A review of the genus Mecistorhabdia (Lepidoptera: Erebidae: Arctiinae: Syntomini) with a description of a new species from the Central African Republic

The genus Mecistorhabdia Kiriakoff, 1953 is reviewed and one new species M. sulaki sp. n. from the Central African Republic and Cameroon (Afrotropical region) described. The holotype female of the type species M. haematoessa (Holland, 1893) is matched with a conspecifi c male using a molecular technique. M. burgessi Kiriakoff, 1957 is removed from the synonymy as it is considered to be a valid species. Habitus and male genitalia of all three species are illustrated along with the female genitalia of M. haematoessa. A key for identifi cation of species based on the structure of male genitalia and a map of their distribution are provided. The results of a study of the genetic diversity of the three species of Mecistorhabdia, based on cytochrome c oxidase subunit I gene, are presented. Relationships between species are briefl y discussed. ZooBank Article Registration: http://zoobank.org/urn:lsid:zoobank.org:pub:350BED21-ED91-4276-90E9-8AF1D2B2CCBC


INTRODUCTION
For a long time the thyretines were thought to be a separate monophyletic lineage and treated as a separate family, the Thyretidae (Kiriakoff, 1949). Following morphological and molecular studies of the entire Noctuoidea it was downgraded to tribal rank by Jacobson & Weller (2002) and fi nally treated as a subtribe Thyretina of Syntomini (Zahiri et al., 2011). Recently a comprehensive molecular study revealed that the Thyretina is not monophyletic (Przybyłowicz et al., 2019). As a result, it was formally synonymized with Syntomini. The group was revised by Przybyłowicz (2009) Until recently the genus Mecistorhabdia included only a single valid species, M. haematoessa, which was originally placed by Holland (1893) in the genus Metarctia Walker, 1855. This species was described based on a single female without description of its reproductive organs. Kiriakoff (1953) revised the status of this taxon and proposed a new sary with the old female holotype of M. haematoessa (specimen No. 18.4.6), which is more than 100 years older than the other material. If one of these reactions was successful, an effort was made to obtain a barcode compliant record (> 650 bp) by amplifying shorter regions of COI. PCR were performed in 25 μl reaction volume using 2 × Phanta Max Master Mix (Vazyme Biotech Co., Ltd, China), 10 mM dNTP Mix, 10 μM of each primer and 4 μl of the genomic DNA. The PCR profi le consisted of an initial denaturation step at 94°C for 3 min, followed by 36 cycles at 94°C for 1 min, 48°C for 1 min, and 72°C for 2 min and a fi nal extension step of 7 min at 72°C. Successful PCR products were purifi ed using the GeneMATRIX PCR/DNA Clean-Up Purifi cation kit (EURx, Poland, following the standard protocol) and sequenced in both directions. All sequences were deposited in GenBank under the accession numbers MN441758 and MN441760-MN441768 (Table 1). The sequence of Mecistorhabdia burgessi from the GenBank database was also included in the analysis (Accession Number MK158553). Arctia caja (Linnaeus, 1758) and Lithosia quadra (Linnaeus, 1758) in the tribe Arctiini and Lithosiini, respectively (both in subfamily Arctiinae), were chosen as outgroup for analysis and the sequences were also taken from the GenBank database (Accession Numbers: HM873709 and KJ547685, respectively).
The obtained nucleotide sequences were edited and assembled into contigs in CodonCode Aligner (CodonCode Corporation). Sequences were also checked for protein-coding frame shifts to detect pseudogenes using MEGA X (Kumar et al., 2018) and compared with sequences from GenBank using a BLAST search. Best-fi t models for Bayesian (BI) and maximum likelihood (ML) analyses were calculated in MrModeltest (Nylander, 2004) using the Akaike information criterion (AIC). ML analyses were carried out in IQTree (Nguyen et al., 2015). Bootstrap support was calculated using 1000 replicates. BI analyses were performed in MrBayes (Ronquist & Huelsenbeck, 2003) with four independent runs, each having three heated and one cold chain. Analyses were run for 6 million generations with trees sampled every 1000 generations. The fi rst 25% of each run was discarded as burn-in. Convergence among the runs was assessed using Tracer (Rambaut et al., 2018). All trees were visualized using FigTree with a midpoint rooting option (Rambaut & Drummond, 2012).
Pairwise sequence divergences were calculated separately for intraspecifi c as well as interspecifi c comparisons with MEGA X using Kimura's two parameter (K2P) distance model. Species delimitation method Automatic Barcode Gap Discovery, ABGD (Puillandre et al., 2012), is an automatic procedure that considers the sequences as hypothetical species based on the barcoding gap. The model fi rst calculates a range of prior intraspecifi c diver-and provide high-quality illustrations of the reproductive organs enabling an easier separation of species.

MATERIALS AND METHODS
In this study we follow the systematic account of Przybyłowicz et al. (2019) for the higher taxonomic levels.
Specimens of the genus Mecistorhabdia were collected in Uganda (July 2011 by Harald Sulak), Central African Republic (September 2000 by local collector) and Republic of the Congo (January to March 1997 by Viktor Siniaev and Vladimir Murzin) and stored in the MWM collection. Additional material from the RMCA and ISEAPAS collections was used for comparison.
The abdomens were removed and placed in cold 5% KOH solution overnight after which the abdominal cuticle was opened by a right lateral cut to the genitalia, which were removed and stained in Evans Blue. The aedeagus was pulled out caudally. Images of genitalia were taken using a Leica DFC450 camera.
Forewing length was measured from the base to the apex of the wings, parallel to the wing costa. Forewing width was measured perpendicularly from the costa to the tornus.
Morphological terms for genitalia are those used by Klots (1970).

Molecular analysis
Total genomic DNA was extracted from leg muscle tissue of dried specimens or specimens preserved in pure ethanol using the GeneMATRIX Bio-Trace DNA Purifi cation kit (EURx, Poland, following the standard protocol for tissue). The cytochrome c oxidase subunit I gene (COI) was amplifi ed by polymerase chain reaction (PCR) using the primers LCO1490 (5'-GGT CAA CAA ATC ATA AAG ATA TTG G-3') and HCO2198 (5'-TAA ACT TCA GGG TGA CCA AAA AAT CA-3') (Folmer et al., 1994). In case of failure in sequencing with these primers, additional PCR reactions were carried out to recover amplicons using pri mers ZBJ-ArtF1c (5'-AGA TAT TGG AAC WTT ATA TTT TAT TTT TGG-3') and ZBJ-ArtR2c (5'-WAC TAA TCA ATT WCC AAA TCC TCC-3') (Zeale et al., 2011). This combination was neces- gence to obtain the maximum limit of the intraspecifi c diversity and then partitions the data on the basis of the fi rst signifi cant gap beyond this limit, which is a potential limit between intra-and interspecifi c diversity. The data set was submitted to the ABGD online website (https://wwwabi.snv.jussieu.fr/public/abgd/) and analysed using the following settings: P (prior intraspecifi c divergence) set from 0.001 (Pmin) to 0.08 (Pmax) and Steps set to 10; X (minimum relative gap width) set to 1; Nb bins (for distance distribution) set to 20 and Kimura (K80) model and TS/TV set to 2.0 were selected.

Diagnosis.
Habitus is quite similar to M. burgessi but distinctly differs from it in the male having a larger wingspan -30-35 mm and 28-30 mm respectively. The structure of male genitalia should be used for safe identifi cation, so in contrast to M. burgessi costal process of valva C-curved, the length of saccular process of valva is longer or equal in length to costal process. Vesica bears minute but distinct cornutus. It is also noteworthy that the coloura- tion of M. haematoessa is much less orange than that of M. burgessi.
Male genitalia (Figs 7,9). Uncus coracoid shaped, extended in medial part; dorsally and laterally covered by elongate setae. Tegumen ribbon-shaped. Valvae symmetrical, divided in to two elongated and slightly curved processes: saccular process, which are the same length or longer than the costal process, covered by setae abdominally; costal process narrows towards top, dorsal edge and tip covered by elongate setae. At the base valvae bear basal appendage, which is almost half the size of costal process, curved dorsally and tip well covered with sclerotized elongate setae. Juxta well developed, almost X-shaped. Saccus rounded at top. Aedeagus tubular, narrows in basal part. Vesica divided in to two lobes -elongate dorsal with fi eld of small sclerotizations on the base and an abdominal one with one or two small cornuti.
Female genitalia (Fig. 8). Papillae anales rounded and covered with setae; apophyses posteriors well developed; VIII segment strongly connected with vaginal plate, which is well developed and forms two bumps on the side of ostium; apophyses anteriores very small, almost three times shorter than apophyses posterior; corpus bursae oviform shape; ductus bursae, ductus seminalis and corpus bursae not sclerotized, without signum. Type locality. Gabon, Valley of the Ogooué River, Lambarene.
Distribution (Fig. 12). Cameroon, Equatorial Guinea, Gabon, Republic of the Congo, Democratic Republic of the Congo, Angola.
type specimen are destroyed so it is not possible to complete the analysis. Detection of a new taxon with unknown females sympatrically distributed with haematoessa raised the question about the real identity of the sexes ascribed until now to already known taxa. Due to the overall similarity of all three Mecistorhabdia species, scarcity of females and signifi cant difference in size between males and females it was not obvious if the female type of haematoessa should be associated with the male morphotypes proposed by Kiriakoff (1953) or with those of the newly discovered new taxon. The only reliable and convincing answer for this question was the analysis and comparison of molecular characters of both male morphotypes and the female holotype of M. haematoessa.
Mecistorhabdia burgessi Kiriakoff, 1957, sp. rev. (Figs 3,4,10,12,13,15) Mecistorhabdia burgessi Kiriakoff, 1957: 96. Diagnosis. This species differs from others in external view of adult, which has a more intensive orange and darkgrey colouration on forewing and pale orange hindwing. Specimens of this species are frequently mixed together with M. haematoessa in collections of Lepidoptera. It differs from the latter taxon by smaller wingspan of male -28-30 mm compared to 30-35 mm in M. haematoessa. The structure of male genitalia should be used for safe identifi cation, so in contrast to M. haematoessa saccular process of valvae is distinctly shorter than costal process and vesica has no distinct cornutus (only tiny sclerotizations). It should also be noted that the costal process of valvae is almost straight in comparison with those of M. haematoessa. We removed the taxon burgessi from the synonymy of M. haematoessa in view of the distinct differences in structures of male genitalia.
Male (Figs 3, 4). Antenna bipectinate, bright yellow. Head and patagia dark brown, tegula, thorax covered by dirty orange and brown elongate scales. Abdomen dorsally dirty orange, ventrally pale brown. Wingspan 28-30 mm. Forewing dirty orange dark-grey, with intense orange costal margin and blurs of orange scales on postbasal, discal, postdiscal areas. Marginal edge of forewing mainly dirty orange. Hindwing pale-orange with grey scales on marginal edge.
Female. Unknown. Male genitalia (Fig. 10). Uncus of coracoid shape, extended in medial part; dorsally and laterally covered by elongate scales. Tegumen ribbon-shaped. Valvae symmetrical, divided into two elongated and slightly curved processes: saccular process, which is distinctly shorter than costal process, slightly curved, covered by setae abdominally; costal process narrows towards top, dorsal edge and tip covered by elongate setae. At the base valvae bear a basal appendage, which is almost half the size of costal process, curved dorsally and tip well covered with sclerotized elongate setae. Juxta well developed, almost Xshaped. Saccus rounded at top. Aedeagus tubular, narrows towards top. Vesica divided in to two lobes -elongated abdomen with fi eld of tiny scrobinations at the base and laterally, which is much shorter and without cornuti. Distribution (Fig. 12). Democratic Republic of the Congo, Uganda, Kenya, Tanzania.
Mecistorhabdia sulaki sp. n. (Figs 5,6,11,12,13,15) ZooBank taxon LSID: 4DA10EF3-74D9-42B6-BF5A-80B506BD1D0A Diagnosis. This species differs from others in its smaller size -wingspan 27 mm, colouration of the forewing, which is distinctly brighter. The structure of male genitalia should be used for safe identifi cation, because in contrast to M. haematoessa and M. burgessi the valvae have no basal appendage, basal part of the aedeagus is distinctly narrow, vesica has one area of tiny sclerotizations and two fl attened round cornuti.
Male (Figs 5,6). Antenna bipectinate, bright yellow. Head, patagia and thorax covered by dirty orange elongate scales. Abdomen dorsally orange, with elongate pale-yellow scales on tip. Wingspan 23-27 mm. Forewing bright orange yellow with deep orange costal margin and blurs of grey scales on postbasal, discal and postdiscal areas. Marginal edge of forewing mainly yellow. Hindwing paleyellow.
Female. Unknown. Male genitalia (Fig. 11). Uncus apparently extends to middle part and bears a small sharpened appendage on the tip; dorsally and laterally covered by elongate scales. Tegumen ribbon-shaped. Valvae symmetrical, divided in two elongated and slightly curved processes: costal process long, slightly extended, rounded and bears setae on distal part; saccular process half the length of costal process, with slightly rounded tip, ventrally covered by elongate setae. Juxta well developed and elongated. Saccus with rounded top. Aedeagus tubular, narrows and slightly curved basal part. Vesica divided in two lobes -elongate dorsal with area of minute sclerotizations and abdominal with two fl at rounded cornutti.   There were no indels and stop codons in the 613 bp of COI. The MrModeltest indicated that the SYM + G (gamma distribution shape parameter G = 0.7231) as the best model of DNA substitution. The phylogenetic trees calculated using different methods had identical topologies, so only the Bayesian tree is shown (Fig. 13). The Bayesian poste-rior probability values for the nodes were generally higher than the bootstrap values. In the genus Mecistorhabdia species were clearly separated and formed a monophyletic group. The phylogenetic analysis divided the genus into three lineages in which Mecistorhabdia sulaki was a sister species to M. haematoessa + M. burgessi.
The K2P distances of the COI gene (barcoding region) between different Mecistorhabdia specimens ranged from 0 to 6% (Table 2). The maximum distance between the specimens of M. haematoessa was 0.3%. The oldest representative of M. haematoessa (the female holotype, specimen No. 18.4.6) was similar to the rest of the specimens of this species. The genetic divergence values between specimens of M. burgessi ranged from 0 to 0.99%. Average distance between M. burgessi and M. sulaki included in the analysis was 6%.
The ABGD method revealed gap between intra-and interspecifi c distance variation in the barcoding (Fig. 14). Hence, ABGD tree also clustered the COI sequences into three Mecistorhabdia groups (putative species) (Fig. 15), which are congruent with the phylogenetic tree (Fig. 13).

DISCUSSION
Mecistorhabdia is one of the smallest genera of Syntomini. However, the thorough investigation based on both morphological and molecular characters revealed hidden diversity within until now a monotypic genus. The previous study of the single types of M. haematoessa (female) and M. burgessi (male) led to the incorrect assumption that they re present a single taxon (Przybyłowicz & Kühne, 2008). This was due to the uniformity of both taxa in terms of pattern, colouration and male genitalia.
It is likely that careful reexamination of material (both existing in collections and new) of several other monotypic genera of Arctiinae from a wide area, including a morphogenetic analysis, would reveal the existence of further undescribed taxa. Such discoveries are already reported for Cacoethes Hübner, 1816(de Freina, 2014 and Neophemula (Ochse, 2017. Our study once again proves that old museum material can be used in molecular procedures to solve long-lasting taxonomic problems such as matching opposite sexes, detecting the molecular characters of unique types or ascribing incomplete types to particular taxa (Strutzenberger et al., 2012;Prosser et al., 2016). This last aspect is especially important when the genitalia are damaged due to mechanical (crushing, tearing) or biological (fungi, pests) factors. In such cases, COI sequences can be extracted from even small fragments of highly degraded DNA and provide crucial information on the taxonomic identity of the specimen analyzed.
Based on the above taxonomic treatment further studies can now be carried out on the genus Mecistorhabdia. In terms of morphology future studies should concentrate on collecting the so far undiscovered females of M. burgessi and M. sulaki. This fresh material will help in defi ning the distributions of the sympatric M. haematoessa and M. sulaki. The question of the intraspecifi c genetic variation of each species should be studied using freshly collected specimens. M. burgessi seems to be the typical representative of an East African fauna with the western border of its range in Ruwenzori Mountains. However, this hypothesis should be confi rmed by extensive collecting in the eastern part of the Congo basin.