The resolution of phylogenetic relationship among animals is still one of the most challenging problems in systematic zoology. Insect wing is a highly valued morphological character in the systematics, but few studies have been conducted to quantify wing shape variations for phylogenetic reconstruction. In this study, with Cantharinae as the subject, we conducted the GM analyses from hindwings of 16 representative genera. Further, we conducted the UPGMA based on Procrustes distance and Euclidean similarity measure of Mahalanobis distance, respectively, and NJ analysis of the Mahalanobis distance, as well as MP analysis using merged landmark dataset. In the meantime, we constructed the phylogenetic relationships among these genera based on the mitochondrial genomes, with a total of 41 sequences novel to Cantharinae, by BI and ML analyses. As a result, the CVA analysis demonstrated that the hindwing shapes of the cantharid genera are significantly different from one another. All the topologies produced by the GM data partially correspond with that of mitogenomic data. The close relationships of some genera are frequently recovered, including Cyrebion + Themus, Cantharis + Taiwanocantharis + Taocantharis, Stenothemus + Falsopodabrus + Habronychus. These results prove the importance and potential application of the hindwing shapes in recovering the relationships among the sibling genera.

Geometric morphometric, genera, hindwing, mitochondrial genome, phylogeny, Soldier beetles

Knowledge of how living (and extinct) taxa are related to one another underpins much of evolutionary biology (Kapli et al. 2020). The resolution of phylogenetic relationship among animals is one of the most challenging problems in systematic zoology (Field et al. 1988). In the premolecular age, organismal phylogenies were generally created based on morphological character states. However, there are very few homologous morphological characters that can be compared among all organisms. With the arrival of DNA sequencing, molecular phylogenetic has become the standard for inferring evolutionary relationships (Ziemert and Jensen 2012). Since that a number of genes with fundamental biochemical functions are found in all species and they can be sequenced, aligned, and analyzed to study phylogenetic relationships at the deepest part of the tree of life (Hillis and Dixon 1991). Moreover, based on the analyses of ribosomal gene sequences, this relationship appeared to be robust to tree-building methods. In recent years, mitochondrial genome has become the most popular molecular marker in inferring the phylogenetic relationships among the animals, especially for various groups of insects (Bajpai and Tewari 2010).

Although many advantages of molecular over morphological phylogenetic have been recognized, morphological analysis still cannot be replaced or neglected in the construction of phylogenetic relationships. For many groups of poorly known organisms, the only known specimens of many species are represented merely by the holotype or type series. Collecting additional material can be prohibitive because of rarity of the species, inaccessibility of the habitat, destruction of known collection localities, legal protection of the habitat or species, or high costs of procurement. A high percentage of recently extant species have been exterminated in this century by human activities, especially through the destruction of tropical rainforests (Myers 1986). Because of this high extinction rate, a majority (or at least a large fraction) of described species may never be collected again and will remain known only from traditionally preserved specimens. So this is the reason why paleontology always has been primarily a morphological endeavor, and the fossils at least represent a set of taxa that provide potential information about evolution (e.g. Patterson and Rosen 1977; Schaeffer et al. 1972), which is nearly limited to morphological analysis. Moreover, an understanding of morphological variation in fossils requires an understanding of the morphology of living species.

It is well-known that wing shape of insects exhibits a high heritability in nature (Bitner-Mathé and Klaczko 1999; Moraes et al. 2004), and wing morphology is of a primary importance to entomologists interested in systematics (Su et al. 2015). It was Comstock (1893) who first popularized the use of insect wing venation for traditional classification (Kunkel 2004). Wing veins and their intersections are unambiguously homologous (Ross 1936), so since the 1970’s, several authors have begun to use the insect wings of morphometrical studies in systematics and phylogeny (Plowright and Stephen 1973; Rohlf 1993; Klingenberg 2003; Gumiel et al. 2003). Geometric morphometrics (GM) utilizes powerful and comprehensive statistical procedures to analyze shape differences of a morphological feature, using either homologous landmarks or outlines of the structure (Rohlf and Marcus 1993; Marcus and Corti 1996; Adam et al. 2004), and it is considered to be the most rigorous morphometric method (Gilchrist et al. 2000; Debat et al. 2003). Compared with other organs, the wing venation is unique and the examination of wing venation pattern shows many methodological advantages, because they are basically 2-dimensional and the venation provides many well-defined morphological landmarks (Gumiel et al. 2003), the interactions of the veins, which are easy for identification and able to capture the general shape of the wing (Bookstein 1991), as well as their rigidity and good conservation in either living or fossil specimens (Pavlinov 2001). Among insects, the use of GM analysis to study wing venation has been useful in identification at the individual level (Baylac et al. 2003; Dujardin et al. 2003; Sadeghi et al. 2009), in distinguishing sibling species (Matias et al. 2001; De la Riva et al. 2001; Villegas et al. 2002; Roggero and Dentrèves 2005; Aytekin et al. 2007; Francuski et al. 2009; Tüzün 2009) and in delimitation among the genera (Baracchi et al. 2011; Su et al. 2015) and higher taxonomic category (Bai et al. 2012, 2013). However, few studies have been conducted to quantify such wing variations for phylogenetic reconstruction. Thanks to the advent of the phylogenetic morphometric (PM) analysis method (Díaz-Cruz et al. 2021), it makes possibility to explore the relationships among the organisms based on the morphometric data.

The beetle family Cantharidae, commonly known solider beetles, is a large group with approximately 6,000 species in the world (Delkeskamp 1977; Kazantsev and Brancucci 2007). It is divided into five subfamilies (Brancucci 1980), based on a comprehensive comparative morphological study. In this classification, hindwing venation is one of the highly valued characters in the subfamilial level, and different from one another in the number of vein, cells and their length. In comparison, within each subfamily, the venation is stable and only exhibits quantitative variations among genera (Lanham 1951). Therefore, it is an ideal material to explore the relationships among the genera based on the hindwing variation through PM analysis.

In the present study, taking the cantharid beetles as the subject, we are going to apply GM to analyze the hindwing shapes of 16 representative genera of Cantharinae, based on which to explore their relationships by the PM analysis. Meanwhile, we shall reconstruct the phylogeny among these genera based on the (nearly) complete mitochondrial genomes by both Maximum-likelihood (ML) and Bayesian inference (BI) analyses. Finally, we will examine the accuracy of PM of hind wing shapes, by comparison with the produced topology of mitogenomes. Based on the results, we are able to evaluate the reliability of the hindwing shapes in inferring phylogenetic relationships among the generic taxa, also shed new lights on reconstruction of phylogenetic relationships, especially for those taxa rare, inaccessible or extinct organism relying on the morphology.

In the present study, the statistical test performed by TpsSmall suggested that our obtained data of hind wings (3 specimens were collected from each of 37 species amount to 111 samples, 13 landmarks for each sample) is acceptable for the geometric morphometric analysis. Further the CVA analysis suggested that all representative genera (a total of 16 genera) of Cantharinae can be distinguished from one another by the hind wing shapes, which is consistent with the previous study (Su et al. 2015).

In insects, the wing shapes of geometric morphometric analyses are usually applied in distinguishing the sibling species or uncovering the cryptic species (Baylac et al. 2003; Pizzo et al. 2006; Gurgel-Goncalves et al. 2011; Muñoz-Muñoz et al. 2011; Mitrovski-Bogdanović et al. 2013), since that wing GM analysis represents a reliable and rapid alternative that yields satisfactory results when discriminating between morphologically analogous species (Champakaew et al. 2021). Although the geometric morphometric data remains controversial in inferring the relationships among the organisms (Palci and Lee 2019), it has been applied in estimating the evolutionary relationships of some animals (Bogan and Roe 2008; Klingenberg 2015; Siriwut et al. 2015; Püschel and Sellers 2016; Hart et al. 2020; Goharimanesh et al. 2022), especially in the higher grades (tribes or subfamilies or families) of some insect groups, based on the shapes of pronotum and elytra (Acevedo 2015; Eldred et al. 2016; Zhang et al. 2019; Tong et al. 2021), as well as hind wings (Abou-Shaara and Al-Ghamdi 2012; Su et al. 2015; Barour and Baylac 2016). However, none was addressed on the phylogenetic relationships in the generic level. Herein, taking the Cantharinae as an example, we are the first to construct the phylogeny among the genera based on the geometric morphometric data of hind wings using different methods.

Prior to estimate the values of hindwing shape in inferring the phylogenetic relationships among the genera of Cantharinae, a phylogenetic analysis of mitochondrial genomes was constructed to make a standard reference. Mitochondrial genomes has been widely used in the phylogenetic studies in various insect groups, not only in higher grades (Negrisolo et al. 2011; Cameron 2014a, b; Li et al. 2015), but also in generic taxa (Coeur d’acier et al. 2007; Su et al. 2018). The mitophylogenetic trees (Fig. 3) showed that Cantharinae was divided into two large clades, which are corresponding to the morphologically defined Podabrini (Podabrus + Asiopodabrus) and Cantharini (the other genera) (Kazantsev and Brancucci 2007). Within Cantharini, the close relationships of several genera complex were recovered respectively, including Cyrebion and Themus (Yang and Yang 2010; Kopetz 2016), Falsopodabrus, Habronychus and Stenothemus (Wittmer 1974; Okushima and Satô 1999; Švihla 2004; Li et al. 2016), Cantharis, Taiwanocantharis and Taocantharis (Švihla 2011), Micropodabrus, Cephalomalthinus and Pseudopodabrus (Wittmer 1983, 1997; Yang et al. 2009, 2012). These genera have similar morphological characters probably resulted from close affinities respectively. Here their relationships are rigorously tested by the molecular phylogenetic analyses for the first time, and our obtained results are congruent with the morphological classification, which suggests that the constructed mitophylogenetic tree is good enough to be as a reference for the following comparison.

Compared with the above mitophylogenetic tree, the topologies produced by geometric morphometric data of hindwing shapes recovered the close relationships of the aforementioned genera using different methods (Fig. 4). Overall, the results of the phenograms based on both Procrustes distance (Fig. 4A) and Euclidean similarity metrics of Mahalanobis distance (Fig. 4B), as well as NJ analyses based on Mahalanobis distance (Fig. 4D) are quite consistent, but significantly different from that of MP analysis based on the two landmark configurations (Fig. 4C). However, the latter showed similar result as that of mitophylogenetic tree in the position of Podabrus. Except the four genera complex, some genera were always grouped together, including Lycocerus and Prothemus, Micropodabrus, Rhagonycha, Pseudopodabrus and Asiopodabrus, which are never recovered as sister groups by the mitophylogeny. This suggested that hindwing shape may be also convergent in the evolution, although which was usually considered as a character of high value in systematics, like other external morphological characters, such as antennae, pronotum, elytra, etc.

Conflict between morphological and molecular studies of phylogeny may be also resulted from differences in assumptions about the evolutionary process and differences in methods of analysis. The reasons for these differences may be allometric effects, homoplasy, accelerated evolution, genetic drift and, of course possible sampling or measurement errors (Cardini and O’Higgins 2004). In the case of such conflicting results, molecular sequence data are often favoured, as they are typically much more numerous and/or arguably perceived as being more objective (e.g. Jin et al. 2020). Given the strong statistical support and most groupings in all molecular analyses, and the quantity and suitability of mitochondrial data to elucidate phylogenetic relationships of closely related taxa, in this case we favor the results of molecular analyses to estimate the dependability of the PM analyses. Although no topology was produced in the PM analyses congruent with the mitophylogenetic tree, the close relationships of the allied genera were recovered in the phylogenetic geometric morphometrics, which suggested that the latter was helpful in inferring the relationships of sister groups. Thus, we propose the GM analyses should be extended to other morphological structures as well in future insect taxonomy research.

In the present study, taking the Cantharinae as an example, we evaluated the taxonomic value of the hindwing shapes in inferring phylogenetic relationships among the generic taxa of subfamily Cantharinae. A total of 111 hindwing samples representing 37 species belonging to 16 genera of Cantharinae were analyzed by GM analysis. The statistical test performed by TpsSmall suggested that our obtained data is acceptable for the geometric morphometric analysis, and the CVA analysis demonstrated that all representative genera of Cantharinae can be well separated by the hind wing shape. With the constructed mitophylogeny as reference, the PM analyses of the hindwing shapes data using different methods (MP analysis of the two landmark configurations, NJ analysis of Mahalanobis distance, phonograms of both Procrustes distiance and Euclidean similartiy metrics of Mahalanobis distance) showed that the sister relationships of allied genera or morphologically defined genera complex are always recovered. However, some genera in distant relationships sometimes are grouped together under PM analysis, probably due to the convergent evolution in the hindwing shapes. No matter how, the landmark-based hindwings shape GM analyses prove to be feasible in phylogenetic reconstruction and be helpful in recovering the sister relationships of allied genera.

The present study was supported by the National Natural Science Foundation of China (Nos. 32270491, 31772507), the Natural Science Foundation of Hebei Province (No. C2022201005), the Interdisciplinary Research Program of Natural Science of Hebei University (No. DXK202103) and the Excellent Youth Scientific Research and Innovation Team of Hebei University (No. 605020521005).

We wish to express our deepest thanks to Prof. Xingke Yang (Institute of Zoology, Chinese Academy of Sciences, Beijing, China) for his guidance to the senior corresponding author in studying on the taxonomy of Cantharidae. We are very grateful to the editors and anonymous reviewers for careful scrutiny and useful comments for improving the manuscript.