Revision of the Conwentzia Enderlein, 1905 (Neuroptera, Coniopterygidae) of China, combining morphological and molecular characters

Abstract The Chinese species of Conwentzia Enderlein are revised by integrating morphological characters and molecular data. Conwentzia yunguiana Liu & Yang, 1993 is proposed as a junior synonym of Conwentzia nietoi Monserrat, 1982, syn. nov. and Conwentzia orthotibia Yang, 1974 is proposed as a junior synonym of Conwentzia pineticola Enderlein, 1905, syn. nov. Moreover, a key to the adult males of the Conwentzia from China and DNA barcodes are provided.


Introduction
The dusty lacewing genus Conwentzia belongs to the subfamily Coniopteryginae of family Coniopterygidae and it is a species-poor genus, including only fourteen described species (Sziráki 2011;Oswald 2020). However, Conwentzia is relatively widespread, and widely distributed in the Palaearctic, Nearctic, Oriental and Afrotropical regions (Meinander 1972;Sziráki 2011). This genus was originally erected by Enderlein (1905) based on the type species Conwentzia pineticola Enderlein, 1905.

Imaging
Specimens were examined with an Optec SZ760 stereomicroscope. Photos were taken with a Nikon D5300 digital camera attached to a Leica DM2500 stereomicroscope. The resulting images were edited and processed with Adobe Photoshop CC 2018.

DNA extraction and sequencing protocols
Total genomic DNA was extracted based on the method of Lu et al. (2018) with the commercial Ezup Column Animal Genomic DNA Purification Kit (Sangon Biotech, China) and following the manufacturer's protocol. The PCR primer and reaction conditions for the COI region followed Folmer et al. (1994) and Lu et al. (2018). Products were sequenced in a single direction by Sangon (Shanghai) Co., Ltd. Sequences were edited and analyzed using the software Chromas version 2.3 and BioEdit 7.0.4.1 (Hall 1999). In addition, sequences were translated into amino acids to check for NUMTS and test for quality.

Sequence analysis
The barcoding gap was assessed by means of the following methods. Pairwise genetic distances for COI genes were computed with the Kimura 2-parameter (K2P) method in the MEGA 6.0 software (Tamura et al. 2011). Finally, all sequences were deposited in GenBank. In order to better analyze the data, sequences of C. pineticola from Bavaria in Germany were downloaded from GenBank. The accession numbers of these sequences are as follows: JN299372, JN299373, JN299374, JN299347, JN299348.
Comments. There are fourteen species in the genus Conwentzia. The species described before 1972 are well known thanks to Meinander's (1972) comprehensive revision. However, Conwentzia inverta Withycombe, 1925 was not redescribed by Meinander (1972) because the type material in the Natural History Museum, London is in rather bad condition (Meinander 1972). However, Monserrat found that Barnard had a specimen collected from the type locality, Pusa, in India, by Withycombe in 1925, which he examined and used for his redescription of C. inverta (Monserrat 1982). Species described after 1972 are well known thanks to Sziráki's (2011) comprehensive revision. Thus, all species in the genus Conwentzia are relatively well known.
Key to Chinese species of Conwentzia (males) Note: Conwentzia fraternalis Yang, 1974 is not included in the key as the specimen is only known based on a single female.
Head (Fig. 5a). Yellowish-brown. Compound eyes large and dark. Antennae 36-37-segmented in males and 32-36-segmented in females. Scape relatively broad and blunt. Pedicel cylindrical, longer than wide. Antennae brown; scape light brown in some specimens. Scape broad and blunt. Pedicel cylindrical and longer than broad. Maxillary and labial palpus brown.

DNA barcoding
For the three Conwentzia species treated in this paper, accession numbers of DNA barcodes are listed in Table 1.

Genetic divergence among species
The average intraspecific genetic distance based on the K2P model was 0.10% for Conwentzia sinica Yang, 1974, 0.10% for Conwentzia orthotibia Yang, 1974, 0.10% for Conwentzia yunguiana Liu & Yang, 1993, and 0.52% for C. pineticola. The average interspecific genetic distance based on the K2P model was 2.19% Table 2. Intra-and interspecific Kimura 2-parameter average divergence values (%) of the COI gene analyzed by the MEGA 6.0 software. * = sequences of C. pineticola from Bavaria in Germany downloaded from GenBank. between C. orthotibia and C. pineticola. The other average interspecific genetic distances based on K2P model were 11.14-14.54%. The results (Table 2) showed that all intraspecific genetic distances were less than 2.0%, and all the interspecific genetic distance values exceeded 10% (except for the C. orthotibia and C. pineticola genetic distance).

Discussion
Conwentzia sinica is similar to C. inverta but differs in the shape of the male genitalia. Conwentzia sinica is characterized by a slender gonapophyses 10 in lateral view (Fig. 2a, b), while it is short (Monserrat 1982: 26, fig. 39) in C. inverta, therefore C. sinica is three times longer than C. inverta for gonapophyses 10. Moreover, the basal part of gonapophyses 10 is broad and blunt ventrally in C. sinica but acute in C. inverta. Furthermore, gonocoxites 10 are rectangular caudally in C. sinica (Fig. 2c, d) but oval in C. inverta (Monserrat 1982: 26, fig. 38). The morphological differences between C. sinica and C. inverta are mainly centered around the gonapophyses 10 and gonocoxites 10. However, both structures are almost transparent, requiring careful examination. For C. yunguiana, we found that those specimens do not have clear differences after comparison of the type specimens of C. yunguiana with the description of Conwentzia nietoi Monserrat, 1982. Nevertheless, the distal part of gonocoxites 11 is blunt laterally in C. yunguiana (Fig. 4a, b), while it is acute in C. nietoi (Monserrat 1982: 26, fig. 34). The differences are mainly centered around the distal part of gonocoxites 11 in lateral view, which may be caused by the arched shape above the gonocoxites 10 in lateral view. Besides, the rim is so obscure for the gonocoxites 11 is membranous and transparent in distal part. And we also discussed with György Sziráki, who examined the type specimen of C. nietoi, and his opinion is the same as ours. Therefore, we ascribe the differences in gonocoxites 11 to intraspecific morphological variation.
We found no clear morphological differences between C. orthotibia and C. pineticola after comparison of the type specimens of C. orthotibia with the description of C. pineticola. Nevertheless, the distal part of gonocoxites 11 is wavy caudally in C. orthotibia (Fig. 6c, d), while it is arched in C. pineticola (Meinander 1972: 300, fig. 195F). We were not sure whether such differences should be ascribed to intraspecific morphological variation between C. orthotibia and C. pineticola. The type species was described from Berlin in Germany (Enderlein 1905) and we obtained DNA barcodes of C. pineticola from Bavaria in Germany from NCBI. The results show that the mean interspecific divergence between C. orthotibia and C. pineticola was 2.19%, which is inconsistent with Morinière et al.'s (2014) suggestion that the mean interspecific divergence is 10-20% in the Coniopterygidae, Hemerobiidae, and Myrmeleontidae. Our results suggest that the differences between C. orthotibia and C. pineticola are intraspecific. Liu & Yang, 1993 is proposed as a junior synonym of Conwentzia nietoi Monserrat, 1982, syn. nov. andConwentzia orthotibia Yang, 1974 is proposed as a junior synonym of Conwentzia pineticola Enderlein, 1905, syn. nov. In this study, we added three species barcodes to the Conwentzia DNA library and the mean intraspecific divergence was 11.14-14.54% for the species analysed.