Revisiting the taxonomy and molecular systematics of Sesamia stemborers (Lepidoptera: Noctuidae: Apameini: Sesamiina): updated classification and comparative evaluation of species delimitation methods

Noémie M. C. Hévin; Gael J. Kergoat; Alberto Zilli; Claire Capdevielle-Dulac; Boaz K. Musyoka; Michel Sezonlin; Desmond Conlong; Johnnie Van Den Berg; Rose Ndemah; Philippe Le Gall; Domingos Cugala; Casper Nyamukondiwa; Beatrice Pallangyo; Mohamedi Njaku; Muluken Goftishu; Yoseph Assefa; Onésime Mubenga Kandonda; Grégoire Bani; Richard Molo; Gilson Chipabika; George Ong’amo; Anne-Laure Clamens; Jérôme Barbut; Bruno Le Ru

doi:10.3897/asp.82.e113140

Abstract

In this study, we reassess the phylogenetic relationships of the genus Sesamia Guenée, 1852 and examine in more detail the members of the nonagrioides species group, for which three distinct species complexes are identified. The calamistis subgroup comprises eight species, of which four new species are described: Sesamia kabirara Le Ru sp. nov., Sesamia kalale Le Ru sp. nov., Sesamia mapalense Le Ru sp. nov. and Sesamia teke Le Ru sp. nov. The incerta subgroup consists of 11 species, of which four new species are described: Sesamia kamba Le Ru sp. nov., Sesamia lalaci Le Ru sp. nov., Sesamia lusese Le Ru sp. nov. and Sesamia msowero Le Ru sp. nov. The nonagrioides subgroup comprises ten species of which two new species are described: Sesamia libode Le Ru sp. nov. and Sesamia satauensis Le Ru sp. nov. Phylogenetic and molecular species delimitation analyses of a multi-marker molecular dataset allow us to investigate and clarify the status of Sesamia species and species complexes. Our results yield a well-supported phylogenetic hypothesis for the genus, which supports the monophyletic nature of all but one species subgroup. The results of 16 distinct molecular species delimitation analyses show some levels of incongruence and, overall, a tendency towards over-splitting. We also present an updated list of species for the genus Sesamia and provide morphological keys based on male and female genitalia to determine the species group of any Sesamia species and to identify all species belonging to the nonagrioides species group.

Key Words

integrative taxonomy, molecular phylogenetics, molecular species delimitation, pest species, species complex, systematics

1. Introduction

The genus Sesamia is the second most diverse genus – after Acrapex Hampson, 1854 – of Sesamiina stemborers (Lepidoptera: Noctuidae: Noctuinae: Apameini), a subtribe of ca. 300 species mainly distributed in the Afrotropics. Sesamia currently consist of more than 60 valid species which are mainly distributed in the Afrotropical region (Moyal 2006; Moyal et al. 2011; Kergoat et al. 2015, 2018; Le Ru et al. 2020, 2022). Like other sesamiine stemborers, Sesamia species tend to prefer open habitats and humid environments (Kergoat et al. 2018); they are highly specialised phytophagous insects, whose larvae develop mainly in the stems of C₄ grasses and sedges (Poales). Several species of Sesamia are infamous for being major pests of various cereal crops (e.g., maize, rice, sorghum, and sugarcane), in particular S. calamistis Hampson, 1910, S. cretica Lederer, 1857, S. grisescens Warren, 1911, S. inferens (Walker, 1856) and S. nonagrioides (Lefèbvre, 1827) (Tams and Bowden 1953; Young and Kuniata 1992; Kuniata 1994, 1998; Kuniata and Sweet 1994; Kfir et al. 2002; Singh and Kular 2015; Goftishu et al. 2016; Dey et al. 2021; Roopika et al. 2022). Until the 1950s, the great superficial resemblance of most species of the genus resulted in many misidentifications and confusions. It was only through the study of the male genitalia by Janse (1937–1939) and of both male and female genitalia by Tams and Bowden (1953) that informative morphological characters were identified, allowing improved species identifications and a better circumscription of sesamiine genera. Tams and Bowden (1953) also showed that Sesamia species may be assigned to at least three distinct morphological groups on the basis of the male genitalia: (i) the coniota species group (not named at the time), characterised by the partially fused sacculus and cucullus and a weakly spinose lamina ventralis of phallus, (ii) the cretica species group, characterised by the fused sacculus and cucullus, a medial or dorsal process to the juxta and the strongly spinose lamina ventralis, and (iii) the nonagrioides species group, characterised by the separated cucullus and sacculus, simple juxta and membranous lamina ventralis.

The three Sesamia species groups have only begun to be revised in depth in recent years, thanks to the combined use of morphological and molecular data within an ‘integrative taxonomy’ framework (sensu Dayrat 2005; Will et al. 2005). Initially, two species complexes belonging to the nonagrioides species group were studied, leading to the description of nine new species (Moyal et al. 2011; Kergoat et al. 2015). In another study, Le Ru et al. (2020) formally defined the coniota species group and described three new species. More recently, the cretica species group was revised by Le Ru et al. (2022), resulting in the description of 23 new species. Molecular phylogenetic studies on the cretica species group also highlighted its polyphyletic nature, as several distinct lineages could be evidenced (Le Ru et al. 2022). In the process of revising Sesamia species groups, molecular species delimitation (SD) analyses were carried out in two studies (Kergoat et al. 2015; Le Ru et al. 2022). However, in doing so, both studies essentially focused on the species groups or subgroups of interest; in addition, only a single SD method was used in Kergoat et al. (2015), whereas up to 17 distinct approaches were used in Le Ru et al. (2022). Given that molecular SD analyses are known to be sensitive to breadth of the samples (see Ahrens et al. 2016), it is of interest to investigate Sesamia species boundaries with a more extensive sampling. This would also disclose whether there are significant discrepancies between the results of analyses carried out on a broad sample of specimens and those focusing only on a single species group (Le Ru et al. 2022) or subgroup (Kergoat et al. 2015).

In this study, we present a morphological revision of the nonagrioides species group, with formal descriptions of new species and supplemental descriptions of previously described taxa. We also provide morphological keys based on male and female genitalia that distinguish the three major Sesamia species groups and facilitate the identification of all species belonging to the nonagrioides species group. A review of all the species currently assigned to Sesamia is also performed. This morphological revision is carried out in conjunction with a molecular study, where the largest molecular dataset ever assembled for this genus (more than 600 specimens sequenced for up to four mitochondrial and two nuclear gene fragments) is analysed. Molecular phylogenetic analyses are used to investigate the phylogenetic placement of Sesamia species. In addition to the morphological study, several molecular SD analyses are used to investigate and clarify the status of Sesamia species after an integrative taxonomy approach. The results of molecular SD analyses are also discussed and compared with the results of previous molecular works on the genus.

2. Methods

2.1. Sampling and morphological study

Several thousand specimens belonging to the nonagrioides species group were obtained from the sampling of larvae from grasses and sedges (Poales) visually determined to have herbivore damage in sub-Saharan Africa (see Table S1), and in the western Palearctic (France, Italy, Iran and Turkey; see Kergoat et al. 2015 for details). Larvae were reared on artificial diets (following Onyango and Ochieng’Odero 1994) until pupation and emergence of adults (e.g., Le Ru et al. 2006a, 2006b). For non-crop Afrotropical species, plant specimens were identified by Simon Mathenge (Botany Department, University of Nairobi, Kenya). Additional adult specimens were collected with light traps set up in Cameroon, Democratic Republic of Congo, Ethiopia, Kenya, Republic of Congo, Republic of South Africa and Zambia. All sampling was carried out in accordance with local insect collection permits or local regulations. Lastly, hundreds of adult specimens (including holotypes and paratypes) were studied in the repository institutions listed below (see section 2.5). Specimens of other Sesamia species (already analysed in previous studies on the genus) were also taken into account when constructing the general key to distinguish the three species groups; this was also the case for specimens originally assigned to Sesamia but which do not belong to this genus (see the corresponding section in the discussion). Genitalia were dissected after immersion of the last abdominal segments in a boiling 10% potassium hydroxide (KOH) bath for a few minutes, then cleaned, immersed in absolute alcohol for a few minutes and mounted on slides in Euparal (after separating the phallus from the rest of the male apparatus). The types of the species new to science were deposited as indicated in the individual descriptions given in the taxonomic section.

2.2. Molecular dataset

For this study, a comprehensive molecular dataset was assembled comprising 603 Sesamia specimens (57 from the coniota species group, 131 from the cretica species group and 415 from the nonagrioides species group) and 12 outgroups (one Apameina and 11 Sesamiina; including two species currently attributed to the genus Sesamia but which belong to a yet undescribed genus, see Table S2). This dataset consists of six gene fragments. We used the following four mitochondrial gene fragments: 658 bp region of the cytochrome c oxidase subunit 1 (COI), 992 bp of the cytochrome b (cytb), 370 bp of the small ribosomal RNA (rrnS), and 533 bp of the large ribosomal RNA (rrnL). Two nuclear gene fragments were also targeted: 1,240 bp of the elongation factor-1a (EF1a), and 860 bp of the 28S ribosomal DNA (28S). Most specimens (580 specimens) were sampled and previously sequenced by our research group (Toussaint et al. 2012; Kergoat et al. 2015, 2018; Le Ru et al. 2020, 2022). For the purpose of this study additional specimens were also processed, including a first batch of eight very old specimens (collected between 1921 and 1964) and a second batch of 38 specimens collected between 2008 and 2014. From these 46 specimens, we were able to generate sequences for 39 specimens (including 30 COI, 22 cytb, 24 rrnS, 30 rrnL, 21 EF1a, and 23 28S gene fragments; see Table S2 for details), including two of the oldest (please see below for details). Lastly, GenBank data was also used to include 24 S. inferens specimens (24 COI and one cytb gene fragments) from China, India, Indonesia and Thailand and one S. grisescens specimen (rrnS gene fragment) from Papua New Guinea (see Table S2).

For the 46 additional specimens processed in this study, DNA was extracted from hind legs using DNAeasy tissue kits (Qiagen, Hilden, Germany). Initially, we attempted standard PCRs (see Kergoat et al. 2012 for details), which proved successful for most of the recently collected specimens (37 out of 38), but failed for all older specimens. In an attempt to get data for the oldest specimens, we used high-throughput sequencing (amplicon sequencing) to target COI, rrnL and rrnS gene fragments. Amplicon libraries were constructed for these three genes following Galan et al. (2017), with slight modifications as described in Hévin et al. (2023); for the COI, two overlapping fragments were targeted following Shokralla et al. (2015). The final library was paired-end sequenced on an Illumina MiSeq flowcell using a MiSeq Reagent Kit v2 (500 cycles) at the AGAP laboratory (Montpellier, France). Using this approach, we were able to obtain COI Illumina reads for a specimen of S. kamba sp. nov. collected in 1955 and for a specimen of S. madagascariensis Saalmüller, 1891 collected in 1954. The corresponding reads were processed using the FROGS pipeline (http://frogs.toulouse.inra.fr; Escudié et al. 2018) on the GenoToul Galaxy server using demultiplexing, pre-processing, clustering and chimera removal tools. Remaining contaminants were further detected using the BLAST tool (available at: https://blast.ncbi.nlm.nih.gov/Blast.cgi) and removed manually. Newly generated sequences were deposited on GenBank (see Table S2 for all GenBank accession numbers).

All mitochondrial and nuclear sequences were aligned using MAFFT v7 (Katoh and Standley 2013) with default option settings and a gap opening penalty of 5.0, and further visually checked using Mesquite. For all protein-coding genes (cytb, COI and EF1a), coding frames and stop codons were also checked with Mesquite to detect potential pseudogenes. The final concatenated dataset consisted of a matrix of 615 specimens and 4,653 bp.

2.3. Phylogenetic analyses

Phylogenetic analyses were carried out under maximum likelihood (ML) with IQ-TREE v.2.1.3 (Minh et al. 2020). The concatenated dataset was divided into 12 partitions, with three partitions (one per codon position) for each coding gene fragment (cytb, COI and EF1a) and one partition for each non-coding gene fragment (rrnL, rrnS and 28S). To infer individual gene trees, additional analyses were also conducted on each gene fragment, with three partitions implemented for the coding genes. Best-fit substitution models and partition schemes were selected with ModelFinder (part of the IQ-TREE software, Kalyaanamoorthy et al. 2017) using the Bayesian Information Criterion (Table S3). Maximum likelihood trees were obtained using heuristic searches implementing 500 random-addition replicates with the following settings: random-starting tree, hill-climbing nearest neighbour interchange (NNI) search (-allnni option), a default perturbation strength of 0.5 (-pers 0.5 option), partition-resampling strategy (-sampling GENE option), best partition scheme allowing the merging of partitions (-m MFP+MERGE option). Clade support for all analyses was assessed using 1,000 replicates for both SH-like approximate likelihood ratio tests (SH-aLRT; Guindon et al. 2010) and ultrafast bootstraps (uBV; Minh et al. 2013). Highly supported nodes were identified by SH-aLRT values ≥ 80% and uBV ≥ 95% as recommended by the authors of IQTREE (http://www.iqtree.org/doc/Frequently-Asked-Questions). All tree files were visualized using FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree).

2.4. Molecular species delimitation analyses

Sixteen distinct molecular species delimitation (SD) analyses were used in this study to better assess the consistency and repeatability of the inferred species delineations.

First, we relied on two tree-based methods: the Poisson-tree-process (PTP) approach of Zhang et al. (2013) and the General Mixed Yule Coalescent (GMYC) model of Pons et al. (2006). These methods were applied on the best-scoring ML tree resulting from the analyses of the concatenated dataset, and also on the cytb and COI gene trees. Analyses were performed on a dedicated web server (https://species.h-its.org) with default parameters. For the GMYC model, we relied on the default single threshold approach (results referred to as ‘GMYCs’) of Pons et al. (2006) as well as the more parameter-rich approach (results referred to as ‘GMYCm’) of Monaghan et al. (2009), which implements multiple thresholds to account for potential heterogeneity of evolutionary rates among lineages. The ultrametric trees required for the GMYC methods were created with treePL (Smith and O’Meara 2012) using default settings.

Second, two distance-based methods were used: the Automatic Barcode Gap Discovery (ABGD) model of Puillandre et al. (2012) and the Assemble Species by Automatic Partitioning (ASAP) model of Puillandre et al. (2021). As these methods are known to be sensitive to missing data, they were only used for the cytb and COI gene fragments for which we had better coverage. ABGD analyses were performed using default settings, a standard Kimura 2-parameter model (K80) and recursive partitioning strategies on a dedicated webserver (https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html). ASAP analyses, which differ from ABGD by including a specific scoring system to identify the best-fitting set of putative species, were carried out on a dedicated web server (https://bioinfo.mnhn.fr/abi/public/asap/#) using default settings and a K80 model.

Third, a multi-locus coalescent-based SD approach implemented in the program tr2 (Fujisawa et al. 2016) was used. This method uses both gene trees and a guide tree as inputs, and relies on a Bayesian model comparison framework with rooted triplets. All six gene trees were used and the topology resulting from the analyses of the concatenated dataset was used as a guide tree.

Finally, LIMES (Ducasse et al. 2020) was used to compare the partitions of all SD analyses using a SPART file (Miralles et al. 2022) as input. All the results of the SD analyses and the morphospecies previously defined were ranked according to the mean congruence (mCtax) obtained. Following Ducasse et al. (2020), specimens with missing gene fragments were excluded, resulting in the removal of five species (S. djenoensis Le Ru, 2022, S. grisescens, S. lusese sp. nov., S. madagascariensis and S. viettei Rungs, 1954) for this specific analysis.

Results of the morphological and molecular studies were considered altogether, with molecular analyses being used to either test species status or determine monophyly of species groups and subgroups. To refine our hypotheses, we proceeded in an ‘iterative’ manner (sensu Yeates et al. 2011), carrying out additional analyses over time, which enabled us to identify with greater precision the species complexes requiring in-depth morphological studies.

2.5. Abbreviations

MCSN – Museo Civico di Storia Naturale, Milan, Italy; MNHN – Muséum national d’Histoire naturelle, Paris, France; NHM – Natural History Museum, London, United Kingdom; NMK – National Museum of Kenya, Nairobi, Kenya; SMLF – the Senckenberg Museum, Frankfurt am Main, Germany; TMSA – Ditsong Museum (formerly Transvaal Museum of South Africa), Pretoria, Republic of South Africa.

3. Results

Reappreciation of the already known species of the nonagrioides species group and study of new material allowed us to arrange this assemblage into three subgroups, namely the calamistis, incerta and nonagrioides subgroups, as detailed herein in sections 3.3, 3.4 and 3.5.

3.1. Phylogenetic analyses

The best-scoring tree from the ML analyses of the concatenated dataset is shown in Figure 1. Overall, 40% and 50% of the nodes are supported (SH-aLRT > 80% and uBV > 95%, respectively), with 67% and 78% of the interspecific nodes supported (SH-aLRT > 80% and uBV > 95%, respectively), and 35% and 45% of the intraspecific nodes (SH-aLRT > 80% and uBV > 95%, respectively) (Fig. 1; see also Figshare files for details). The cretica species group is recovered as paraphyletic (with five distinct lineages), while the coniota species group and the nonagrioides species group are recovered as monophyletic with high support (SH-aLRT ≥ 97% and uBV of 100%). Except for the fuscifrontia subgroup, which is found to be paraphyletic, all cretica subgroups (albivena, cretica, salama and wiltshirei) and nonagrioides subgroups (calamistis, incerta and nonagrioides) are recovered as monophyletic with high supports (SH-aLRT values comprised between 85.1 and 100% and uBV ≥ 98%). Within these subgroups, all species are also recovered as monophyletic with high support (SH-aLRT ≥ 80% and uBV ≥ 95%, except S. inferens, for which SH-aLRT = 79.9%). The results of the separate gene tree analyses further used by several SD approaches are available online on Figshare.

Figure 1.

Maximum likelihood topology resulting from the analysis of the concatenated dataset, supplemented by the results of the SD analyses. The tree is shown on the left along with information on clade support for major nodes (S = supported, NS = not supported). The results of molecular analyses are displayed on the right, sorted by mCtax values. The bars are coloured in green, yellow or red, depending on whether a given analysis matches, splits or merges the number of described species, respectively. Analysis type, mCtax, Ctax (between morphology and a given analysis) and the number of species found (Nb species) are summarised in the panel at bottom left.

3.2. Species delimitation analyses

Depending on the methods and loci analysed, between 57 and 125 putative species are inferred by the SD analyses (Fig. 1; see Table S7 for details). Of the 59 sampled Sesamia “species”, 32 are recovered by at least 75% of the SD analyses, while 15 are recovered by less than 50% of the SD analyses. Five SD analyses have mCtax values that are close to the one inferred for the morphospecies (in mCtax order: GMYCs.concatenated, PTP.concatenated, GMYCs.COI, GMYCs.cytb, and ASAP.cytb). Eleven species (S. calamistis, S. coniota Hampson, 1902, S. epunctifera Hampson, 1902, S. incerta (Walker, 1856), S. inferens, S. kafulo Le Ru, 2022, S. kamba sp. nov., S. nonagrioides, S. poephaga Tams and Bowden, 1953, and S. salama Le Ru 2022) are often split by the SD analyses. Of the 16 SD analyses (only ten for S. inferens from India and Indonesia), S. inferens is found only once as a distinct species; S. inferens from China is recovered as a putative species by 81% of the analyses, S. inferens from Thailand by 81%, S. inferens from Indonesia by 70%, and S. inferens from India by 90% (with seven of the ten SD analyses finding different putative species). Sesamia jansei Tams and Bowden, 1953 and S. schoenoplectus Le Ru, 2020 are often split and sometimes some of the specimens are lumped together. Two sister species are sometimes lumped together by the SD analyses: S. libode sp. nov. with S. typhae Le Ru, 2015, and S. capensis Le Ru, 2015 with S. natalensis Le Ru, 2015. Finally, S. msowero sp. nov. is sometimes lumped with S. lalaci sp. nov. or S. pennipuncta Moyal, 2011 by the SD analyses.

3.3. calamistis subgroup

3.3.1. Morphology and taxonomy of calamistis species subgroup

The calamistis subgroup consists of S. calamistis, S. kabirara sp. nov., S. kalale sp. nov., S. mapalense sp. nov., S. madagascariensis, S. monodi Rungs, 1963, S. oriaula Tams and Bowden, 1953 and S. teke sp. nov.. This subgroup is characterised by the following combination of characters: (i) tegumen with large peniculi, vinculum with a large saccus; (ii) valve triangular with sacculus and cucullus separated; costa short, heavily sclerotised ending with a long curved costal spine; sacculus well sclerotized, ovoid at base, ending with a narrow, more or less elongated apical extension, its inner margin extended almost to costal margin, bearing outwardly several rows of short and stout spines; cucullus weakly sclerotized, elongated, clavate at apex with scattered and papillated hairs; (iii) juxta trapezoidal, except for S. madagascariensis; (iv) uncus angled at base, stout tapered to a fine point, tufted with long hairs dorsally; (v) phallus short; lamina ventralis with an elongate carinal crest produced into paired lateral lobes; vesica with a semi-circular or narrow flat cornutus or without cornutus; (vi) ventrolateral plates of female segment A8 elongated, sclerotized, at least twice as long as wide; (vii) ostium bursae small, transverse, strongly sclerotized; (viii) ductus bursae short and broad, generally strongly sclerotized posteriorly. The descriptions of the four new species belonging to the calamistis subgroup are presented below: S. kabirara sp. nov. from Kenya and Uganda, S. kalale sp. nov. from Zambia, S. mapalense sp. nov. from Kenya and S. teke sp. nov. from the Republic of Congo. A supplemental redescription is also provided, where appropriate, for species previously described in this subgroup.

Sesamia calamistis Hampson, 1910

Figures 2A–D; 3A, H; 4A; 5A; 6

= Sesamia mediastriga Bethune-Baker, 1911

Sesamia calamistis – Hampson (1910: 325, Plate CXLIV, fig. 18), Gaede (1934–1935: 97, Tafel 10, Reihe i), Tams and Bowden (1953: 664, figs 18–21, 47–48), Viette (1967: 703, figs 545–546), Poole (1989: 907 [catalogue]). Sesamia mediastriga – Bethune-Baker (1911: 518). Sesamia vuteria sensu Hampson (1910: 324, text figure 142, ♂) nec Stoll, 1790 (pro parte, misidentification).

Type material

Holotype ♂, [REPUBLIC OF SOUTH AFRICA], Cape Colony, Grahamstown, De. 20.03, 98–167, 1924/247, 1905.316, Agrotidae genitalia slide No. 1259, ♀ abdomen stuck on ♂ type, (NHM).

Other material

ANGOLA: one ♂, North Angola, N’dalla Tando, 2700 Ft, X 1908, W. Ansorgei, Type Sesamia mediastriga Bethune-Baker, 1911, (NHM); BOTSWANA: four ♂, Chobe District, Kasane, 17°47′38″S, 25°09′10″E, 929m a.s.l., iii.2016, ex larvae in stems of Pennisetum macrourum Trin., (B. Le Ru leg.) (MNHN); CAMEROON: two ♀, Northwest Region, Babessi, 06°01′55″N, 10°33′07″E, 1203m a.s.l., xi.2013, ex light trap, (B. Le Ru leg.) (MNHN); four ♀, Northwest Region, Ndop, 05°58′40″N, 10°24′25″E, 1182m a.s.l., xii.2013, ex light trap, (B. Le Ru leg.) (MNHN); DEMOCRATIC REPUBLIC OF CONGO: three ♂, four ♀, Province Orientale, Yangambi, Yangole, 00°41′16″N, 24°18′10″E, 375m a.s.l., v.2010, ex larvae in stems of Paspalum virgatum L. (B. Le Ru leg.) (MNHN); ETHIOPIA: three ♂, five ♀, Jimma, Waro Kolobo, 07°36′49″N, 36°49′59″E, 1726m a.s.l., ix.2005, ex larvae in stems of Cenchrus unisetus (Nees) Morrone, (B. Le Ru leg.) (MNHN); KENYA: three ♂, two ♀, Eastern Province, Mtito Andei, 02°40′37″S, 38°11′43″E, 739m a.s.l., iv.2004/x 2003, ex larvae in stems of Sorghum arundinaceum (Desv.) Stapf, (B. Le Ru leg.) (MNHN); one ♀, Eastern Province, Mtito Andei, 02°40′37″S, 38°11′43″E, 739m a.s.l., xii.2003, ex larvae in stems of Cyperus articulatus L., (B. Le Ru leg.) (MNHN); one ♀, Eastern Province, Mtito Andei, 02°40′37″S, 38°11′43″E, 739m a.s.l., iv.2004, ex larvae in stems of Eriochloa meyeriana (Nees) Pilg., (B. Le Ru leg.) (MNHN); one ♂, Eastern Province, Mtito Andei, 02°40′37″S, 38°11′43″E, 739m a.s.l., xii.2003, ex larvae in stems of Schoenoplectus corymbosus (Roth ex Roem. & Schult.) J. Raynal, (B. Le Ru leg.) (MNHN); two ♂, one ♀, Eastern Province, Mtito Andei, 02°40′37″S, 38°11′43″E, 739m a.s.l., iv.2004, ex larvae in stems of Cyperus rotundus L., (B. Le Ru leg.) (MNHN); three ♂, two ♀, Central Province, Mai-Mahiu, 01°02′04″S, 36°26′25″E, 2042m a.s.l., xii.2009, ex larvae in stems of Panicum deustum Brickell & Enslin ex Muhl., (B. Le Ru leg.) (MNHN); MADAGASCAR: one ♀, Tananarive, Anombalahivato, 18°59′24″S, 47°28′03″E, 1395m a.s.l., i.2004, ex larvae in stems of Pennisetum polystachion (L.) Schult., (B. Le Ru leg.) (MNHN); REPUBLIC OF CONGO: four ♂, Kouilou, Kimbakala, 04°21′04″S, 12°08′50″E, 46m a.s.l., Iv.2013, ex larvae in stems of Pennisetum polystachion (L.) Schult., (B. Le Ru leg.) (MNHN); two ♂, Kouilou, Bas Kouilou, 04°28′19″S, 11°42′58″E, 11m a.s.l., iv.2013, ex larvae in stems of Jardinea gabonensis Steud., (B. Le Ru leg.) (MNHN); one ♂, four ♀, Kouilou, Lac Nanga, 04°53′48″S, 11°56′37″E, 14m a.s.l., iv.2013, ex larvae in stems of Echinochloa pyramidalis (Lam.) Hitchc. & Chase, (B. Le Ru leg.) (MNHN); one ♀, Bouenza, Nkaye, 04°10′17″S, 13°19′18″E, 189m a.s.l., iii.2013, ex larvae in stems of Imperata cylindrica (L.) P. Beauv., (B. Le Ru leg.) (MNHN); REPUBLIC OF SOUTH AFRICA: four ♀, Kwazulu Natal, Duku Duku, 28°26′46″S, 32°17′26″E, 27m a.s.l., iii.2016, ex light trap, (B. Le Ru leg.) (MNHN); TANZANIA: one ♀, Iringa province, Sao Hill, 08°23′03″S, 35°10′45″E, 1938m a.s.l., ii.2013, ex larvae in stems of Hyparrhenia dregeana (Nees) Stapf ex Stent, (B. Le Ru leg.) (MNHN); one ♀, Zanzibar, Mtwango, 08°23′03″S, 35°10′45″E, 1938m a.s.l., ii.2013, ex larvae in stems of Hyparrhenia dregeana (Nees) Stapf ex Stent, (B. Le Ru leg.) (MNHN); one ♂, two ♀, Zanzibar, Kikobwa, 05°57′48″S, 39°19′30″E, 85m a.s.l., iii.2012, ex larvae in stems of Andropogon schliebenii Pilg., (B. Le Ru leg.) (MNHN); ZAMBIA: one ♂, two ♀, North-Western Province, Rwanko Azhi, 12°13′13″S, 25°39′04″E, 1413m a.s.l., iii.2012, ex light trap, (B. Le Ru leg.) (MNHN); two ♂, five ♀, Luapula Province, Ngwenya, 12°58′32″S, 28°27′19″E, 1243m a.s.l., iii.2012, ex light trap, (B. Le Ru leg.) (MNHN).

Diagnosis

(See also the identification key of calamistis species subgroup, section 3.3.2.). This species can be distinguished from the other known members of the calamistis subgroup by the combination of the following characters of the male and female genitalia: vinculum u-shaped at the outer margin, slightly w-shaped at the inner margin without indentation, with a large saccus; apical extension of the sacculus shorter than the cucullus; juxta large and trapezoidal, the inferior plate produced into a blunt point, the sides pointed, the superior plate short and wide; phallus short and thick, vesica with a semi-circular large cornutus; ventrolateral plates of female segment A8 almost twice as long as wide; ostium bursae small, transverse, strongly sclerotized, with short and bifid tip on each side; ductus bursae short and broad, strongly albeit asymmetrically sclerotized posteriorly.

Supplemental description

(Fig. 2A–D). The external morphology of adults of both sexes has been described in great detail by Tams and Bowden (1953) and Viette (1967). — Forewing length: male 27–33 mm (x̅ = 29.2 mm, N = 20); female 32–38 mm (x̅ = 34.0 mm, N = 20). — Male genitalia (Fig. 3A, H). Tegumen with large peniculi; vinculum u-shaped at the outer margin, slightly w-shaped at the inner margin, with a large saccus. Valve with sacculus and cucullus separate; costa short and narrow, heavily sclerotized, ending with a short and thin straight spine, with an apical tooth, sometimes bifid; sacculus heavily sclerotized rounded at base, a thick apical extension, slightly curved inwards, shorter than the cucullus, with a crest of several rows of short stout spines, blunt at apex; cucullus long, weakly sclerotized, clavate at apex, with scattered and papillated hairs. Juxta large and trapezoidal, the inferior plate produced into a blunt point, the sides pointed, the superior plate short and wide, with pointed terminations; uncus small and stout, angled at base, tapered near apex to a fine curved point; phallus short and thick, dilated at base; lamina ventralis with an elongate carinal crest, produced into paired lateral lobes; vesica with a semi-circular large cornutus. — Female genitalia (Fig. 4A). Apophyses anteriores with spatulate tips; ventrolateral plates of segment A8 weakly sclerotized, elongated, almost twice as long as wide; ostium bursae small, transverse, strongly sclerotized, the posterior side slightly flared with short and bifid tip on each side; ductus bursae short and broad, strongly albeit asymmetrically sclerotized posteriorly; corpus bursae long, without signa; ovipositor lobes at least 2.5 times longer than wide with dorsal surface bearing numerous short and stout setae, the ventral side of each lobe slightly curved and tooth-shaped; apophyses posteriores more slender than apophyses anteriores. — L5 instar larva (Fig. 5A). Length, 35–40 mm, breadth, 4.0 mm; head smooth, light brown, prothoracic shield pale salmon beige; body with ground colour salmon pink, pinacula and caudal plate pale yellow-brown. Young larvae are very similar in appearance to mature ones.

Figure 2.

Adults of the calamistis subgroup (S. calamistis, S. kabirara sp. nov., S. kalale sp. nov., S. madagascariensis, S. mapalense sp. nov., S. monodi, S. oriaula and S. teke sp. nov.). S. calamistis: A male upper side; B male underside; C female upper side; D female underside; S. kabirara sp. nov.: E male upper side; F male underside, G female upper side; H female underside; S. kalale sp. nov.: I male upper side; J male underside; S. madagascariensis: K male upper side; L male underside, M female upper side; N female underside; S. mapalense sp. nov.: O male upper side; P male underside, Q female upper side; R female underside; S. monodi: S male upper side; T male underside; U original label type from MNHN; S. oriaula: V male upper side; W male underside; X female upper side; Y female underside; S. teke sp. nov.: Z female upper side; AA female underside. Scale bar = 10 mm.

Figure 3.

Male genitalia of members of the calamistis subgroup, with the vesica inside the phallus. S. calamistis: A apparatus; H phallus; S. kabirara sp. nov.: B apparatus; I phallus; S. kalale sp. nov.: C apparatus; J phallus; S. madagascariensis: D apparatus; K phallus; S. mapalense: E apparatus; L phallus; S. monodi: F apparatus; M phallus (poorly prepared, damaged); S. oriaula: G apparatus; N phallus. Scale bar = 1 mm for apparatus and 0.5 mm for phalli.

Figure 4.

Female genitalia of members of the calamistis subgroup. A S. calamistis; B S. kabirara sp. nov.; C S. madagascariensis; D S. mapalense sp. nov.; E S. oriaula; F S. teke sp. nov. Scale bar = 1 mm.

Figure 5.

Last instar larvae of members of the calamistis, incerta and nonagrioides subgroups. A S. calamistis; B S. mapalense sp. nov.; C S. oriaula; D S. epunctifera; E S. firmata; F S. incerta; G S. kamba sp. nov.; H S. penniseti; I S. roumeti. Scale bar = 10 mm.

Distribution

All sub-Saharan Africa (including Madagascar and Mascarene Islands), recorded from all vegetation mosaics (White 1983) (Fig. 6) belonging to the corresponding bioregions (sensu Linder et al. 2012).

Figure 6.

Distribution map of sampled specimens of the calamistis subgroup, with the exception of S. calamistis specimens.

Ecology

Larvae were collected on young stems and shoots of 57 host plant species either belonging to Cyperaceae (12 species from three genera) or Poaceae (45 species from 23 genera) (Table S4).

Remarks

This important pest species of cultivated grasses (maize, pearl millet, rice, sorghum, sugarcane and wheat; Gahukar 1984; Leslie 1994; Haile and Hofsvang 2001; Kfir et al. 2002) is very common and widespread in sub-Saharan Africa. As reported by Tams and Bowden (1953), the species is variable in size, wing patterns and colours, depending on its geographical origin. In addition to the specimens recorded above, respectively seven, 13 and 33 S. calamistis specimens were found in the MCSN, NHM and TMSA. Unlike S. nonagrioides that is frequently associated with Typha domingensis Pers. in sub-Saharan Africa, S. calamistis larvae have never been reared from this plant.