The Chimarra lehibemavo species-group , new and endemic to Madagascar ( Trichoptera , Philopotamidae )

Abstract. The Chimarra lehibemavo group is described to include thirteen new species: Chimarra lehibemavo sp. nov., C. cebegepi sp. nov., C. fenoevo sp. nov., C. forcellinii sp. nov., C. fotobohitra sp. nov., C. gattolliati sp. nov., C. gensonae sp. nov., C. jejyorum sp. nov., C. hamatra sp. nov., C. makiorum sp. nov., C. moramanga sp. nov., C. saha sp. nov. and C. tamara sp. nov. The adults are easily recognizable by their large size, yellow colour and the structure of the male genitalia. The membranous tergum IX and the absence of the mesal lobe of tergum X are observed in other lineages, but the strong asymmetrical deformation of the phallotheca is apomorphic. The group is monophyletic with unknown affinities, but a preliminary phylogenetic placement is suggested following genetic analysis of two specimens. With one exception, the species have restricted geographical distributions in Madagascar and inhabit rivers in eastern pristine rainforests.


Introduction
The genus Chimarra Stephens, 1829 is the largest genus in the Trichoptera (Morse 2015).It is almost cosmopolitan in distribution, only absent from the southernmost areas of South America, New Zealand and many oceanic islands.The Neotropical and Oriental regions harbour the main known diversity.For a long time ignored, the rich Malagasy fauna of Chimarra begins to be studied.Ten species are reported to date from the island, but already collected material has permitted to estimate a richness exceeding sixty species (Gibon et al. 1999).The morphological diversity offered by this material has revealed the presence of distinct lineages and this was confirmed by the DNA-based phylogenies published by Wahlberg & Johanson (2014) and Kjer et al. (2014).
The long-term objective of our studies was to contribute to the conservation of the Malagasy fauna, not only for emblematic, economic or with high touristic value species, but also taking into account little known taxa important for ecosystem functioning.For this purpose, history of settlement, endemism, vicariance and redundancy are key data, these studies are much more informative at infrageneric levels (Resh & Unzicker 1975).
The genus Chimarra is monophyletic and clearly characterized by morphological synapomorphies.Three of the four recognized subgenera are reported only from the New World (mainly the northern part of South America and the Antilles).All the Old World species belong to the nominal subgenus Chimarra, which secondarily invaded North America (Ross 1956;Blahnik 1998), an event dated by Wahlberg & Johanson (2014).A few large lineages were considered by Ross (1956) and redescribed by Blahnik et al. (2009Blahnik et al. ( , 2012)).These lineages illustrate the long-term history of the genus.The species group defined here should be included, thereafter, in the great historical lineages, that would require other approaches such as molecular studies or larval morphology.
I have recently described two species belonging to the minima species group and widely distributed in the open lowland landscapes of Madagascar (Gibon 2015).By contrast, the lehibemavo group includes numerous species which are, with the exception of Chimarra fenoevo sp.nov., restricted to the pristine rainforests and microendemic (sensu Gibon 2000 and Wilmé et al. 2006).The monophyly of the group is inferred from male genital structure, which includes an unique combination of characters and a remarkable apomorphy.No closely relative group could be identified.

Material and methods
The material was collected during the project "Biodiversité et Biotypologie des eaux continentales malgaches" jointly conducted by the ORSTOM and the CNRE (Antananarivo).The equipment, the sampling method and the study areas were described by Elouard & Gibon (2001).Unless otherwise stated, the material was collected by the author.Some field campaigns benefited from the logistic of the WWF; in those cases, the main results and detailed ecological descriptions of the regions were published (Andringitra: Goodman 1996;Andohahela: Goodman 1999;Marojejy: Goodman 2000).Specimens were captured using a portable light trap, which was composed of a black light and a gas lamp, and subsequently preserved in 75% ethanol.The male genitalia of some specimens were cleared in a solution of potassium hydroxide, studied under the microscope in cedar oil, and mounted on slides in Euparal ® .The holotypes, paratypes and other specimens are deposited in the CBGP (Montferrier).
The storage in ethanol (75%) at room temperature is not favourable to the long-term preservation of the DNA (K.Kjer, personal communication during the XV th International Symposium on Trichoptera), our colleague Gwenaelle Genson (INRA / CBGP) managed to obtain sequences from two specimens.The extraction of genomic DNA was performed on whole specimens, using "DNeasy 96 Blood & Tissue extraction (QIAGEN)", following standard protocol.A primer cocktail described by Cruaud et al. (2010) was used to sequence the standard bar-code sequence of the mitochondrial COI gene (cytochrome c oxidase I).Voucher sequenced specimens and their DNA are deposited in the collections of the CBGP.
The terminology used thereafter is that of Blahnik (1998).Schematic ecological profiles of the capture sites were established according to altitude, distance to the source and watershed, their construction, use and limitations were described and discussed by Randriamasimanana & Gibon (2001).The Chimarra lehibemavo group

Diagnosis
All the Old World species of the genus belong to the nominal sub-genus Chimarra Stephens, 1829, characterized by a curvature of the stem of Rs vein in the forewing.In some groups, as minuta and georgensis, this curvature is weakly marked.In the lehibemavo group, the stem of Rs vein is not curved but clearly angular (Fig. 1).This constitutes an accentuated state of the character.The lehibemavo group differs from the large digitata lineage by the absence of the median lobe of tergum X and the membranous tergum IX; it differs from the tsudai lineage by the undivided lateral lobes of tergum X.These lateral lobes are U-shaped or boomerang-shaped with one dorso-distal and one ventral branch.This is also observed in the georgensis species group but species of this latter group have a dorsally membranous tergum X and a reduced but sclerotized tergum IX.Above all, the phallic apparatus is the best characteristic of the group: the phallotheca is laterally split in two plates (thereafter named dorsal and ventral lamina); each of these plates is distorted in its own way, making the apparatus asymmetric.This phallotheca, associated with a variable internal sclerite and a large phallotremal sclerite, is a remarkable apomorphic character, the best way to identify the group.Each species exhibits specific forms of phallotheca and phallic sclerites, making the phallic apparatus the best diagnostic character.
In Madagascar, alive adults of the group are easily identifiable by the combination of the yellow colour and the large size.Most of the Malagasy Chimarra are black or brown, but the species of the minima and dybowskina groups are yellow or pale brown; the forewing length varies from 4 to 6 mm.In Africa, some species from mountain areas are quite large (for exemple 6.2 mm in Chimarra philipponi Gibon, 1986 from Guinea or 7.0 mm in Chimarra calidopectoris Wahlberg et al., 2014 from Malawi).The
Abdominal segment VIII complete, smaller than the VII, sternite VIII with a ventral keel.Sternite IX with a ventral keel and large antero-lateral lobes deeply invaginated in sternite VIII; in lateral view: latero-distal margin convex, sub triangular or rounded.Tergite IX membranous.Tergum X completely divided in two lateral lobes, median part membranous without noticeable median lobe.When viewed laterally, each lateral lobes v-or boomerang-shaped with one dorsal and one ventral branch.Preanal appendages small, simple and generally knoblike.Inferior appendages one-segmented, shape and size variable depending on the species; when viewed laterally, rectangular or triangular with small lobes along the caudal margin, most species having an additional lobe on the inner side (internal lobe, Fig. 4A,  C).Phallic apparatus comparatively large, comprising a bulbous base prolonged by a tubular phallotheca (structure sometimes described as pistol-shaped), remarkably long (longer than abdominal segments VIII, IX and X).Phallotheca sclerotized, distal part split, forming two conspicuous, asymmetrically enlarged or distorted extensions (dorsal and ventral lamina, Fig. 2D-E).Endotheca with: (1) a well developed phallotremal sclerite complex, (2) a large, variable in shape, internal sclerite, usually spiniform but sometimes globular or bullet-shaped.

Biomolecular data
The CO1 barcoding sequence is known for two specimens only, one male of C. fenoevo sp.nov.and one male of C. tamara sp.nov.The identification research in the database BOLD (Ratnasingham & Hebert 2007) indicates for C. fenoevo sp.nov. a similarity percentage of 95.54 with C. OF sp.MG9, an unidentified Malagasy species included in the phylogeny published by Kjer et al. (2014, supplementary material).For C. tamara sp.nov., the maximum percentage is only 83.31, but again with C. OF sp.MG9.In the electronic supplement file of Kjer et al. (2014), the nearest relative of C. OF sp.MG9 (in a COI based phylogeny) is C. orumbera Cartwright, 2002 from Australia.Chimarra orumbera is easily distinguished from any species of the lehibemavo group by the presence of the dorsal lobe of tergum X (although quite simple and poorly sclerotized), and the absence of distortion of the phallotheca.These data will be more useful for future study, when the afrotropical fauna will be better known.For the moment, I would not say that this preliminary result confirms or, at least, does not refute the conclusion of the morphological study: the lehibemavo group is not close to any other among the afrotropical species.which is stout and shorter than the phallotremal sclerite, whereas longer in Chimarra cebegepi sp.nov.

Etymology
From the Malagasy 'lehibe' ( = big) and 'mavo' ( = yellow), referring to the colour and large size of the species.
inferior appendageS.Roughly triangular, anterior margin length about twice the ventral margin length; hypotenuse sinuous; internal lobe small and thumb-shaped on lateral view, wide and sinuous on dorsal view.phallic apparaTuS.Little modified, ventral lamina separate and forming long digitiform plate (Fig. 2C-D), dorsal lamina wide and sub rectangular; internal sclerite short and stout, bifid lateral view; phallotremal sclerite with one central peaked element and two hook-shaped lateral elements.C. cebegepi sp.nov.differs from C. saha sp.nov.by the distal margin of the inferior appendage slightly concave and regularly rounded in lateral view (Fig. 3A).Moreover C. saha sp.nov. is characterized by the lateral deflection of the dorsal lamina of the phallotheca (Fig. 13E), which is more developed than in C. cebegepi sp.nov.(Fig. 3E).

Etymology
The name is designed from the acronym of the Centre de Biologie pour la Gestion des Populations.inferior appendageS.Dorsal branch elongate, regularly narrowing; ventral branch short; internal lobe constituted by six small indentations (Fig. 3C); distal margin concave and regularly rounded in lateral view (Fig. 3A).phallic apparaTuS.Phallotheca split from mid-length in two asymmetrical laminas of equivalent length; internal sclerite long (longer than the laminas of the phallotheca) with curved apex and, before midlength, short and triangular protuberance (Fig. 3D); phallotremal sclerite present with two lateral, hookshaped elements.(Fig. 5A).

Etymology
The name is that of the nearest city to the type locality.
inferior appendageS.Triangular, hypotenuse sinuous, with half a dozen of small lobes (lateral view), internal lobe large, very apparent on the dorsal view (Fig. 4C).phallic apparaTuS.Distal part of dorsal lamina widened and rounded, ventral lamina digitiform, slightly longer than dorsal lamina; internal sclerite short with a rough surface and a small lateral extension (Fig. 4D); phallotremal sclerite relatively small and thin.

Diagnosis
Chimarra forcellinii sp.nov. is a close relative of C. fenoevo sp.nov.These two species differ by the shape of the inferior appendages (lateral view), L-shaped in C. forcellinii sp.nov.(Fig. 5A), whereas triangular in C. fenoevo sp.nov.(Fig. 4A).Another useful diagnostic character is the maximum width of the phallotheca: that is at the apex in C. fenoevo sp.nov.(Fig. 4D), whereas in C. forcellinii sp.nov. it is in the middle of the tubular part (Fig. 5D).

Etymology
This species is dedicated to Maxence Forcellini in tribute to his work for the aquatic biomonitoring of La Réunion.
European Journal of Taxonomy 319: 1-31 (2017) Tergum X.Both dorsal and ventral branches elongated, ventral branch almost straight, dorsal branch slightly and regularly curved ventrad in lateral view, strongly angled inward in dorsal view (Fig. 5B).
inferior appendageS.L-shaped (lateral view), anterior branch twice as long as ventral one, distal margin sinuous.phallic apparaTuS.Ventral lamina shorter than dorsal lamina, digitiform (dorsal view); dorsal lamina with prominent median bulge and straight distal edge (Fig. 5D); internal sclerite strong and globular; phallotremal sclerite well developed including an indistinct dorsal element (Fig. 5E) and a broad plate ending in two hooks (Fig. 5D).

Distribution
Madagascar (endemic), Rianila River basin.where it is around 35° (Fig. 7A, 2A).Another remarkable character of this species is the strong median narrowing of the dorsal lamina (Fig. 6D), also observed in C. gensonae sp.nov., but with a different shape (Fig. 8E).

Etymology
The name is that of the river where the holotype was collected.
inferior appendageS.L-shaped, dorsal branch twice as long as ventral branch, internal lobe rounded and hump-shaped (Fig. 6C).phallic apparaTuS.Ventral lamina as long as dorsal lamina, but thinner on the lateral and dorsal views; dorsal lamina with a median narrowing (Fig. 6D); internal sclerite long, nearly as long as ventral lamina, with a little, lateral expansion; phallotremal sclerite with a basal globular part and a distal pair of hooks.

Diagnosis
Chimarra gattolliati sp.nov. is most closely related to C. jejyorum sp.nov.In both species, the lateral lobes of tergum X are short, the dorsal branches are small and thumb-shaped.The angle between dorsal and ventral branch is more acute in C. gattolliati sp.nov.(Fig. 7A) than in C. jejyorum sp.nov.(Fig. 9A).But this is not easy to observe, the two species are more quickly distinguished by the inferior appendages or by the phallic apparatus.In C. gattolliati sp.nov., the ventral branch of the inferior appendage is distally protruding, the dorsal branch is thin and arched (Fig. 7A); in C. jejyorum sp.nov., the ventral branch is not clearly apparent and the dorsal branch is wide and only curved towards the apex (Fig. 9A).The phallothecae of both species are split, in C. gattolliati sp.nov., this splitting occurs after mid-length (Fig. 7D), whereas, in C. jejyorum sp.nov., it occurs somewhat after the base (Fig. 9D).

Etymology
This species is dedicated to Jean-Luc Gattolliat in tribute to his work on the Malagasy may flies.

Description
Size.Forewing 6.6 mm, hind wing 5.5 mm.Tergum X. Lateral lobe with wide basal part, short ventral branch and barely longer dorsal branch; when viewed dorsally, dorsal branch angled with, towards the apex, a small protuberance bearing the sensillae.
inferior appendageS.L-shaped (lateral view); ventral branch distally protruding; dorsal branch thin and regularly arched.European Journal of Taxonomy 319: 1-31 (2017) phallic apparaTuS.Phallotheca split after mid-length; ventral lamina longer than dorsal lamina, dorsal lamina with a lateral fold (Fig. 7E) and an apical curvature (Fig. 7D); phallotremal sclerite with basal part flanked by two indistinct small needles and extended by a pair of hooks; internal sclerite long and stout, with apical curvature, slightly shorter than phallotremal sclerite.

Etymology
It is a pleasure to dedicate this species to my colleague Gwenaelle Genson in recognition of her efforts for the barcoding of caddisflies.

Description
Size.Forewing 7.6 mm, hind wing 6.0 mm.Tergum X. Dorsal branch wide and short, thumb-shaped when viewed laterally, toward the apex a small protuberance with two sensillae; ventral branch thin as long as dorsal branch; dorsal and ventral branches forming an angle of 65°.
inferior appendageS.Roughly triangular, anterior margin almost twice as long as ventral margin; dorsal part curved distad, inner lobe long, thin, slightly protruding.phallic apparaTuS.Short compared to other species of the group.Phallotheca split nearly from the base; ventral lamina as long as dorsal lamina, thinner and digitiform; when viewed dorsally, dorsal lamina with a lateral fold and a sudden narrowing after mid-length (Fig. 8D); internal sclerite short, stout, GIBON F.-M., Chimarra lehibemavo group from Madagascar bullet-shaped (Fig. 8E); phallotremal sclerite including an ovoid basal part, a thin spine and a pair of hooks.

Etymology
From the Malagasy 'jejy' (a traditional musical instrument).It is a reference to the name of the Mountain Marojejy, in translation: where there are many guitars.
inferior appendageS.Ventral branch reduced, dorsal apex with small denticles.phallic apparaTuS.Phallotheca split according to a sagittal plane, right lamina large with a big dorsolateral bump, left lamina nearly as long but thinner (dorsal view); internal sclerite sickle-shaped (lateral view); phallotremal sclerite present, sclerotized but indistinctly shaped.GIBON F.-M., Chimarra lehibemavo group from Madagascar these appendages have a distinct and heavily sclerotized supplementary branch on the distal edge.This branch is inserted at mid-height in C. hamatra sp.nov.(Fig. 10A), whereas almost apically in C. makiorum sp.nov.(Fig. 11A).

Etymology
The name is a Malagasy adjective that means 'uncommon'.

Description
Size.Forewing 7.1 mm, hind wing 5.6 mm.Tergum X. Lateral lobe reduced; dorsal branch straight (lateral view), relatively short with, at midlength, a small lateral bump bearing two sensillae; ventral branch short and thin.inferior appendageS.Wide, roughly rectangular in lateral view; ventral branch weakly developed; median branch protruding, inserted on distal edge (lateral view), making the apex claw-shaped on dorsal view (Fig. 10C).phallic apparaTuS.Long, distal half of the phallotheca poorly sclerotized with the exception of a strip terminated by a triangular tip; internal sclerite long, with a basal triangular extension and curved apex; phallotremal sclerite relatively small, ending in a pair of hooks.

Diagnosis
Chimarra makiorum sp.nov.differs from C. hamatra sp.nov.by the apical insertion of the median branch of the inferior appendage (at mid-height in C. makiorum sp.nov.).

Etymology
The name is designed from that of the river where the type was collected (Rivière des makis).It is used according to the Latin rules.

Diagnosis
Chimarra moramanga sp.nov.differs from the other species of the lehibemavo group by the strongly bent base of the tubular part of the phallic apparatus (Fig. 12D).

Etymology
The name is that of the nearest city to the type locality.

Description
Size.Forewing 6.4 mm, hind wing 5.0 mm.Tergum X. Dorsal branch of the lateral lobe long (twice as long as the ventral branch), basal part straight, proximal part gently curved dorsad (lateral view), a small bump with two sensillae just before the curvature.
inferior appendage.Triangular with long anterior margin and short ventral margin, dorsal apex slightly curved distad, inner lobe large and rounded (dorsal view).phallic apparaTuS.Phallotheca strongly angled at base, then gently bent ventrad, distal part (one third of the length) poorly sclerotized with the exception of a dorsal strip with triangular apex (lateral view, Fig. 12D); internal sclerite as long as the modified distal part of the phallotheca, with enlarged base and curved apex; phallotremal sclerite shorter than the internal sclerite.

Diagnosis
By most of its characters, C. saha sp.nov. is closely related to C. moramanga sp.nov.The remarkable structure of the phallic apparatus of this latter prevents any confusion between the two species.Moreover, C. saha sp.nov. is characterized by the laterally deflected dorsal lamina of the phallotheca.This deflection makes the distal part of the apparatus wider than the median part (dorsal view).

Etymology
From the Malagasy 'saha' ( = valley), referring to the name of the type-river, used as a noun in apposition.

Description
Size.Forewing 7.5 mm, hind wing 5.9 mm.Tergum X. Dorsal branch of the lateral lobe long (twice as long as the ventral branch), bent ventrad at mid-length with a dorsad curved apex (lateral view).
inferior appendage.Triangular with long anterior margin, short ventral margin and a small triangular point on distal edge, dorsal apex slightly curved distad, inner lobe large and rounded (dorsal view).

Etymology
The name is that of the Tamara River, one of those where the species was collected.

Description
Size.Forewing 7.3 mm, hind wing 5.9 mm.Tergum X. Ventral branch of lateral lobe small, dorsal branch in direct continuity from the base; when viewed dorsally, apex slightly enlarged, bearing a small protuberance with two sensillae.inferior appendage.Dorsal apex curved distad and protruding ventral part (lateral view), inner lobe long forming a serrated big bump on the dorsal view (Fig. 14C).phallic apparaTuS.Phallotheca slim and long, split after mid-length; ventral lamina twisted with acute apex, slightly longer than dorsal lamina; phallotremal and internal sclerites relatively small; internal sclerite slightly longer than phallotremal sclerite, apex curved.

Geographic Data
Geographic and ecological information on the lehibemavo group are summarized on figures 15 to 18.With one exception discussed thereafter (C.fenoevo sp.nov.), the species live in small streams of the oriental rainforests, where they have narrow latitudinal distributions.From the north to the south of the Island, appears a sequence of species or small groups of species:     Such an eco-geographic pattern was already observed, described and discussed for the Rossinae and for the Philopotaminae (Gibon & Elouard 1996;Gibon 2013Gibon , 2014)).Compared to the genera Rossodes Oezdikmen & Darilmaz, 2008, Wormaldia McLachlan, 1865and Ranarijaodes Gibon, 2014, the lehibemavo group offers a less extreme situation.One species, C. tamara sp.nov., was captured in two adjacent areas (Andringitra and Ranomafana).Moreover, where it can be observed, the altitudinal distributions of the species are wider.In the Marojejy National Park, C. cebegepi sp.nov.was collected from 400 to 700 m a.s.l., C. jejyorum sp.nov.from 400 to 1200 m a.s.l. and C. lehibemavo sp.nov.from 400 to 1600 m a.s.l.In the same area, the six species of Wormaldia were strictly restricted to one altitudinal zone (Gibon 2014, table 1).C. fotobohitra sp.nov.was collected outside the large area of pristine forest, but on a small stream coming from a forest remnant and is possibly a relic species.
Finally, one species is widely distributed, C. fenoevo sp.nov.was collected from the Marojejy in the north to Andohahela in the South, from 70 to 1700 m a.s.l.It was also recorded from the western slope, on headwaters of the Mangoky, Betsiboka and Tsiribihina Rivers.This geographical exception coincides with a different ecological profile.Present on small tributaries, C. fenoevo sp.nov.colonizes also broader rivers, further away from the sources, without, however, being a true potamic species (it has never been captured on true large rivers).Another characteristic of this species, compared with other members of the group, is the vegetation of its capture sites.Precisely, the vegetal landscape of the streams (the terrestrial biome sensu Ross 1963) and not the riparian vegetation.Chimarra fenoevo sp.nov. is the only species of the group that is not associated with the evergreen rainforest.On the eastern slope, it was mainly encountered in the forest transition or in secondary forests; in the Central Highlands, in vestigial or relict gallery forests, above rice field areas.

Discussion
Species distributions in the lehibemavo group are similar to those described for the other subfamilies, Philopotaminae and Rossinae (Gibon 2013(Gibon , 2014)), very different from that described for the minima group (Gibon 2015).We can assume that the widely distributed Malagasy species of the minima group, morphologically close to their continental relatives, have recently colonized the Island.In comparison, the numerous microendemic species of the C. lehibemavo species group possibly belong to a much older fauna.Validation of this hypothesis would require genetic research such as that undertaken by Monaghan et al. (2005), who have demonstrated that the Malagasy fauna of Baetidae (Ephemeroptera) is combining archaic and invasive lineages.