Protagonista lugubris, a cockroach species new to China and its contribution to the revision of genus Protagonista, with notes on the taxonomy of Archiblattinae (Blattodea, Blattidae)

Abstract The blattid genus Protagonista Shelford, 1908, which is tentatively assigned to the subfamily Archiblattinae (= Planeticinae), is reported from China for the first time with illustrations and description of Protagonista lugubris Shelford, 1908. It is a wood-dwelling and potentially a wood-feeding species. The male and female genitalia of Protagonista are described and illustrated for the first time. The species Protagonista pertristis Hanitsch, 1923 is revived from the synonymy of Protagonista lugubris, and the remaining three nominal species that were also considered as synonyms of Protagonista lugubris are now recognized as synonyms of Protagonista pertristis. In agreement with Princis (1965), we propose that Eroblatta Shelford, 1910, a genus closely related to Protagonista, should be placed in the subfamily Archiblattinae rather than Blattinae. However, the taxonomy of Archiblattinae is problematic and awaits revision. Photos and a key to species of Protagonista and Eroblatta are provided, including photos of the holotypes of the synonymized nominal species. In addition, although Planeticinae is the senior synonym of Archiblattinae, the priority of the latter should be maintained since it is in prevailing usage based on the Article 40.2 in ICZN 4th edition.


Introduction
The cockroach subfamily Archiblattinae (= Planeticinae) belongs to the family Blattidae and is distributed in southeast Asia. Archiblattinae has a controversial taxonomic history. Planeticidae was erected by Walker (1868) based on the genus Planetica Saussure, 1863. This was subsequently synonymized with the genus Archiblatta Snellen van Vollenhoven, 1862 by Saussure (1869). It was Kirby (1904) that then established the subfamily Archiblattinae (family: Blattidae) on the basis of the genus Archiblatta, making Planeticidae a synonym. Meanwhile, Kirby (1904) included Catara Walker, 1868in this subfamily. Handlirsch (1930 raised Archiblattinae to Archiblattidae, which was assigned to Blaberoidea by Princis (1960). Shelford (1908Shelford ( , 1910 established Protagonista and Eroblatta, respectively, and placed them in the subfamily Blattinae (note all cockroaches were included in Blattidae). Princis (1965) included the four genera listed above in Archiblattidae. Grandcolas (1996) synonymized Archiblattidae with Blattidae. Recently Roth (2003) consented to this placement according to the male and female subgenital plates of three genera (Archiblatta, Catara, Protagonista) out of the four, but listed Eroblatta under Blattinae. From then on, the subfamily Archiblattinae was accepted as comprised of 3 genera (Archiblatta, Catara and Protagonista).
The genus Protagonista was established by Shelford (1908). He described Protagonista lugubris from the Manson Mountains, Tonkin (i.e. northern Vietnam) without any description of the male genitalia and designated it as the type species of Protagonista. Later Hanitsch (1923Hanitsch ( , 1925Hanitsch ( , 1929Hanitsch ( , 1931 described another four species belonging to the genus Protagonista from Southeast Asia: P. pertristis, P. fusca, P. aterrima and P. laeta. But Bruijning (1948) synonymized these 4 species with Protagonista lugubris Shelford, 1908 according to the difference in the depth of color of the tegmina and the whole body, which has less taxonomic value. Until now, the genus Protagonista was only reported from Vietnam, Malaysia, Indonesia and Singapore.
Previous studies of the subfamily Archiblattinae are fairly limited. In this paper, we report one known species P. lugubris newly discovered from China, distributed in Hainan and Guangxi, of which the male and female genitalia are described in detail for the first time. A key to all species of Protagonista and the related genus Eroblatta is given. The taxonomic status of this subfamily and the genus Eroblatta, as well as the validity of the name Archiblattinae, are discussed. We also deal with the synonymy of Protagonista lugubris based on the examination of holotypes, geographical distribution and original descriptions.

Material and methods
The terminology for the body, male and female genitalia used in this paper mainly follows McKittrick (1964) and Roth (2003). Terminology of veins follows Haas and Kukalová-Peck (2001) with modification by Li and Wang (2015). The specimens are deposited in the College of Plant Protection, Southwest University, Beibei, Chongqing, China (SWU), unless otherwise noted. Measurements are based on specimens examined. Widths of pronota and tegmina are based on their widest portion. The genital segments of the examined specimens were macerated in 10% NaOH and observed in glycerin jelly using a Motic K400 stereomicroscope. All drawings were made with the aid of a Motic K400 stereomicroscope. All specimens deposited in SWU were photographed using a digital camera (Canon EOS 50D) coupled with a macro lens (Canon EF 100mm f/2.8 USM). The photographs were processed in Helicon Focus software.
The terminology of Roth (2003) is used in describing the spines (armament) on the antero-ventral margin of the front femur, where type A refers to a row of stout or "heavy" spines which decrease gradually in size distad, terminating in two or three large spines, rarely up to five large terminal spines. The number of stout terminal spines are indicated by subscripts so that one or two terminal spines are Type A 1 or A 2 .
The standard barcoding sequences of the mitochondrial COI gene (658 bp) of Protagonista lugubris from Hainan and Guangxi are approved, which are deposited in GenBank under the accession numbers KU511283, KU511284, KU511285 and KU511286.

Taxonomy
Subfamily Archiblattinae Kirby, 1904Kirby, (1868 Protagonista Shelford, 1908: 158;Shelford 1910: 22;Hanitsch 1923: 443;Bruijning 1948: 117;Princis 1965: 388;Grandcolas 1996: 520;Roth 2003: 33. Type species: Protagonista lugubris Shelford, 1908. Generic diagnosis. The genus Protagonista is remarkable on account of the shape of the pronotum (as long as broad, quadrangular, with rounded angles, sides not de-flexed), and the pubescence on its pronotum and tegmina (after Shelford 1908). The other three genera of Archiblattinae differ from it by the apterous female and the unarmed or weakly-armed femur (Archiblatta and Catara) or by the tibia having three rows of spines (Eroblatta). Description. Antennae slightly moniliform. Ocelli present. Pronotum as long as broad, quadrangular, with rounded angles, sides not deflexed and not covering vertex (Figs 1-6 appears to show it covering the vertex, however this is an artifact of the photo angle). Pronotum and tegmina with fine pubescence. Tegmina and hind wings fully developed in the male, exceeding the apex of the abdomen. Tegmina short and truncated in the female, hind wing vestigial to a small lobe. Styli present and cerci moderate. Legs slender; front femora Type A 2 ; hind tibia with 2 rows of spines along outer margin; hind metatarsus very long, considerably exceeding the remaining joints in length; the tarsal pulvilli present on the proximal four tarsomeres; arolia minute.
Female. Body black (Figs 3-4, 48). Eyes, ocelli and antennae similar to those of male. Vertex and face reddish brown. Labial palpi and maxillary palpomeres brown. Pronotum black. Abdominal terga black, but with the last segment brown. Abdominal sterna black and center reddish brown. Legs and cerci brown.
Vertex exposed, with 3 longitudinal shining stripes. Face punctuated. Tegmina short, just exceeding the metanotum, with punctures and scattered erect pubescence, heavily sclerotized with metallic shine. Hind wings much reduced. Legs slender, front femur Type A 2 . Each hind tibia with 2 rows of spines along outer margin. Hind metatarsus exceeding the remaining joints in length.
Nymph. Body color, characters of pronotum and antennae similar to those of adults. Legs light brown. Cerci reddish brown. Infraspecific variation. The individual differences in morphological characters mainly involve: 1) the number and shape of smooth areas of pronotum (Figs 7,16,28,37); 2) the dentate tine close to the largest tine of the serrated edge of sclerite R1 sclerotized (Figs 12,22) or not sclerotized (only one case, Figs 33, 43); 3) body color (Figs 1-2, 5-6). We provide pictures for detailed comparison (one male from Hainan, one male from Guangxi) illustrating the appearance of individual differences (Figs 1-2, 5-6, 7-12, 13-22, 28-33, 34-43). These infraspecific variations cannot separate the populations from each other into different species and the key morphological characters strongly suggest they are conspecific. However, their COI genes show a great genetic divergence among them: the standard barcoding sequence of one (Baisha) of the three Hainan populations has a distance of 3.0% and 3.1% from the other two (Wuzhishan, Baoting) respectively, and it is far distant (4.6%) from the Guangxi population, which in turn is very distinct from the two remaining Hainan populations (Wuzhishan, 6.1%; Baoting, 6.2%).
Male measurements (mm). Body length: 17.5-21.0. Total length including tegmen: 19.0-24.5 . Pronotum length × width: 4.9-5.5 × 5.5-6.5. Tegmen length × width: 15.0-20.0 × 5.0-6.5. Habitat. The adult P. lugubris were observed in shrubs at night by the collectors who also found the nymphs and adults in rotten wood. Their rugged pronotum with thickened and raised margins, which resembles that of Cryptocercinae and Panesthiinae is conducive to moving about in rotten wood. However, if Protagonista utilizes the wood tunneled by other organisms or if they bore the wood themselves is yet to be seen. Additionally, wood feeding has not been observed but is still a possibility in Protagonista and the other morphologically similar Archiblattinae.

Discussion
Validity of the name Archiblattinae. As explained in the introduction, the scientific names Archiblattinae and Planeticinae are synonyms. According the Principle of Priority in ICZN, Archiblattinae should be abandoned and the earlier name Planeticinae is valid although its type genus is no longer valid. But the name Planeticinae/-idae has been ignored for a long time and the substitute name Archiblattinae/-idae is in prevailing usage since Princis's (1965) catalogue; therefore the priority of the latter should be maintained based on the ICZN rule 40.2.
Taxonomic status of subfamily Archiblattinae and its genera. The subfamily Archiblattinae has a controversial taxonomic history since it was established. Although Kirby (1904) erected the subfamily Archiblattinae based on the genus Archiblatta, Shelford (1910), Hanitsch (1915) and Bruijning (1948) placed Archiblatta in Blattinae. Subsequent authors also have different suggestions on the taxonomic status of family Archiblattidae. Princis (1965) listed it as a family, but Grandcolas (1996) synonymized Archiblattidae with Blattidae and assigned Archiblatta and Catara to Blattidae. Roth (2003) suggested that Archiblatta, Catara and "?Protagonista" should be in their own subfamily because of the absence or greatly reduced femoral armament and reserved the subfamily Archiblattinae (Fam. Blattidae). Inward et al (2007) and Legendre et al (2015) found the subfamily Archiblattinae (Archiblatta) and Blattinae to be respectively monophyletic. Klass and Meier (2006) placed Archiblatta as sister to Polyzosteriinae + Blattinae. Djernaes et al. (2015) indicated that the structuring of Blattidae into the subfamily Polyzosteriinae (Drymaplaneta, Eurycotis), Archiblattinae (Archiblatta), and Blattinae (Periplaneta, Deropeltis) may be artificial since Archiblattinae were placed within Blattinae. In spite of the few studies on the Archiblattinae, our knowledge about the genera other than Archiblatta is still so lacking that it is reasonable to question their classification in Archiblattinae. There is uncertainty regarding the taxonomic status of Catara, Protagonista, and Eroblatta. When comparing the male genitalia of P. lugubris with those of Archiblatta hoeveni (illustrated by Klass 1997) and other species (e.g. Periplaneta americana, Periplaneta brunnea, Periplaneta ceylonica, Blatta orientalis, Neostylopyga rhombifolia, Melanozosteria nitida) in the subfamilies Blattinae and Polyzosteriinae, we find that the male genitalia of Archiblattinae and Blattinae are closer to each other than to those of Polyzosteriinae. Furthermore, Protagonista and Archiblatta are more similar to each other than they are to the genera of Blattinae. However, we failed to find independent, distinct morphological features separating the two taxa; thus the male genital differences between them might not be adequate as diagnostic characters in subfamily-group taxonomy. Our observations coincide, in a phylogenic sense, with Djernaes et al. (2015). However, Archiblattinae is easily distinguished from other blattid cockroaches by the special pronotum (hardened and rugose, sides thickened and not deflexed) and the special tibia which are extraordinarily cylindrical with sparse spines. Additionally, the cladistic results themselves are still in dispute. Yet we should not simply rely on the cladistic results solely to alter the classification, even if a widely accepted cladistic conclusion is demonstrated in the future. Therefore we propose to retain the validity of the subfamily Archiblattinae and the arrangement of the 4 genera mentioned above before a comprehensive taxonomic and phylogenetic study has been conducted. If done, this future study should on one hand confirm whether the subfamily Archiblattinae is monophyletic, and should on the other hand discern the relationships among the 4 genera and give an acceptable arrangement of them.  Hanitsch (1932) synonymized P. aterrima Hanitsch, 1929with P. fusca Hanitsch, 1925. Bruijning (1948 synonymized 4 species (P. aterrima, P. fusca, P. pertristis Hanitsch, 1923 (Figs 44-47) and P. laeta Hanitsch, 1931) with P. lugubris Shelford, 1908, as he considered a slightly differing coloration not an important specific character that is coincident with Hanitsch (1932). We also consider the four nominal species P. aterrima, P. fusca, P. pertristis and P. laeta to be the same species, but separate from P. lugubris Shelford, 1908. We base this on evidence from holotypes, original descriptions and geographical distribution. They are all from a limited region (Sunda Shelf) far away from the localities of P. lugubris (Northern Vietnam and South China). We agree with Hanitsch's (1932) synonymy based on the holotypes: the distal half of coxae and the base of femora of P. aterrima are light testaceous, as is P. fusca. Meanwhile we indicate that Bruijning's (1948) viewpoint should be revised: P. aterrima, P. fusca, P. pertristis and P. laeta are identical indeed, but they are not synonyms of P. lugubris. Therefore P. aterrima, P. fusca and P. laeta are the junior synonyms of P. pertristis. P. pertristis is distinguished from P. lugubris by the following characters: 1) coxae of all the legs with the distal half and the base of femora are orange yellow on P. pertristis and P. fusca, but all the legs of P. lugubris are uniformly brown (Note: The character of difference in color depth of tegmina and the whole body has less taxonomic value, in accordance with Hanitsch (1932) andBruijing (1948). However, the distinct coloration of part of the body (such as this case) should be the criteria for species differentiation.); 2) cerci of P. pertristis and P. fusca slender, yellow, but those of P. lugubris thicker and shorter, brown. In conclusion, the genus Protagonista is comprised of 2 species, P. lugubris Shelford, 1908 andP. pertristis Hanitsch, 1923, the latter with three junior synonyms: P. aterrima, P. laeta and P. fusca.

Synonymy of Protagonista lugubris.
Taxonomic status of the genus Eroblatta. The genus Eroblatta was erected based on Protagonista borneensis Shelford, 1908by Shelford (1910. The genera Eroblatta and Protagonista differed from each other in the spines on outer margin of tibiae according to his original description. Eroblatta has 3 rows of spines along outer margin of tibia, whereas there are only 2 rows in Protagonista. Eroblatta was also placed in Archiblattidae by Princis (1965) but Roth (2003) listed it under Blattinae without any explanation; within the period between these two publications (i.e. Princis 1965 andRoth 2003), works with respect to Eroblatta were absent, the change on its taxonomic arrangement is deemed to be simply an inadvertent error. After examining the holotype (Fig. 50) and according to original description, it is seen clearly that Eroblatta borneensis has the typical archiblattid pronotum (hardened and rugose with sides thickened and not deflexed) and the special tibia, which are extraordinarily cylindrical with sparse spines. Therefore Eroblatta is exactly, for the time being, a member of Archiblattinae. Figure 50. Eroblatta borneensis (Shelford 1908), holotype and labels. Scale bar: 5.0 mm. Shelford, 1908 Vietnam (Tonkin, type locality), South China P. pertristis Hanitsch, 1923, stat. rev. Sunda Shelf including Malay Peninsula (type locality) syn. P. fusca Hanitsch, 1925 Sarawak syn. P. aterrima Hanitsch, 1929Sumatra syn. P. laeta Hanitsch, 1931 Singapore Eroblatta E. borneensis Shelford, 1908 Borneo (Sarawak, type locality)