Chromosomal data are important for taxonomists, cytogeneticists and evolutionary biologists; however, the value of these data decreases sharply if they are obtained for individuals with inaccurate species identification or unclear species identity. To avoid this problem, here we suggest linking each karyotyped sample with its DNA barcode, photograph and precise geographic data, providing an opportunity for unambiguous identification of described taxa and for delimitation of undescribed species. Using this approach, we present new data on chromosome number diversity in neotropical butterflies of the subfamily Biblidinae (genus Vila Kirby, 1871) and the tribe Ithomiini (genera Oleria Hübner, 1816, Ithomia Hübner, 1816, Godyris Boisduval, 1870, Hypothyris Hübner, 1821, Napeogenes Bates, 1862, Pseudoscada Godman et Salvin, 1879 and Hyposcada Godman et Salvin, 1879). Combining new and previously published data we show that the species complex Oleria onega (Hewitson, [1852]) includes three discrete chromosomal clusters (with haploid chromosome numbers n = 15, n = 22 and n = 30) and at least four DNA barcode clusters. Then we discuss how the incomplete connection between these chromosomal and molecular data (karyotypes and DNA barcodes were obtained for different sets of individuals) complicates the taxonomic interpretation of the discovered clusters.

karyotype, DNA barcoding, COI, meiosis, metaphase, Lepidoptera, Nymphalidae, Biblidinae, Danainae, Ithomiini, Peru

Chromosomal data are an important source of information for taxonomic, evolutionary and comparative phylogenetic studies (White 1973). However, the application of these data is often difficult because of unclear taxonomic identity (e.g. Petrova et al. 2015) or doubtful species identification (or even due to the lack of a species identification) of the samples that were used as vouchers for karyotype analysis [e.g. some samples and identifications in Robinson (1971) and Brown et al. (2004)]. Theoretically, one can try to find these samples, provided that they were neatly labeled and can be recognized, are stored in accessible museums and have not been lost, and then check their identification using taxonomic literature or comparison with type specimens. However, it is complicated and almost impossible in practice.

To avoid this problem, here we suggest linking each karyotyped sample with its DNA barcode. It was empirically demonstrated that the mitochondrial DNA barcode, a relatively short fragment of the mitochondrial COI gene (658 base pairs) (i.e., a negligible part of the genome in terms of size), could differentiate up to 95% of species in many taxa (Hebert et al. 2003, 2004; Hebert and Gregory 2005; Hajibabaei et al. 2006; Lukhtanov et al. 2009). In addition, the barcoding DNA protocol provides a standardized system for storing information on vouchers that served as the basis for DNA barcoding, including the image, the exact label and the storage location of the samples. This makes it possible, if necessary, to relatively easily find and re-examine a voucher.

Obtaining barcodes is currently a simple technical task, which can be carried out in almost any laboratory or on a commercial basis. Our personal experience, based on a molecular analysis of the fauna of Central Asia, Eastern Europe and Western Asia (Lukhtanov et al. 2009, 2016; Lukhtanov 2017; Pazhenkova and Lukhtanov 2019), shows that if there are barcode libraries (Ward et al. 2009; Dincă et al. 2011) for a given region and for a given taxonomic group, barcodes ensure almost 100% success of species identification. Even if such a library is not currently available for a group or region under study, the presence of a barcode makes it possible to reliably identify the sample in the future. Thus, linking karyotypes with DNA barcodes resolves the problem of reliable species identification.

Additionally, combination of DNA barcodes and karyotypes represents a powerful tool for detection, delimitation and description of unrecognized species (Lukhtanov et al. 2015; Vishnevskaya et al. 2016, 2018). Therefore, linking karyotypes with DNA barcodes, potentially resolves the problem of unclear species identity in chromosomal studies.

The approach based on combination of chromosomal and DNA barcode data has been already used in different studies on butterflies (Lukhtanov et al. 2014, 2015; Lukhtanov and Dantchenko 2017), fish (Marques et al. 2013), lizards (de Matos et al. 2016), mammals (Tavares et al. 2015) and mussels (Garcia-Souto et al. 2017). However, its principles have not been explicitly formulated.

In this paper, we demonstrate the algorithm, features and capabilities of the proposed approach with the butterflies of the Neotropical fauna.

The Neotropics is one of the most species-rich regions of the world, and the nymphalids are the most speciose butterfly family (Van Nieukerken et al. 2011). Therefore, it is not surprising that the neotropical fauna of Nymphalidae is very rich in species (site (https://www.butterfliesofamerica.com/L/Nymphalidae.htm).

Chromosomal studies represent only a small part of the Neotropical nymphalid diversity (de Lesse 1967, 1970a, b; de Lesse and Brown 1971; Wesley and Emmel 1975; Suomalainen and Brown 1984; Brown et al. 1992, 2004, 2007a, b; McClure et al. 2017; Lukhtanov 2019a). However, they demonstrate an extremely high level of the interspecific karyotype variation and a potential for solving taxonomic problems within the South American nymphalid species. This potential is practically not used (but see: Suomalainen and Brown 1984; Constantino and Salazar 2010; McClure et al. 2017) in opposite to the numerous chromosomally based taxonomic studies in palearctic butterflies (Lorković 1958; de Lesse 1960; Lukhtanov et al. 2011, 2015; Talavera et al. 2013).

In this study we suggest a plan for further analysis of the Neotropical Nymphalidae based on a parallel analysis of chromosomal and molecular markers.

Using this approach, we confirm the previously published data on the karyotypes of Godyris dircenna (n = 36), Hypothyris euclea (n = 14), Napeogenes sylphis (n = 14) and Oleria gunilla (n = 11) (Brown et al. 2004).

Haploid chromosome number n=30 is found by us in Pseudoscada timna, whereas n = 31 was reported for this taxon by Brown et al. (2004).

We provide the first data on karyotypes of Vila emilia and demonstrate a high interspecific chromosome number variation in this genus (previously n = 15 was reported for an unidentified Vila species from western Brazil; Brown et al. 2007a).

We show chromosome number n = 14 for Hyposcada kena confirming high level of interspecific variation in the genus Hyposcada (from n = 12 to n = 19) (Brown et al. 2004).

Different chromosome numbers were previously reported for Godyris zavaleta by Brown et al. (2004): n = 46 (on the page 220–221), n = 35–45 (p. 222), n = 36–46 (p. 224), n = 40 (p. 229). However, the credibility and the reason for this variation were not discussed. We provide n = 33 for this species and point out the need for further study of this taxon.

Even more interesting data were obtained regarding the species Oleria didymaea (Hewitson, 1876) and O. onega. We found n = 22 in the taxon identified by us as Oleria didymaea ramona (Haensch, 1909), whereas n= 15 was reported for taxon identified as Oleria alexina didymaea (Brown et al. 2004) raising the question of further study of the complex Oleria didymaea – alexina.

Based on chromosome numbers, we hypothesize that Oleria onega is a complex of at least three species with different chromosome numbers: n = 15 (our data), n = 22 and n = 30 (Brown et al. 2004). A similar conclusion can be made on the basis of molecular data that show the presence of at least four clusters of DNA barcodes in this complex (Figs 20, 21). The status of the detected chromosomal races and mitochondrial clusters could be theoretically resolved based on analysis of: (1) congruence of chromosomal and molecular characters in different sets of individuals, or (2) pattern imitating (vs not imitating) linkage of chromosomal and mitochondrial markers that are known to be unlinked (Lukhtanov at al 2015; Vishnevskaya et al. 2016, 2018; Lukhtanov 2019b). Unfortunately, the previously karyotyped samples (Brown et al. 2004) were not studied with respect to molecular markers, and vice versa, the vouchers for molecular studies were not karyotyped (de-Silva et al. 2010).

The incomplete connection between the chromosomal and molecular data (karyotypes and DNA barcodes were obtained for different sets of individuals) complicates the taxonomic interpretation of the discovered clusters. Nevertheless, we predict that in future linking karyotypes with DNA barcodes will result in a significant rearrangement of taxonomy of the genus Oleria.

We thank Fedor Konstantinov (St. Petersburg State University) for help in butterfly collecting and research. The study was supported by the projects RFBR 18-04-00263-a, RFBR 17-04-00828-a and state research project AAAA-A19-119020790106-0.