taxonID	type	description	language	source
03DA2F25FFADFFE6FC84FBB6526C51A5.taxon	description	Morphological traits Morphological traits used for species and subspecies circumscriptions relied primarily on wing morphology, particularly the: 1) extent (including presence / absence) of silvery-blue wing scaling (which creates iridescence structurally) 2) extent of “ solid blue ” of the hindwing upper side 3) density of silvery-blue on both wings (thereby allowing or preventing visibility of underside markings through to the upper side) 4) clarity of hindwing underside markings (plain vs. well marked) 5) pallidity of wings (hindwing underside, or in general) 6) hindwing and forewing shape 7) length of the hindwing “ tail ” Some of these traits are useful only to a limited degree (either for a single taxon or a small group of taxa, e. g., hindwing shape for C. chrysaor natalensis, forewing shape for C. brooksi and C. palmus, and short hindwing “ tail ” for C. chrysaor, C. chrysaor natalensis, and C. phosphor). Preliminary investigations suggest that the length of the upper row of lateral setae in early first instar larvae appears to be fairly constant within a given species, and the length and shape of the tubercles housing the tentacular organs appear to differ significantly among Chrysoritis species (details in HEA 23 a). Thorough investigation of these traits is beyond the scope of this paper but should be considered in future taxonomic efforts. Ecological traits Heath & Pringle (2007: 38) suggested that any species within the thysbe clade could utilise the host plant of any other species within the clade, and so all species in the thysbe clade have been treated as polyphagous. Thus, host plant data are considered to be of little use in Chrysoritis taxonomy (see also Cottrell, 1984: 41). TEA 20 showed that Chrysoritis species show remarkable overlap in climatic niche, thus that trait is also not useful for taxonomy. The following ecological traits were examined for their potential in delimiting species: 1) host ant species, 2) male patrolling terrain, and 3) range overlap (sympatry vs. allopatry; see Note S 1) with other Chrysoritis taxa. These traits are discussed below.	en	Quek, Swee-Peck, Pringle, Ernest L., Heath, Alan (2022): Chrysoritis Butler (Papilionoidea: Lycaenidae: Aphnaeinae) - Part I: Molecular phylogenetic analyses of a South African genus of myrmecophilous butterflies. Metamorphosis 33 (1): 107-126, DOI: 10.4314/met.v33i1.13, URL: https://doi.org/10.4314/met.v33i1.13
03DA2F25FFADFFE6FC84FBB6526C51A5.taxon	biology_ecology	Host ants and plants Host plants and ants are listed together with original sources in Heath et al. (2008), and TEA 20 provided a broad summary of host plant genera in their Table S 5. Following revised IDs of the host ants by B. Blaimer (see below) a revised list of plant and ant hosts is given in Table S 1 in HEA 23 b. The two ant genera recorded as associated with Chrysoritis species are the ‘ Droptail’ ants Myrmicaria Saunders, 1842 and the ‘ Cocktail’ ants Crematogaster Lund, 1831, both in Myrmicinae. The majority of Chrysoritis species are associated with Crematogaster; only two species are recorded with Myrmicaria – C. oreas and C. pyroeis. Host ant identification The absence of the " correct " species of host ant, assumed to be chemically detected by the gravid female, may deter oviposition. Dr Hamish Robertson provisionally identified the ants mentioned herein from samples collected in the field that were found in association with Chrysoritis larvae and adults. Some of these identifications were subsequently revised by B. Blaimer (Smithsonian National Museum of Natural History) and endorsed by H. G. Robertson. The authors also referred to Peter Slingsby’s recent guidebook (2017) to confirm ant identifications. The Crematogaster species currently known to associate with Chrysoritis are:	en	Quek, Swee-Peck, Pringle, Ernest L., Heath, Alan (2022): Chrysoritis Butler (Papilionoidea: Lycaenidae: Aphnaeinae) - Part I: Molecular phylogenetic analyses of a South African genus of myrmecophilous butterflies. Metamorphosis 33 (1): 107-126, DOI: 10.4314/met.v33i1.13, URL: https://doi.org/10.4314/met.v33i1.13
03DA2F25FFAEFFE7FF3FF85852C6502E.taxon	description	Unresolved cryptic species diversity is prevalent among ants. Even putatively well-known ant species like Lasius niger and Lasius alienus in Europe have turned out to be species complexes. Few of the ant species-level identifications listed above have been scrutinised by myrmecologists, and thus may be suspect (Fiedler, 2021). For example, Crem. peringueyi and Crem. liengmei are widespread common species that could potentially contain cryptic species (see also Bickford et al., 2007) but they have yet to be studied in detail to determine if this could be the case here (B. Blaimer, pers. comm.). Male patrolling terrain (MPT) MPT is an informative trait in circumscribing Chrysoritis species. In some taxa, males congregate at specific topographic features (e. g., hilltop, gulley etc.), awaiting the arrival of newly eclosed virgin females. Likewise, virgin females will seek out particular terrain features to find their mate soon after eclosing. This phenomenon, possibly a form of lekking, is sometimes known as " hilltopping. " Specificity in MPT maximises the chances of finding a mate of the same species and could serve as a prezygotic barrier to gene flow between species in sympatry. MPT specificity appears to be characteristic of thysbe clade species and has so far not been observed in species outside the thysbe clade. Males of C. dicksoni (outside the thysbe clade) show aggregating behaviour but it seems not to be associated with a consistent topographical feature (see HEA 23 a). In non- thysbe clade species (noting the exception of C. dicksoni), most males and females congregate and mate near their host plants, and some species (e. g. C. zonarius) prefer to stay very close to their host plant. Analyses of molecular data Phylogenetic analyses Molecular data for population and phylogenetic analyses are from TEA 20. The 406 Chrysoritis samples reported therein were found to contain a few samples that were sequenced twice for COI. After removing the duplicates, the total number of Chrysoritis samples was 399 (excluding 4 outgroup samples), of which all have sequences of COI (1220 base pairs [bp]), 97 have EF (1039 bp) and CAD (745 bp) sequences, and 98 have H 3 (328 bp) sequences. We also discovered that the COI sequence for sample AH 06 M 581 (identified as C. pan but appearing as part of C. stepheni) was likely the result of contamination and we deduced the correct sequence (see Note S 1). The trees presented by TEA 20 based on all four genes combined lacked well supported resolution in the thysbe clade. Examination of the nuclear gene trees (CAD, EF and H 3, unpublished, courtesy of G. Talavera) showed scant resolution within the thysbe clade. Thus we performed further analyses on the COI data alone (1220 nucleotides). Identical sequences were identified using Arlequin 3.5. (Excoffier et al., 1992; Excoffier & Lischer, 2010) and removed prior to phylogenetic analyses, resulting in a sample set of 270. Maximum likelihood (ML) analyses were performed on the unpartitioned COI dataset using IQTree (Nguyen et al., 2015; Minh et al., 2020) run on IQTree’s web server (Trifinopoulos et al., 2016) at https: // www. hiv. lanl. gov / content / sequence / IQTREE / iqtre e. html using the “ find best and apply ” substitution model setting (ModelFinder, Kalyaanamoorthy et al., 2017, resulting in the TVM + F + R 3 model selected using AIC, BIC and AICc), and branch support was obtained from 1000 repetitions of ultrafast bootstrap (Hoang et al., 2018). ML analyses were also performed on the three nuclear genes separately using the same software and settings as for the COI data. The aligned and combined nuclear gene dataset matrix is available for download at https: // doi. org / 10.6084 / m 9. figshare. 19225203 and the COI dataset matrix at: https: // doi. org / 10.6084 / m 9. figshare. 19225101. Haplotype network construction A statistical parsimony network (Templeton et al., 1992) of COI haplotypes (with duplicates removed as described above) from thysbe clade (sensu TEA 20) was constructed using TCS 1.21 (Clement et al., 2000), and the network visualised using the program tcsBU (Murias dos Santos et al., 2016). The 334 samples within the thysbe clade collapsed into 200 unique haplotypes; however, some haplotypes differed only by missing data. A list of samples sharing identical haplotypes is provided in Table S 1, available at: https: // doi. org / 10.6084 / m 9. figshare. 19225038. AMOVAs To ascertain the degree to which genetic structure in the COI data of the thysbe clade can be explained by taxonomic designation versus geographic distribution, we performed analyses of molecular variance (AMOVAs) using Arlequin 3.5. (Excoffier et al., 1992, Excoffier & Lischer, 2010). AMOVAs were run using four grouping schemes: 1) by species, 2) by subspecies, 3) by region, and 4) by locality group (comprising a cluster of localities in the same vicinity – see Table S 2). In addition to the overall fixation index (ΦST), fixation indices for all pairwise comparisons were calculated to determine how differentiated a species or subspecies was from its sister taxon. Two samples were excluded from the AMOVAs: AH 06 M 581 (C. pan which had a contaminated COI sequence, see above) and AH 12 C 011 (a brooksi x rileyi hybrid). Genetic distances In this study, the authors used COI data from TEA 20 to generate a pairwise COI distance matrix of Chrysoritis samples. Genetic distance is not used to determine rank designations or species-level splits. This table is available as a. xlsx file at: https: // doi. org / 10.6084 / m 9. figshare. 17241566. v 1.	en	Quek, Swee-Peck, Pringle, Ernest L., Heath, Alan (2022): Chrysoritis Butler (Papilionoidea: Lycaenidae: Aphnaeinae) - Part I: Molecular phylogenetic analyses of a South African genus of myrmecophilous butterflies. Metamorphosis 33 (1): 107-126, DOI: 10.4314/met.v33i1.13, URL: https://doi.org/10.4314/met.v33i1.13
