The story of the Malagasy devils (Orthoptera, Tetrigidae): Holocerus lucifer in the north and H. devriesei sp. nov. in the south?

Abstract Madagascar is home to some of the largest and most colorful pygmy grasshoppers (Tetrigidae) in the world, known as ‘Malagasy Metrodorinae’. Among them, Devil’s pygmy grasshoppers (genus Holocerus Bolívar, 1887) are unique in having two long spines on the back, which are modified internal lateral pronotal carinae. The genus Holocerus was composed of two species – H. lucifer (Serville, 1838) and H. taurus Rehn, 1929 syn. nov., but here it is evidenced that the latter represents a junior synonym of the former. Simultaneously, H. devrieseisp. nov. is described as a species new to science. Holocerus lucifer is a northern species of paler coloration and longer spines (distributed from Marojejy and Maroantsetra in the north to Zahamena in the south), whereas H. devrieseisp. nov. represents the southern and darker species (distributed from Vohimana and Andasibe-Mantadia in the south to the Antongil Bay in the north). There are potential overlaps in the distribution of the two species, but without more georeferenced localities, it is impossible to discriminate whether they occur only sympatrically or also syntopically.

We know about the existence of two Holocerus species, as defined by Rehn (1929). Those are Serville's H. lucifer, which is, according to Rehn, a darker species with shorter dorsal spines, and Rehn's H. taurus, a yellowish-green species with long spines. After the insight into the entomological collections of the MNHN in Paris and the ANSP in Philadelphia, it became evident that both Serville and Rehn described the very same species. In this study, we provide photographs of Tetrix lucifer and Holocerus taurus type specimens, as well as living specimens, and we aim to clarify once and for all what the epithet 'lucifer' refers to. We present an annotated distribution of H. lucifer, a pale colored species with long spines; synonymization of H. taurus Rehn, 1929 syn. nov. with H. lucifer; and an annotated distribution of a newly described darker species with shorter spines, H. devriesei sp. nov.

Identification, nomenclature and taxonomy
The two taxa within the genus Holocerus can be easily distinguished based on the characteristics presented by Rehn (1929). However, Rehn's nomenclature is incorrect. The two could be treated as species or subspecies, but based on the overlaps in their distributions and no evidence of intermediate forms, we treated them as separate species. Nomenclature follows the International Code of Zoological Nomenclature (ICZN 1999), whereas Tetrigidae taxonomy follows the Orthoptera Species File (Cigliano et al. 2020).

Morphological terminology and measurements
We followed Tumbrinck (2014) for the description of general morphology; Devriese (1991Devriese ( , 1995Devriese ( , 1999 for the nomenclature of pronotal carinae; and Pushkar for terminology of pronotal projections (Storozhenko and Pushkar 2017). Measurements were taken on museum specimens in ImageJ2 (Rueden et al. 2017) on the traits previously measured in Holocerus taxonomy (Rehn 1929). The accuracy of the measurements was 0.1 mm.

Photography
Several photos of living individuals were obtained online. These are included in the study with the permission of the photographers (Rowe Becky, Paul Bertner, Marc Hoffmann, and Frank Vassen). (2 spp.), Notocerus (2 spp.), Rehnitettix (1 sp.), and Silanotettix (3 spp.), which share the lack of a tegminal sinus; the lack of humeral carinae; rounded tarsal pulvilli; modified antennal segments; and pronotum with projected and undulated parts (Devriese 1991 Ecology and habitat. Records of adults and nymphs in different parts of the year indicate that the species may be active throughout the year. Devil's pygmy grasshoppers are rainforest dwellers and they inhabit primary and secondary rainforests of Madagascar, where they can be found standing on both wet and dry tree bark of species such as the traveler's palm (Ravenala madagascariensis; order Zingiberales: family Strelitziaceae) ( Figure 4). Holocerus lucifer and H. devriesei sp. nov. are good fliers and can glide for long distances (> 10 m) between rainforest trees and bushes.

Taxonomic treatment
Generic diagnosis and affinity to other genera. The genus Holocerus is similar to the genera Notocerus (2 spp.) and Eurybiades (1 sp.), which are both endemic to Madagascar. Unlike the clearly separated dorsal spines in Holocerus, members of the genus Notocerus (N. cornutus Hancock, 1900 andN. formidabilis Günther, 1974) have a highly elevated dorsum between the spines (making the spines connected). With the morphology of its spines, Notocerus cornutus is more similar to the members of the genus Holocerus than to its own congener, N. formidabilis (which has a high and warty hump). Holocerus can also easily be distinguished from Notocerus by the more prominent eyes. The only species from the genus Eurybiades, E. cerastes Rehn, 1929, is much smaller than the members of the genus Holocerus, and is easily distinguished from both Holocerus and Notocerus members by the long promedial projection, which is spiky and directed forwards. Holocerus and Notocerus have a short and tooth-like promedial projection of the prozona.  Rehn, 1929syn. nov.: Rehn (1929; detailed description of the species based on seven specimens, ♂ HT, 1 ♀ PT and 3 ♂♂ PTs from unknown locality, and 2 ♀♀ PTs from Maroantsetra, depicted) ( Fig. 1E-G); Günther (1939: 91; mentioned, new records); Günther (1959: 11;included in key, depicted); Günther (1970: 79-92; discussed); Günther (1974: 966, 967; new records, depiction); Yin et al. (1996: 876;listed

Additional examined material.
There are a few records of the species since its description -by Rehn (1929) under the name H. taurus; a photographic record from Flickr; a photographic record from Twitter; and several specimens from a museum collection in MNCN, Madrid. For detailed data on all records, see Table 1.

Annotated specific diagnosis.
Holocerus lucifer is similar to H. devriesei sp. nov., which is the only other species of the genus Holocerus. Holocerus lucifer is easily distinguished from H. devriesei sp. nov. with the following set of characteristics: (i) Holocerus lucifer has slenderer femora of fore and mid legs than that of H. devriesei sp. nov.; (ii) dorsal spines are slenderer, more elongate and decurved in H. lucifer than those in H. devriesei sp. nov., and in profile they are, as described by Rehn, 'distinctly falcate, scimitar like'; (iii) middle prozonal spine (promedial projection) is blunter and lower in H. lucifer than that in H. devriesei sp. nov., and (iv) H. lucifer generally has more pale colored parts than H. devriesei sp. nov.
Variability. (Figure 2). Variability is evident in (1) coloration, as there are darker and paler specimens, (2) size, as there are larger and smaller specimens, and (3) the shape of the dorsal spines (elevated internal lateral carinae of the pronotum), as there are specimens in which the spines are more decurved and specimens with almost straight projections.
Distribution and habitat. The species is known from the rainforests of Madagascar, from Marojejy and Maroantsetra in the north to the rainforests of Zahamena in the south. A specimen with the label 'Tamatave' could have been collected in the rainforests in the vicinity of the city of Toamasina, but also anywhere in the wide Toamasina region. The species inhabits primary and secondary rainforests and is probably a good flier, taking into account the observations of its sibling species' ecology. As only one photo of a living individual of this species has been taken to date, hardly anything can be concluded about the species' natural history. Despite being described for already 180 years, this species is less known and understood than H. devriesei sp. nov.
Distribution, habitat and threats. Holocerus devriesei sp. nov. inhabits eastern Malagasy rainforests, from Ranomafana in the south, via Analamazaotra, rainforests around Lakato, Vohimana and Andasibe-Mantadia, all the way to the Antongil Bay, where it has an overlap in distribution with H. lucifer. The natural habitat of the species are primary and secondary rainforests, but on account of forest depletion and habitat degradation, population decline is expected, as well as extinction of its subpopulations in certain territories where the rainforest is undergoing degradation.

Discussion and conclusions
We know about the existence of two species of Malagasy pygmy devils within the genus Holocerus. A pale colored species with longer decurved spines (Figs 1, 2) inhabits the rainforests of Marojejy and Masoala, from around the Antongil Bay in the north to the Zahamena NP in the south (Figure 5), and should be referred to as Holocerus lucifer. A darker species with shorter angular spines (Figs 3, 4), which should from now on be referred to as H. devriesei sp. nov., inhabits the rainforests from Ranomafana in the south to the Bay of Antongil in the north ( Figure 5). There are distribution overlaps between the two species, but as there are no evidences of hybridization, we treated them as separate species.
The name confusion originated when Rehn (1929) described the pale-colored species with longer spines (the true H. lucifer) as a new species under the name H. taurus, while at the same time applying the name 'lucifer' to designate the darker specimens with shorter and angular spines (H. devriesei sp. nov.). Subsequent authors followed Rehn's application of these names (Günther 1939(Günther , 1959(Günther , 1970Yin et al.1996;Otte 1997). This was a serendipitous error, as Rehn (1929) did not have a possibility to check what Serville (1838) described under the name H. lucifer. As we had the name-bearing specimens of both species before us, it became clear that H. taurus syn. nov. represents a synonym of H. lucifer. This confusion was furthermore inflamed by Skejo (Skejo and Caballero 2016;Skejo 2017) who applied the epithet 'taurus' to a dark species with angled spines. Now, with the description of this dark Holocerus species as H. devriesei sp. nov., we hope that this nomenclatural knot has been untangled and that both experts and citizen scientists will be able to correctly name Holocerus specimens. The IUCN Red List assessments (Danielczak et al. 2017a, b) should be amended accordingly. tion in the Natural History Museum in Madrid, not only for checking the specimens numerous times, but also for all her help and a great time; thanks to Laure Desutter, Simon Poulain, and Tony Robillard for providing the facitilies for our work in the MNHN collections in Paris and help with labeling type specimens; thanks to Jason D. Weintraub, a the curator in the Entomological collections of the ANSP Philadelphia, for photographing H. taurus type; thanks to non-anonymous reviewers (Josef Tumbrinck and Arne W. Lehmann) for their time, comments and suggestions; thanks to Michael Hendriksen and Dora Papković for proof reading. Open access was funded by the Heinrich -Heine Universität Düsseldorf: Open Access Publishing.