Contributions to Zoology, 85 (1) 1-11 (2016) Intraspecific morphological variation of the scorpionfly Dicerapanorpa magna (Chou) (Meco­ptera: Panorpidae) based on geometric morphometric analysis of wings Mei Liu1, Na Ma1, 2, Bao-Zhen Hua1, 3 1 State Key Laboratory of Crop Stress Biology for Arid Areas, Entomological Museum, Northwest A&F University, Yangling, Shaanxi 712100, China 2 School of Agricultural Engineering, Henan Provincial Key Laboratory of Funiu Mountain Insect Biology, Nanyang Normal University, Nanyang, Henan 473061, China 3 E-mail: [email protected] Key words: canonical variates analysis, individual variation, sexual dimorphism Abstract Stillwell et al., 2005; Neto et al., 2013). Morphological variation provides a major source of characters and Wing shape variation was investigated between the sexes and character states in traditional taxonomy (Michaux, Dicerapanorpa mag- among four populations of the scorpionfly 1989; McPeek, 1990). However, it is often difficult to na (Chou, 1981) endemic to the Qin-Ba Mountains area, China through the landmark-based geometric morphometric approach. determine whether variation is intraspecific or inter- The results show that sexual dimorphism exists both in wing specific in taxonomic practice (Nedeljković et al., 2013; size and shape in D. magna. Significant differences exist in fe- Riedel et al., 2013; Barão et al., 2014). male wing size and shape among D. magna populations. The Landmark-based geometric morphometrics is an possible reasons of the wing variation are discussed based on the effective tool to quantify and analyze the overall shape divergence time of D. magna in combination with the tectonic and climatic events in the Qin-Ba Mountains during the late of biological structures based on two- or three-dimen- Miocene-Pleistocene period. Whether reproductive isolation ex- sional Cartesian coordinates of landmarks (Bookstein, ists between different populations needs further research. 1991; Adams et al., 2013). It has been widely applied in insect taxonomy and systematics to detect subtle mor- phological variation (Francuski et al., 2009; Baracchi Contents et al., 2011; Neto et al., 2013; Pepinelli et al., 2013), which is often undetectable with classical morphomet- Introduction ......................................................................................... 1 ric methods (Zelditch et al., 2004). Material and methods ........................................................................ 2 Insect wings are two-dimensional flattened struc- Insect sampling ............................................................................. 2 Geometric morphometric and statistical analyses .............. 3 tures that bear many homologous landmarks and are Results ................................................................................................... 3 suitable for geometric morphometric analysis (Grod- Variation between D. magna and D. baiyunshana ............... 3 nitsky, 1999; Zelditch et al., 2004). They have been Sexual dimorphism of D. magna ............................................... 5 successfully used to resolve taxonomic problems in Wing size variation among populations of D. magna ......... 6 Wing shape variation among populations of closely related species groups (Palmer, 2004; Aytekin D. magna ..... 7 et al. et al et al. Discussion ............................................................................................ 8 , 2007; Baracchi ., 2011; Neto , 2013; et al. Acknowledgements ............................................................................ 9 Barão , 2014) and to discriminate populations References ............................................................................................ 9 within a species (Haas and Tolley, 1998; Pepinelli et al., 2013). The scorpionflies Panorpidae (Insecta: Mecoptera) Introduction usually live in large populations in moist forested hab- itats (Byers and Thornhill, 1983). Since the adults are Morphological variation is prevalent in wide-ranging relatively weak fliers (Thornhill, 1980; Byers and species (Mayr, 1963) and is related to phenotypic plas- Thornhill, 1983), gene exchange between discrete pop- ticity, physiological response to environmental factors ulations is greatly limited, if not completely broken and adaptation to local environments and divergent off. In this case, morphological variation is prevalent selection, potentially leading to speciation through the in Panorpidae, especially the wing markings and geni- evolution of reproductive barriers (McPeek, 1990; talia (Ward, 1979, 1983; Jones, 2010; Ma et al., 2014). Downloaded from Brill.com10/05/2021 11:29:46AM via free access 2 Liu et al. – Morphological variation of Dicerapanorpa magna Fig. 1. Localities of the populations analyzed: JLJ (Jialingjiang Source For- est Park, Shaanxi); TB (Mt. Taibai, Shaan­xi); HDT (Huoditang Forest Farm, Shaan­xi); NGS (Nangongshan National Forest Park, Shaanxi); BYS (Baiyunshan, Henan). The Qin-Ba Mountains area is composed of the Qin- using landmark-based geometric morphometrics to ling Mountains and Daba Mountains in central China quantify wing size and shape variations between the (Liang, 2002). These two mountains are continuous sexes and among allopatric populations. and are divided by the Hanshui River and a series of basins. The Qinling Mountains are the prominent boundary between the Oriental and Palearctic Regions Material and methods in China and are rich in the biodiversity of Mecoptera (Chou et al., 1980; Ying, 1994; Yuan et al., 2012). Insect sampling Dicerapanorpa magna (Chou in Chou et al., 1981) was originally described based on a female holotype To assess intra- and interspecific variation of D. magna, collected from Huoditang at the southern slope of the the closely related species D. baiyunshana Zhong & Qinling Mountains. Later, this scorpionfly species was Hua, 2013 was included included in the analysis. Owing found to be widely distributed in the Qin-Ba Moun- to limited sample size of D. baiyunshana males, inter- tains (Hua and Chou, 1997; Nie and Hua, 2004). Previ- specific analysis of wing variations was conducted ous research reveals that individuals of D. magna from only for females. Ten female specimens of D. baiyun- different areas exhibit variation both in external mor- shana were collected from Baiyunshan (BYS, 1300 phology and internal anatomy, including the number m), Henan Province in 2001. Specimens of D. magna of female ovarioles and male salivary gland tubes (76 males and 172 females) were collected from four (Hou and Hua, 2008; Ma et al., 2011). This led us to regions in Shaanxi Province from June to August in hypothesize whether these variations are intraspecific 2013, including 24 males and 48 females from the or interspecific. Huoditang Forest Farm (HDT, 1500-1800 m), 22 males In this paper, different populations of D. magna and 40 females from the Jialingjiang Source Forest collected from the Qin-Ba Mountains were examined Park (JLJ, 1500-1700 m), 8 males and 44 females from Fig. 2. Right forewing of D. magna, showing the landmarks used in this study. Downloaded from Brill.com10/05/2021 11:29:46AM via free access Contributions to Zoology, 85 (1) – 2016 3 Nangongshan National Forest Park (NGS, 1700-1800 rotated against the consensus configuration using m), and 22 males and 40 females from Mt. Taibai (TB, Generalized Procrustes Analysis (GPA) procedure to 2300-2500 m). The localities of populations analyzed remove the nonshape effects of size, position and ori- are marked on the map (Fig. 1). entation (Rohlf and Slice, 1990; Dryden and Mardia, 1998). The resulted matrix (w; ‘weight matrix’ sensu Geometric morphometric and statistical analyses Rohlf et al., 1996) was used for shape analysis. For a shape, a principal component analysis (PCA) was car- Right forewings were carefully removed from 75% al- ried out to determine the explained percentage of each cohol-preserved specimens and temporarily mounted principal component (PC) of the total variation. The on micro-slides. Photographs with the same scale were total shape variables were used for the multivariate taken with a digital camera attached to the Nikon analysis of variance (MANOVA) to test wing shape SMZ­1500 stereo zoom microscope (Nikon, Tokyo, Ja- differences within and among species/population. Ca- pan). nonical variate analysis (CVA) and linear discrimi- Twenty-three homologous landmarks (LMs) at vein nant analysis (LDA) were used to discriminate popu- intersections or terminations that could be reliably lations and provide shape variations associated with identified were selected (Fig. 2), and can be considered canonical variates (CVs). The percentages of correct type I landmarks (Bookstein, 1991). Data for wing size classification (hit ratio, HR sensu Gerard et al., 2015; and shape were obtained by positioning landmarks on Huberty and Olejnik, 2006) based on a leave-one-out digitized wings using tpsDig2 (Rohlf, 2004). cross-validation procedure in LDA were used to eval- Measurement error. An analysis of measurement uate the discriminatory power of the wing. The allo- error was conducted on a subsample of 30 specimens metric effect or the change in shape associated with of D. magna, which were chosen randomly and repeat- size differences was evaluated with a multivariate re- ed 2 times to obtain landmark data. Procrustes ANO- gression of shape variables onto size. Morphometric VA (analysis of variance) was performed for landmark and statistical analyses were computed using the