Zoological Journal of the Linnean Society, 2018, 182, 38–49. With 2 figures. Parasitoid–host associations of the genus Coccophagus (Hymenoptera: Aphelinidae) in China QING-SONG ZHOU1,2, ANDREW POLASZEK3, YAO-GUANG QIN1,2, FANG YU1,2, XU-BO WANG4, SAN-AN WU4, CHAO-DONG ZHU1,2, YAN-ZHOU ZHANG1,2* and PAOLO ALFONSO PEDATA5 1Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, 1 Beichen West Road, Chaoyang District, Beijing 100101, China 2University of Chinese Academy of Sciences (UCAS), No. 19A Yuquan Road, Beijing 100049, China 3Department of Life Sciences, Natural History Museum, Cromwell Road, London SW7 5BD, UK 4The Key Laboratory for Silviculture and Conservation of Ministry of Education, Beijing Forestry University, No.35 Tsinghua East Road Haidian District, Beijing 100083, China 5Istituto per la Protezione Sostenibile delle Piante CNR, SS Portici, Via Università 133, 80055 Portici (NA), Italy Received 1 December 2016; revised 13 February 2017; accepted for publication 19 March 2017 Host relationships among many Aphelinidae are complex due to their heteronomous reproductive behaviour, where males have different host relationships from females. Heteronomous parasitoids present a fascinating problem in host selection and sex ratio decision making. Accurate identification of insect parasitoids is a prerequisite in the determination of parasitoid–host relationships. With the goal of disentangling the parasitoid–host associations in the genus Coccophagus, we compared the performances of the COI barcode and of the 28S-D2 rRNA region for iden- tification and delimitation of 17 morphospecies of Coccophagus parasitoid wasps collected during a 9-year rearing programme of scale insects in China. Molecular data were analysed by two different methods of species delimitation, the automatic barcode gap discovery and the general mixed yule coalescent. Both methods were effective in discrimi- nating all previously recognized morphospecies. The congruence of morphospecies delimitation with that obtained by DNA barcode and nuclear gene data greatly enhanced our ability to unravel the parasitoid–host associations of genus Coccophagus. Most Coccophagus species (11 out of 17) can use host species belonging to different genera, with different levels of host specificity, particularly in host selection to produce female offspring. Sex-related differences in host relationships have also been detected and discussed. ADDITIONAL KEYWORDS: Aphelinidae – generalist parasitoid – mitochondrial gene –nuclear gene – species delimitation. INTRODUCTION Wagner, 1988; Jervis, Kidd & Walton, 1992), the practi- cal reasons mainly include (1) the need for an exhaus- The parasitoid–host relationship has long been a tive survey of potential hosts for a given parasitoid; focus of behavioural ecology, evolutionary biology and (2) the immense effort required for identification of community ecology (Price, 1980; Futuyma & Moreno, parasitoids, hosts and host plants and (3) interspe- 1988; Kawecki, 1998; Derocles et al., 2014). Limited cies/intraspecies variation, a long-standing problem parasitoid–host data, particularly host range and in species delimitation. Among them, the fundamental host-use patterns, are often an obstacle to under- issue is the accurate identification of the interacting standing true relationships (Hawkins, 1994; Shaw, partners, particularly the identification of sexually 1994; Quicke, 1997). As previously noted (Whitfield & dimorphic parasitoid wasps. Because descriptions for many parasitoids are based on the female sex only, *Corresponding author. E-mail: [email protected] and descriptions of males are often uninformative due © 2017 The Linnean Society of London, Zoological Journal of the Linnean Society, 2018, 182, 38–49 38 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/182/1/38/3871338 by Institute of Zoology, CAS user on 14 December 2017 PARASITOID–HOST ASSOCIATIONS OF COCCOPHAGUS 39 to a general lack of useful characters, biological data whiteflies, whereas males may develop as ectoparasi- associated with the male sex in species inventories toids of the same (or closely related) host species or as are scarce or unreliable. DNA barcoding, as an identi- hyperparasitoids of conspecific females or other endo- fication tool, has greatly facilitated the discrimination parasitoid species (Williams & Polaszek, 1996; Hunter of both parasitoids (Zhang et al., 2011; Chesters et al., & Woolley, 2001). While there is increasing informa- 2012) and host species (Deng et al., 2012, 2016; Wang tion on host-use patterns for some parasitoid groups et al., 2015, 2016). Recent efforts to estimate parasi- (Smith et al., 2006, 2008; Zhang et al., 2011; Chesters toid–host relationships show that without DNA-based et al., 2012), few such studies have been conducted on discrimination of parasitoid and host species, errone- Aphelinidae. Their actual host range remains to be ous parasitoid–host relationships may be diagnosed investigated for many species because of very scant due to inadvertent labelling of morphologically simi- knowledge of the systematics of both hosts and para- lar, but genetically isolated, lineages as a single spe- sitoids (Viggiani, 1984; Hayat, 1998; Kim & Heraty, cies (Smith et al., 2006, 2008; Chesters et al., 2012). 2012). Consequently, sex-specific host usage, which The combination of DNA-based discrimination of par- may follow independent evolutionary trajectories in asitoids with morphological data and host records has the two sexes (Williams & Polaszek, 1996; Hayward, greatly improved our knowledge of the ecological rela- McMahon & Kathirithamby, 2011), has not yet been tionships of Diptera and Hymenoptera parasitoids exhaustively investigated. (Smith et al., 2006, 2008; Zhang et al., 2011; Chesters In China, an intensive survey of numerous scale et al., 2012). insects, and rearing of insect parasitoids, has been The Aphelinidae (Hymenoptera: Chalcidoidea) is conducted (Zhang et al., 2011; Chesters et al., 2012). an important group with remarkable economic impor- From 2007 to 2015, we processed more than 2000 col- tance as biological control agents against whitefly and lections of scale insects, gathering data on host plants scale insect pests (Clausen, 1978; Greathead & Waage, and parasitoids belonging to the genus Coccophagus. 1986). The subfamily Coccophaginae has host relation- We selected species of Coccophagus because, in addi- ships that are different for males and females, and tion to their economic importance as natural enemies therefore are defined as heteronomous (Walter, 1983; and their heteronomous lifestyle, many host records Williams & Polaszek, 1996). These unusual host rela- are available from the literature (Compere, 1931; tionships were described by Flanders (1936, 1937), with Annecke & Insley, 1974; Huang, 1994; Hayat, 1998). subsequent contributions by Zinna (1961), Yasnosh In this study, we sought to ascertain whether pat- (1976) and Viggiani (1981). Walter (1983) gave a sim- terns of sequence variation in COI and 28S-D2 rRNA plified but comprehensive classification, proposing the are congruent with each other, and with morphology, term ‘heteronomous hyperparasitoids’ for those spe- in revealing species boundaries. In particular, we cies in which males develop as hyperparasitoids at the employed molecular evidence to test whether mor- expense of the preimaginal stages of conspecific and/or phologically similar males from different hosts repre- heterospecific parasitoid larvae, either on the same or sented a single species and consistently matched the different primary hosts as conspecific females. He also putative corresponding females. On the basis of solid proposed two further categories: (1) diphagous para- taxonomic work, we investigated the host range and sitoids, in which sexes are both primary parasitoids, host-use pattern of Coccophagus species occurring but females develop as endoparasitoids and males as in China, to gain more data on biased sex ratios and ectoparasitoids and (2) heterotrophic parasitoids, in differential host usage related to their heteronomous which conspecific females and males develop on com- reproductive behaviour. pletely different hosts (e.g. whiteflies and Lepidoptera eggs, respectively), both as primary parasitoids. Outside Coccophaginae, heteronomous parasitoids are MATERIAL AND METHODS known only in the family of Myrmecolacidae in the ‘twisted winged flies’ Strepsiptera (Kathirithamby, SAMPLING, REARING AND 2009), in this case as heterotrophic parasitoids. MORPHOLOGICAL IDENTIFICATION Most hosts of Coccophaginae belong to the Hemiptera All samples of Coccophagus species (Supporting Sternorryncha (Aleyrodoidea, Coccoidea, and rarely Information, Table S1) were reared from wild-collected Aphidoidea and Psylloidea) (Evans, Polaszek & plant material infested with scale insects across China Bennett, 1995; Polaszek, Evans & Bennett, 1992). In from 50 sites in 22 provinces (Supporting Information, addition, systematically unrelated hosts such as eggs Fig. S3). Collection of hosts and rearing of parasitoids of Heteroptera, Auchenorrhyncha and Lepidoptera followed methods described in Zhang et al. (2011). may be used (Yasnosh, 1979; Viggiani, 1981; Polaszek, Parasitoids and representatives of their scale insect 1991; Polaszek & Luft Albarracin, 2011). Females are hosts were stored in 95% ethanol for subsequent usually primary endoparasitoids of scale insects or study. Preliminary morphological identification