs (details of structure and arrangement of metatibial spurs and setae) and, to a lesser extent, also the forewings (e.g. structure of costal margin and pterostigma) and terminalia of adults (mainly aedeagus), as well as the tarsal arolium and some setal characters of the immatures. From an evolutionary perspective, the question arises whether the states of these characters are homologous or not in different Liviinae taxa. A way to evaluate homoplasy is the reconstruction of ancestral states of the morphological characters on to a solid topology inferred from an independent dataset (Al Sayad and Yassin 2019). In our molecular trees, many nodes are poorly supported and the topologies are not solid; therefore, we cannot use this approach. In fact, we use the morphological analysis to illuminate the poorly supported nodes in the molecular analyses to derive some reasonable formal classification of genera, implicitly considering our morphological dataset as trustworthy.

Biogeographic and host patterns

The distribution and host plants of the species are given in the Supporting Information, File S3 and are summarized in Table 3. The clade Liviini (Fig. 11) is north temperate or subtropical with the exceptions of Anomoterga and Aphorma. The former has species in the Afrotropics, Indomalaya, Oceania, and the Palaearctic, the latter in Indomalaya and the Palaearctic. Camarotoscena and Syntomoza are restricted to the Palaearctic, and Livia is Holarctic being the only genus of Liviini with species in the Americas. The clade Paurocephalini is exclusively tropical and subtropical. Two genera are restricted to the Old World: Haplaphalara widespread in the Afrotropics, Australasia, and Indomalaya, and Paurocephala in Australasia, Indomalaya, and Oceania. The clade comprising the two small genera, Klyveria and Woldaia, is exclusively Neotropical. Diclidophlebia, Liella, and Melanastera are pantropical/subtropical, but the first two are most diverse in the Old World and the last in the New World. There are no clear-cut biogeographic patterns, but it is interesting to note the near vicariance between the two tribes (temperate versus tropical) (Figs 10, 11) and the two pairs of sister-genera (based on morphology) (Fig. 11): Liella–Paurocephala (Afrotropics versus Australasia, Indomalaya, and Oceania) and Diclidophlebia–Melanastera (Old World versus New World).

In the Liviini, Anomoterga, Camarotoscena, and Syntomoza are restricted to Salicaceae (Malpighiales), with the possible exception of Anomoterga tahuata (see Supporting Information, File S3). The non-monophyly of the three genera in all analyses (Figs 10, 11) may result from an ancestral host association of Liviini with Salicaceae (and subsequent host shifts to other plant taxa in Aphorma and Livia) or from multiple colonizations of Salicaceae. Species developing on Salicaceae occur also in other psyllid families, e.g. Bactericera Puton, 1876, Egeirotrioza Boselli, 1931 (both Triozidae), and Cacopsylla Ossiannilsson, 1970 (Psyllidae). Aphorma is associated with Ranunculaceae (Ranunculales), and Livia with the two monocot families Cyperaceae and Juncaceae (Poales). Both Ranunculaceae and monocots are unusual psyllid hosts (Ouvrard et al. 2015). All genera of Paurocephalini have at least one species that develops on Malvaceae (Malvales) (Fig. 11). Klyveria and Woldaia are restricted to the malvaceous genus Luehea. Diclidophlebia, Haplaphalara, and Liella species are associated mostly with Malvaceae but a few species develop on the unrelated families Chrysobalanaceae, Euphorbiaceae, Hypericaceae, Irvingiaceae (Malpighiales), Connaraceae (Oxalidales), and Rhamnaceae (Rosales). The single Old World species of Melanastera develops on Malvaceae, a host family not used by the New World species, which are mostly associated with Melastomataceae (Myrtales) and one with Cannabaceae (Rosales). Some undescribed Melanastera species from Brazil and Central America develop also on Annonaceae, Myristicaceae (both Magnoliales), and Asteraceae (Asterales) (unpublished BMNH and NHMB data). Two of the species groups of Paurocephala defined by Mifsud and Burckhardt (2002) develop on Malvaceae and one on Cannabaceae, Moraceae, and Urticaceae (Rosales). Malvales and Rosales also provide hosts for psyllid species from other families, e.g. all species of Carsidaridae and Pauropsylla Rübsaamen (Triozidae), as well as some Cacopsylla spp. (Psyllidae) (Ouvrard 2023).

Species diversity

The number of described Liviini species is only slightly more than half that of Paurocephalini (Table 3). With 25 described species, Livia is the largest genus within Liviini followed by Camarotoscena. Anomoterga and Aphorma have few species, and Syntomoza is monotypic. In material at hand, there are five undescribed species of Anomoterga and one of Aphorma. All the new taxa are extra-Palaearctic. Paurocephalini is represented by 89 described species. Paurocephala, which was monographically revised by Mifsud and Burckhardt (2002), includes currently 46 species. No comparable revisions exist for the other genera. Haplaphalara, Liella, and Melanastera comprise around a dozen described species each, and the other genera less than 10. Woldaia is monotypic. Except for Woldaia, all genera of Paurocephalini are represented in the collections at hand by undescribed species. Melanastera, in particular, is very species-rich in tropical and subtropical America, with many undescribed species associated with Annonaceae, Melastomataceae, and other plant families.

CONCLUSION

Morphological homoplasy is a major problem in psyllid taxonomy and was responsible in the past for a great deal of taxonomic instability and confusion (White and Hodkinson 1985, Burckhardt and Ouvrard 2012). The broad molecular study of the world fauna by Percy et al. (2018) corroborated the monophyly of many taxa previously diagnosed with morphological characters but indicated also a series of previously misplaced taxa (Burckhardt et al. 2021). The present study, using an approach of reciprocal illumination between molecular and morphological data, confirms the monophyly of Liviinae and three of its six previously recognized genera, while the other three are polyphyletic. Phylogenetically informative morphological characters were found on the head, thorax, and hindlegs and, to a lesser extent, also the forewings and terminalia of adults, as well as the tarsal arolium and some setal characters of the immatures. Some of these characters, e.g. details of the metatibial spurs, have not been previously used for phylogenetic reconstruction and are potentially useful to the systematics of other groups of psyllids. Ward and Boudinot (2021), in a study of the systematics of the ant genera Camponotus Mayr, 1861 and Colobopsis Mayr, 1861, succinctly emphasized the value of reciprocal illumination between molecular and critical morphological analysis. In Liviinae, the molecular data provide a strong phylogenetic signal (high support values) at subfamily and tribal levels, whereas morphological characters show more phylogenetic signal at generic level. Despite the high degree of homoplasy in morphological characters, they remain valuable for both reconstruction and evolutionary interpretation of the psyllid phylogeny, particularly in parallel with molecular data.

The Psylloidea is most species-rich in the tropics and southern hemisphere, but these faunas are less studied than those of the Holarctic. Material in major psyllid collections (BMNH, DMPC, MHNG, and NHMB) suggests that the global psyllid diversity comprises at least as many species as have been described to date, and that a large number of the undescribed species originate from the tropics (Burckhardt and Queiroz 2012, Burckhardt et al. 2021). This situation is well reflected in the Liviinae. The predominantly north temperate tribe Liviini comprises only slightly more than half the number of species of the tropical and subtropical Paurocephalini. In the former, the number of new species represented in the material at hand constitutes around 12% of the described species, whereas in the Paurocephalini, there are slightly more undescribed than described species. Future targeted fieldwork will certainly substantially increase the number of new species of Paurocephalini and probably of Anomoterga, but not of the genera Aphorma, Camarotoscena, Livia, and Syntomoza.

Most species of Liviinae are, as far as known, monophagous or oligophagous, and six of the genera are restricted to a single plant family (only confirmed hosts considered) (Table 3). Confirmed hosts of the Liviinae occur in five major clades of angiosperms: magnoliids (Annonaceae, Myristicaceae), monocots (Cyperaceae, Juncaceae), Ranunculales (Ranunculaceae) at the base of eudicots, rosids (Malpighiales: Chrysobalanaceae, Euphorbiaceae, Hypericaceae, Irvingiaceae, Salicaceae; Oxalidales: Connaraceae; Malvales: Malvaceae; Myrtales: Melastomataceae; Rosales: Cannabaceae, Moraceae, Rhamnaceae, Urticaceae; Sapindales: Sapindaceae), and asterids (Asterales: Asteraceae) (Table 3, Fig. 11). No species of Liviinae are known to develop on Fabaceae or Myrtaceae, which range among the most important psyllid hosts (Ouvrard et al. 2015). The occurrence of malvaceous hosts in all genera of Paurocephalini suggests that this may be the original host association from which shifts to other, unrelated taxa occurred, rather than cospeciation.

The majority of genera of Liviinae occur in more than one biogeographic realm (Table 3; Fig. 11) but two genera each are restricted to the Neotropics (Klyveria, Woldaia) and the Palaearctic (Camarotoscena, Syntomoza), respectively. No clear-cut biogeographic patterns can be found. The tribes show an imperfect vicariance between north temperate (Liviini) and tropical/south temperate realms (Paurocephalini). Similarly imperfect is the Old World/ New World vicariance of Diclidophlebia and Melanastera and the Afrotropics/Australasia, Indomalaya, and Oceania vicariance of Liella and Paurocephala.

Zoobank registration: urn:lsid:zoobank.org:pub:535B2566-E5C4-4C18-BCDC-550464F33B1E.

ACKNOWLEDGEMENTS

For the loan of material we thank Paul Brown, Jon Martin (BMNH), and David Preston (BPBM). We are grateful to two anonymous reviewers for their comments on a previous manuscript version. We warmly acknowledge the great help of Lenka Dušátková, Kristína Křížová, and Kristýna Koukalová with organizing laboratory work and molecular sequencing at Masaryk University. We are grateful to Lenka Dušátková, David Ortiz, and Andrea Špalek Tóthová who kindly helped with improving the molecular techniques and the phylogenetic analyses of DNA sequence data. Alexandre Domahovski (Universidade Federal do Paraná, Brazil) helped with TNT and Jowita Drohojowska (Silesian University, Katowice, Poland) commented on the mesosternum for which we are very grateful.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

FUNDING

L.Š.S. was funded partly by a grant of the Swiss National Science Foundation (SNSF; P2BSP3_168733). G.C. was partly funded by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2020R1I1A1A01074074).

DATA AVAILABILITY

The data underlying this article are contained within and in the online Supporting Information.

REFERENCES

Author notes