Mythomantis serrata, Schwarz, Christian J. & Helmkampf, Martin, 2014

Schwarz, Christian J. & Helmkampf, Martin, 2014, A remarkable new species of Mythomantis Giglio-Tos, 1916 from northern Borneo, with notes on the systematics of Deroplatyinae Westwood, 1889 (Mantodea: Mantidae), Zootaxa 3797 (1), pp. 120-129 : 121-127

Mythomantis serrata sp. nov.

( FIGURES 1–7 )

Type material. Holotype: m#, Sarawak, Gunung Mulu Nat. Park, R. G. S. Exped. 1977–78, J. D. Holloway et al., B. M. 1978-206; Site 19, March, W. Melinau Gorge, 100 m, 427567, Alluvial forest, Acl-understorey (NHM). Paratypes: m#, same data as holotype; m#, Sarawak, Gunung Mulu Nat. Park, R. G. S. Exped. 1977–78, J. D. Holloway et al., B. M. 1978-206; Site 21, March, W. Melinau Gorge, 130 m, 423576, Alluvial/kerangas bank, Aclunderstory (NHM); m#, Sarawak, Gunung Mulu Nat. Park, R. G. S. Exped. 1977–78, J. D. Holloway et al., B. M. 1978–206; Site 23, April, W. Melinau Gorge, 250 m, 430558, FEG 4 Limestone forest, MV, canopy-understory (this specimen will be deposited in the MNHN, with kind permission of George Beccaloni, NHM); m#, MY, Sabah (Borneo), Danum Valley, 340 m, N04°58’06.9”, E117°50’16.9”, 16. Mar. 2003, col. M. Helmkampf, Select. logged forest, UV light-trap (this specimen will be deposited in the SMNH); m#, Malaysia, Sarawak, Gunung Mulu National Park, Head Quarters, 125 m, N04°02’32,97”, E114°48’52,20”, 1–25 July 2009, G. J. Svenson leg.; GSMC 000080; MN393 Primary Voucher (this specimen will be deposited in the CMNH).

Description (based on complete type series). Body brownish ( Fig. 1 ), paler parts possibly with greenish tinge in life, 56.5–62.5 mm long. The specimen from Sabah is smaller than those from Sarawak.

Head ( Fig. 2 ) 3.6–4.0 mm long and 5.9–6.2 mm wide, yellowish brown; frontal shield, area between the ocelli, labrum and juxta-ocular protuberance mottled with dark brown. Eyes round, exophthalmic, traversed by an irregular yellowish band. Scutellum pentagonal, wider than long. Ocelli large. Vertex with four sulci, juxta-ocular protuberance weakly pronounced. Antennae about 25 mm long, filiform, proximal ten segments pale, the following becoming successively darker. Maxillary palps pale, labial palps pale with distal segment darkened.

Pronotum ( Figs. 1 , 4 ) slender, 24.0– 28.1 mm long, metazona 3.8–4.1 times as long as prozona. Prozona 4.9–5.7 mm long, granulate, with denticulate margins. Metazona triangular in cross section, dorsally with a sharp ridge and occasional granulations, 3.1–3.7 mm wide at supracoxal dilation and 1.5–1.8 mm at the middle (excluding lateral lobes). Lateral margins of metazona with 4–6 (only 3 on the right side of the Sabah specimen) large, triangular, sawtooth-like lobes interspersed with smaller teeth. Prosternum exhibiting an inconspicuous, longitudinal brownish band.

Forecoxae 12.0– 13.5 mm long, slender. Posterior margin irregularly denticulate, anterior margin with 6 dark teeth. Dorsal surface indistinctly banded, ventral surface dark brown with a light band in the distal third. Trochantera light brown, with an indistinctive dark spot near base of femur. Anterior femora ( Fig. 3 ) 14.8–16.6 mm long (the Sabah specimen showing a regenerated left femur), slender, slightly sinuate, with 4 posteroventral, 14–15 anteroventral and 4 discoidal spines, all with dark apices. Posterior surface with several dark tubercles, anterior surface smooth, sometimes with 3 indistinct dorsoventral bands. Claw groove located in distal half of femur, featuring a dark spot. Anterior tibiae 6.4–7.2 mm long, with 5 posteroventral and 12 (occasionally 13) anteroventral spines; posteroventral spines curved, all spines with dark apex. Anterior tarsus long, yellowish, ventrally darkened.

Middle and hind coxae longitudinally striped. Metathoracic ear present, of the DK type ( Yager & Svenson 2008). Middle and hind femora with thickened bases, 12.3–13.5 mm and 13.6–15.0 mm long, respectively, with interrupted longitudinal brown bands and pale annulations in the proximal and distal third. Apical lobes ( Fig. 5 ) long, acute, with a dark longitudinal stripe. Middle tibiae about the same length as femora, hind tibiae with 15.5–17.0 mm longer than corresponding femora. Proximal tarsomere of hind legs about twice as long as the remaining segments combined.

Tegmina 36.5–42.5 mm long, 1.5 times as long as pronotum. Costal area opaque, discoidal area subhyaline, with small smoky spots and larger opaque yellowish patches, the latter concentrated in the distal portion of the tegmen. Stigma conspicuous and shiny, of ivory color (sometimes encircled by a large whitish patch) and bordered by a dark spot on each end. Alae 33.7–38.0 mm long, smoky with preapical hyaline band, apical part opaque. Small veinlets of anal area whitish.

Abdomen with small lateral lobes and slightly larger medio-ventral lobes. Anterior part of tergites ochraceous, posterior part dark. Sternites brownish, proximal segments sometimes mottled with pale markings, projecting medio-posteriorly into a lobe. Supra-anal plate ( Fig. 6 a) wider than long, with concave margins and rounded apex. Cerci ( Fig. 6 a) pale, setose, moniliform, 5.3-5.4 mm long, with 18–19 cercomeres; distal cercomere slender, acute, longer than preceding segment. Subgenital plate ( Fig. 6 b) stout, slightly asymmetrical, with a dark, irregular median stripe in the anterior half and two even darker bands on each side. Styli long and slender.

Male genitalia ( Fig. 7 ) typical for the genus, apical process of left dorsal phallomere strongly curved, phalloid apophysis digitiform, slightly acuminated, moderately pigmented, except for a small, heavily sclerotized and spiny dorsal lobe just before its distal end; distal process of ventral phallomere bilobate (boat-shaped), heavily sclerotized, irregularly dentate, left lobe pointed; ventral process of right phallomere strongly sclerotized, bulged, apex roughly acuminate.

Female: unknown.

Diagnosis. The new species can be easily distinguished from both M. confusa and M. gracilis by the unique shape of the pronotum, exhibiting rather large triangular lobes, a pattern not observed in any other mantid species known to date, and by the concave margins of its supraanal plate. Additionally, its body length exceeds that of both congeners. Mythomantis serrata resembles M. gracilis in the shape of the pronotum, the ratio length/width exceeding 7.1 in both species, whereas the pronotum of M. confusa is more robust than that of the two other species (P length/P width of males: 6.2–6.8). Mythomantis serrata also differs from M. confusa and M. gracilis in having the distal process of the ventral phallomere pointed on its left side, the same structure being blunt in the two latter species, and by having a heavily sclerotized dorsal lobe on the phalloid apophysis of the left dorsal phallomere. The genus remains poorly known due to the paucity of specimens, particularly females, which are known only for M. confusa . The males of the three species of Mythomantis can be distinguished by the following key (after Roy 2004, modified):

1. Margin of metazona with large sawtooth-like lobes..................................................... serrata

- Margin of metazona more or less denticulate, but never with sawtooth-like lobes................................... 2

2. Slender species, width of pronotum not exceeding 2.6 mm at supracoxal dilation and 1.4 mm at middle of metazona. gracilis

- Robust, width of pronotum exceeding 2.9 mm at supracoxal dilatation and 1.9 mm at middle of metazona......... confusa

Etymology. The species is named for the conspicuous, large, sawtooth-like lobes on the margins of the metazona (serratus [Latin]: toothed, serrate).

Distribution. Mythomantis serrata is known only from Malaysian Borneo (NE Sarawak and Sabah, Fig. 9 ). Mythomantis confusa is definitely recorded from Java and Sumatra ( Roy 2004). Older Sulawesian and Bornean records dating back to De Haan (1842) and quoted by Giglio-Tos (1927), Beier (1935b), and Ehrmann (2002) need confirmation, as De Haan did not specify which of the locations have to be attributed to the male (= Euchomenella heteroptera ) and which to the female (= Mythomantis confusa ) of his Mantis heteroptera . Mythomantis gracilis is only known from the (probably lost) type specimen from Ambon Island and an additional male from Java deposited in the SMNK ( Roy 2004).

Biological notes. Due to its rarity and restricted distribution not much is known about the ecology of M. serrata , except that it seems to be confined to mature dipterocarp forests. As this biologically highly diverse area is severely threatened by deforestation and expanding oil palm plantations, the species might be endangered by habitat loss.

Systematic placement. When Giglio-Tos (1916) erected the genus Mythomantis for Euchomena confusa Westwood, 1889 , he regarded it as related to Pseudempusa Brunner von Wattenwyl, 1892. Later on, Giglio-Tos (1927) placed the two genera more precisely among Eufischeriellinae, which largely correspond to today’s Rivetinini (see Ehrmann 2002). Beier (1935b) did not retain this systematic placement, but moved Mythomantis to the pantropically distributed tribe Angelini, which was later elevated to subfamily level ( Beier 1964), a position which remained unchanged until this study. Currently, the Angelinae are assigned to the family Mantidae and comprise the following genera: Angela Audinet-Serville, 1839 (Neotropical) , Stenopyga Karsch, 1892 , Agrionopsis Werner, 1908 and Leptocola Gerstaecker, 1883 (all Afrotropical), and Euchomenella Giglio-Tos, 1916 , Indomenella Roy, 2008 , Cotigaonopsis Vyjayandi, 2009, Tagalomantis Hebard, 1920 and Mythomantis Giglio-Tos, 1916 (all Oriental) ( Ehrmann 2002, Ghate & Mukherjee 2004, Otte & Spearman 2005, Roy 2008b, Vyjayandi et al. 2009). Roy (2002) also presumed Rhodomantis Giglio-Tos, 1917 to belong to Angelinae . One additional, often cited, Neotropical monotypic genus, Thespoides Chopard, 1916 , turned out to be a chimeric individual ( Rivera 2014). Members of this subfamily are characterized by their slender, more or less elongate bodies, comparatively short foretibiae with respect to the forefemora, and brachypterous females. The only exception is Mythomantis confusa (and presumably its congeners, whose females, however, are unknown), which has macropterous females and a slender, but not especially elongate body. Beier (1935b) assumed a close relationship between Mythomantis and Euchomenella based on the fact that, in contrast to other Angelinae , both genera exhibit closed inner apical lobes on the forecoxae.

Roy (2001) doubted the affiliation of Mythomantis with the Angelinae , but reconfirmed its systematic position after a thorough revision of the genus ( Roy 2004). However, recent morphological and molecular studies ( Yager & Svenson 2008, Svenson & Whiting 2009) presented evidence for Angelinae to be polyphyletic. According to the latter study (in which Mythomantis was not included) the group consists of several superficially similar but not directly related groups of species that have independently evolved into the same ecomorph. The Neotropical genus Angela plesiomorphically lacks the metathoracic hearing organ (“cyclopean ear”, see Yager & Hoy 1986, 1987, 1989, Yager 1999) used in bat attack avoidance by “higher” mantids (Yager & May 1990, Yager et al. 1990, Cumming 1996, Triblehorn & Yager 2001, 2005, Triblehorn et al. 2008, Yager & Svenson 2008). In contrast, the Oriental and Afrotropical members of Angelinae all have a functional ear, albeit secondarily more or less reduced in brachypterous females ( Svenson & Whiting 2008). Nonetheless, the Oriental Angelinae do not group with their Afrotropical counterparts in the molecular phylogeny, although they are members of the same major clade (clade 253 in Svenson & Whiting 2009).

Instead, the two analyzed oriental Angelinae , Euchomenella and Indomenella , formed a clade sister to Deroplatys Westwood, 1839 , an Indomalayan genus of macropterous, leaf-like looking mantids with enlarged, foliaceous pronota and subapical lobes on the ventral side of mid- and hind femora (with the exception of D. indica Roy, 2007 ). Deroplatys , along with the Madagascan Brancsikia Saussure & Zehntner, 1895 , are classified in the subfamily Deroplatyinae Westwood, 1889 (Fam. Mantidae ). Beier (1935a, 1964, 1968) also assigned the Indomalayan genus Parablepharis Saussure, 1870 to Deroplatyinae due to its foliaceous pronotum and subapical lobes on mid and hind legs, while Ehrmann & Roy (in: Ehrmann 2002) transferred Parablepharis to Hymenopodidae , Epaphroditinae .

Comparison of genitalia of Mythomantis with those of Pseudempusa ( Fig. 8 ), Deroplatys ( Anisyutkin 1998, Roy 2007) and members of oriental “ Angelinae ” ( Roy 2001, 2008b, Vyjayandi et al. 2009) indeed suggests a close relationship between Pseudempusa and Mythomantis , as first proposed almost hundred years ago by Giglio-Tos (1916). Both genera share most genital similarities with Deroplatys , and to a lesser degree with oriental “ Angelinae ”, but are very different from those of Rivetinini (compare Figs. 7 and 8 and Roy 2004 with La Greca & Lombardo 1983, Roy 2001, 2007, Vyjayandi et al. 2009 and Shcherbakov 2012). There is a sclerotized ridge on the left side of the ventral phallomere which extends onto the distal process, ending in a small lobe. It is well visible in Pseudempusa and in at least two Deroplatys species ( D. desiccata Westwood, 1839 and D. indica , see Anisyutkin 1998 and Roy 2007), and it is also present in Mythomantis , although without extending onto the distal process. On the right side of the distal process of the ventral phallomere there is a sclerotized lobe in Mythomantis and Pseudempusa , but not in Deroplatys . It may have been lost in the latter genus or represent an autapomorphy of Pseudempusa + Mythomantis . The distal process of the ventral phallomere is elongate and strongly curved in D. desiccata , D. indica and Pseudempusa , while in Mythomantis it shows considerable shortening, with the “primary” hook (i.e. the distal process) being reduced in size and curvature. On the other side, the lobe on the right side is more elongate in Mythomantis . It seems to have adopted the functions of the primary hook to a large extent. The phalloid apophysis is digitiform with acute or subacute apex in Mythomantis , Pseudempusa , and some Deroplatys species like D. desiccata and D. indica .

While Giglio-Tos (1916) did not specify why he regarded Pseudempusa and Mythomantis as closely related, Beier (1964) assigned Pseudempusa to Miomantinae due to crenulated dorsal margins of the forefemora. Ehrmann (2002) placed the genus more specifically among Miomantinae-Rivetinini. The character used by Beier is highly homoplastic, though, since it evolved independently in different unrelated mantodean lineages (e.g. among Acromantinae , Amelinae , Miomantinae and Stagmomantinae , Giglio-Tos 1927, Ehrmann 2002). Another morphological trait that on a first glimpse may support a placement of Pseudempusa among Rivetinini , the eyespot on the alae, does not resemble the ones seen in Rivetinini , and has most likely evolved independently. No author so far considered the genital morphology of Pseudempusa , which is depicted here for the first time. The similarities of the genitalia of Pseudempusa to those of Mythomantis on the one hand, and Deroplatys on the other, and the difference between the genitalia of these three genera and those of other mantodeans, indicate that Deroplatys , Pseudempusa and Mythomantis most likely form a monophyletic unit, despite rather different external morphology. Therefore, we transfer both Mythomantis (from Angelinae ), and Pseudempusa (from Miomantinae Rivetinini ) to Deroplatyinae . Other potential characters supporting this relationship are the basal, partly sclerotized lobe on the right side of the ventral phallomere, closed apical lobes on the forecoxae, elongate apical lobes on the walking legs (most strongly expressed in Mythomantis ), and an ivory-white stigma. Deroplatys and Pseudempusa also share the foliaceous expansion on the pronotum (in Pseudempusa around supracoxal dilatation, in Deroplatys along entire length of pronotum), and two oblique dark stripes on the metazona, a consistent pattern independent of body coloration; the latter feature is also discernible in females of Mythomantis confusa . Deroplatys and Mythomantis , on the other hand, have the pale subapical band on the hindwings in common. This feature is obscured in Pseudempusa by the eye-spot present at the same place; it remains unclear whether the eye-spot is homologous to the pale band.

When compared to other groups, the genitalia of Deroplatyinae in its new sense most closely match those of the oriental “angeline” genera Indomenella and Tagalomantis ( Roy 2007, Shcherbakov 2012). The genitalia of Euchomenella and Cotigaonopsis are much more simplified ( Roy 2001, Vyjayandi et al. 2009, Shcherbakov 2012), but not fundamentally different from those of Deroplatyinae , Indomenella and Tagalomantis . Whether this reflects a close relationship, as indicated by molecular phylogeny ( Svenson & Whiting 2009), remains an open question until more morphological and molecular data become available.

In contrast, we believe that both Brancsikia and Parablepharis are not members of Deroplatyinae but more closely related to Hymenopodidae . This was suggested for Parablepharis before by Roy (2008a), and is supported, among other things, by the simple structure of their genital organs, devoid of the structures discussed above, and by the configuration of the postero-ventral spines on the anterior tibiae. Parablepharis also has a bifid process on the vertex, while Brancsikia has strongly conical eyes ending in a small tubercle ( Paulian 1957, Roy 2008a, Schwarz pers. obs.). None of these structures are found in Deroplatys , but are frequently seen in Hymenopodidae ( D. truncata Guérin-Méneville, 1843 has convergently evolved slightly conical eyes of a different shape than those of Brancsikia ). The two genera also lack other features characteristic for Deroplatyinae , like elongate apical lobes on mid and hind legs, the oblique stripes on the metazona or the pale subapical band on the alae. Their phylogenetic position among Hymenopodidae , however, needs to be more thoroughly investigated.

Mythomantis serrata represents an intriguing addition to the known Southeast Asian mantid fauna. Despite its striking morphology, the species had so far escaped formal description, highlighting how poorly the diversity of this group is still understood even in the face of accelerating habitat destruction. Future studies of Mythomantis focusing on the analysis of both morphological and molecular characters in a phylogenetic framework will not only allow to corroborate the systematic rearrangement outlined here, but also to resolve the phylogenetic relationships both within and near the Deroplatyinae , as well as yield valuable insight into the evolution of mantid characters.

Figures

FIGURE 8. Ventral phallomere and phalloid apophysis of Pseudempusa pinnapavonis (scale: 0.1 mm)

FIGURE 9. Known distribution of the genus Mythomantis (after De Haan 1842, Werner 1921, 1922, 1925, 1928, Roy 2004 and this study): circles = M. confusa; squares = M. serrata; triangles = M. gracilis. Question marks indicate ambiguous locations (see main text). Map by Thomas Lessman (www. WorldHistoryMaps. info), used with permission.

FIGURE 1 – 3. 1. Habitus of holotype, dorsal view (large scale units: 10 mm). 2. Head of holotype, anterior view (scale: 1 mm). 3. Left fore femur and tibia of holotype, posterior view (scale: 2 mm).

FIGURE 4 – 7. 4. Pronota. (a) Holotype; (b) paratype from Sarawak; (c) paratype from Sabah (scale: 5 mm). 5. Apex of left hind femur of holotype (scale: 1 mm). 6. Abdomen apex of holotype. (a) Supra-anal plate with cerci, dorsal view (hairs omitted); (b) subgenital plate, ventral view (scale: 5 mm). 7. Male genitalia. (a) Genitalia of holotype, ventral view (scale: 1 mm); (b) detail of phalloid apophysis of holotype, dorsal view; (c) and (d) phalloid apophyses of two paratypes, dorsal and ventral view, respectively (scale: 0.1 mm).