Taxonomy and historical biogeography of the genus Rhyacophila PICTET , 1834 on the Philippines ( Insecta , Trichoptera ) With 30 figures and 3 tables WOLFRAM MEY

Die Gattung Rhyacophila ist auf den Philippinen mit 9 Arten vertreten. Davon werden hier fünf neue Arten beschrieben: R. baldensis sp. n., R. banahawmontis sp. n., R. amoyanensis sp. n., R. pilifera sp. n., und R. cordilliera sp. n. Die Genitalapparate der Männchen und die Imagines von vier Arten werden abgebildet. Die Arten verteilen sich auf die castanea-Gruppe und die neu aufgestellte spinosellata-Gruppe. Zwei Punktkarten zeigen die Verbreitung der Gruppen auf den Philippinen. Die Verwandtschaftsverhältnisse werden in drei Kladogrammen dargestellt, die auf der Analyse genitalmorphologischer Merkmale beruhen. Die Arten der castanea-Gruppe sind über Borneo auf die Philippinen eingewandert. Die spinosellata-Gruppe ist endemisch auf den Philippinen. Ihre Arten werden als Reliktendemiten angesehen, da die nächsten Verwandten in Ostasien und Japan vorkommen, und die Artengruppe ansonsten in ganz Südostasien fehlt. Die möglichen Ausbreitungsmechanismen werden kurz diskutiert.


Introduction
The Rhyacophilidae and its large genus Rhyacophila PICTET, 1834 represent one of the most ancestral branches within Trichoptera.The group has retained many primitive characters especially in wing venation and genital architecture.According to the fossil record the genus is at least Eocene in age, and probably much older (ULMER 1912, ULMER 1912, ULMER MEY 1988, MEY 1988, MEY WICHARD et al. 2009).Since most of the Philippine Islands originated only in the Miocene (HALL 1996) the existence of the genus HALL 1996) the existence of the genus HALL predates the formation of this archipelago.The presently known Rhyacophila species on the islands Rhyacophila species on the islands Rhyacophila are the result of colonisation processes which started as soon as the islands appeared from the sea.The colonisation history, however, seems to be very complex.It is closely connected with the complicated and puzzling development of maritime South East Asia, including Taiwan, Sundaland and Sulawesi.The Trichoptera fauna of the past was surely as dynamic as it is today.The presentday pattern of its composition and distribution in the area is the result of dispersal, vicariance and subsequent evolution, with dispersal as the principal process in an oceanic environment (COWIE & HOLLAND 2006).These processes are more or less imprinted in the distribution pattern in the archipelago, and their analyses could possibly enable the reconstruction of their colonisation history.The genus Rhyacophila appears to be a promising candidate for undertaking such an analysis.The quality of the study depends on the available data.A comprehensive inventory and an advanced level of faunistic research could provide the necessary facts.Both information sources are, however, weakly developed in the Philippines.Present knowledge is inadaquate, and we know only parts of the fauna, which are but large enough to recognise the presence of a diverse fauna with a high percentage of endemism.This poor faunistic and taxonomic basis prompted the author to perform a series of field trips to the Philippines from 1994 to 2001.Samples of caddisflies were taken on all major islands.This material forms the principal basis of the present study.In the 1990s the environmental devastation of large stretches of rainforests and rivers on the Philippines was already at an advanced stage.Many visited areas were devoid of suitable habitats for stenoecious caddisflies and already possessed an impoverished fauna.An unknown number of Rhyacophila populations and probably even endemic species must surely have gone extinct.By Rhyacophila populations and probably even endemic species must surely have gone extinct.By Rhyacophila now it is no longer possible to achieve a substantial survey of the original distribution of species and communities.The fauna is much reduced and consequently the collected data, although considerable, cannot be significantly increased.Any faunistic or biogeographic research in the Philippines today is actually the study of a fragment out of a fragment.Devastation processes go on in all parts of the country (BROAD & CAVANAGH 1993, HEANEY & HEANEY & HEANEY REGALADO 1998).Nearly 20 years ago WILSON (1992) wrote: "Today the Philippine fauna stands on the eve of comprehensive destruction caused by deforestation and the poverty of an increasing human population."The eve is over.The loss of Philippine diversity is now a reality.Nevertheless, the study of the fauna remains an urgent task, at least to recover the information that still can be saved.

Collecting:
Adult caddisflies were collected using light.Usually, a light-tower (with 2 superactinic tubes, 12 W) was placed in the vicinity of rivers, and the arriving caddisflies were picked up from the gauze sheet that surrounds the tubes.Collecting usually started at sunset and was finished after 3-4 hours.The collected specimens were partly put in 75 % alcohol and partially pinned the next day.During daytime, specimens were captured using a hand-net in the bank vegetation.

Genital preparations:
The abdomen was boiled in KOH for some minutes to macerate the muscles, body fat and internal, unsclerotised structures.The abdomens were stained with Chlorazol Black after which the preparation was washed in 70 % ethanol.At this point the abdomen was studied in detail.It was turned in all directions in order to understand the shape and position of its individual components, and all the relevant abdominal and genital structures were now drawn in dorsal, ventral, lateral and/or caudal view.A drawing tube attachment to the binocular microscope (LEICA, Wild M8) was used.After study the genitalia were deposited in polyethylene genitalia vials (BIOQUIP Products) filled with glycerin.The vials were attached to the pin.

Deposition:
Most of the material was preliminarily stored in the Museum für Naturkunde Berlin.Individuals of all species, including paratypes, are to be deposited later in a museum in the Philippines.

Derivatio nominis:
The species is named after the Bald Mountains, south of the Central Cordilliera, where R. baldensis sp.n. was collected the first time.Length of forewing 6 -6.5 mm, wing-span 13 -14 mm; head and thorax dark brown; antenna brown, the basal third yellow-brown; ocelli white; forewings dark brown with numerous small white spots scattered over the whole forewing; basal parts of veins with erected, black bristles; fork 4 in forewing as long or longer than its stalk; hindwings pale brown; abdomen black dorsally; abdominal segment VII with short sternal process.Male genitalia (Figs 1-3): Segment IX forms a complete ring with tergal part large and sternal part much narrower and constricted on the latero-ventral sides; segment X narrow and compact arranged in dorso-lateral position, dorsal side elevated, rectangular in dorsal view, caudal side with broad tooth directed ventrad; anal sclerites large, curved ventrad and with short root; apical band of segment X short, connected basally with the long tergal strap produced distally in short protruding process; coxopodite of inferior appendages angulated ventrally, with long and slender basal apodemes articulating with dorsal base of the phallotheca; ventral lobe of harpago elongate; phallic apparatus long, aedeagal complex with short dorsal process and broad, membranous ventral plate; ventral plate shortly excised apically, small and slender lateral lobes present; parameres long, crossing each other and bearing numerous bristles on the clubbed ends.

Derivatio nominis:
The specific epithet is derived from the name of the terra typica, the lower elevations of Mount Banahaw in southern Luzon.
Length of forewing 5.5 -6 mm, wing-span 11 -12 mm; head and thorax dark brown; antenna brown, the basal third yellow-brown; ocelli white; forewings dark brown with numerous small white spots scattered over the whole forewing, especially dense in costal area; basal parts of veins with erected, black bristles; fork 4 in forewing as long or longer than its stalk; hindwings pale brown; abdominal segment VII with short sternal process.Male genitalia (Figs 4-7): Segment IX complete ring with tergal part large and sternal part much narrower, constricted on the latero-ventral sides; segment X narrow and compact arranged in dorso-lateral position, dorsal side elevated, but flat and oval in dorsal view, caudal side without lateral teeth; anal sclerites moderately large, curved ventrad and with short root; apical band of segment X inconspicuous, connected basally with the long tergal strap which is produced distally in a short, freely protruding process; coxopodite of inferior appendages angulated ventrally and with long and slender basal apodemes articulating with dorsal base of phallotheca; ventral lobe of harpago elongate, rounded apically; phallic apparatus long, aedeagal complex with broad dorsal process and wide, membranous ventral plate; ventral plate shortly up-curved apically, lateral lobes absent; parameres long, crossing each other and bearing 6-8 bristles on the clubbed tips.

Remarks:
Rhyacophila banahawmontis sp.n. is a member of the Rhyacophila banahawmontis sp.n. is a member of the Rhyacophila banahawmontis castanea-group.It is obviously related to R. negrosana MEY, 1998.Both species lack the lateral lobes of the ventral plate of the aedeagal process.

Rhyacophila spinosellata -group
The group is established here to accommodate the following three new species and two previously known species.The latter species were assigned by MEY (1999) to a basal position within MEY (1999) to a basal position within MEY the yosiiana-group.The study of the new species revealed two synapomorphic characters (rectangular, plate-like dorsal process of phallotheca and large, un-rooted anal sclerites) which exclude the species from the yosiiana-group and necessitate the establishment of a group in its own.All included species are endemic to the Philippines.

Derivatio nominis:
The specific epithet is derived from the name of the small river, where the new species was collected.Length of forewing 5.5 mm, wing-span 12 mm; head and thorax dark brown; frons and setal warts paler, vestiture black; ocelli white; forewings brown with darker pterostigma; legs light brown, spurs black; abdominal segment VII with short sternal process. Male

Derivatio nominis:
The specific epithet is composed of the Latin words "pilus" (hair) and "fer" (carry) referring to the tuft of long hairs on the harpago.both indistinctly separated by a short suture on the latero-ventral side; segment X arranged in dorso-lateral position, dorsal portion compact with short medial process in dorsal view, lateral flanges with small tooth; anal sclerites moderately large, separated and without root; apical band of segment X separated basally and articulating only with paired tendons of the phallotheca; tergal strap absent; dorsal process of phallotheca produced distally in a broad, freely protruding process; coxopodite of inferior appendages with slightly concave ventral margin and with slender basal apodemes articulating with lateral suture of segment IX; ventral lobe of harpago elongate, with a filiform appendage originating from medial base and running in distal direction, its apex with 2-3 long bristles; dorsal lobe short, with a conspicuous tuft of long hairs on inner side; phallic apparatus short, aedeagus a simple, slender tube without dorsal or ventral plates; parameres short, bearing 8-10 black bristles.

Derivatio nominis:
The species is named after its range in the large mountain chain in northern Luzon which is known as Central Cordillera.Length of forewing 7 -7.5 mm, wing-span 15 -16 mm; head and antennae dark brown; setal warts pale brown; thorax black on dorsal side; forewings brown with darker pterostigma; fork 4 in forewing shorter than its stalk; in anal field of hindwing crossvein Cu2-A1 oblique; legs light brown, spurs black; abdominal segment VII with short sternal process.Male genitalia (Figs 18-20): Segment IX forms complete ring with tergal part large and sternal part much narrower, both indistinctly separated by short suture on the latero-ventral side; segment X arranged in dorso-lateral position, dorsal portion divided into a pair of bulbous parts, each with short, ventrally directed tooth; lateral flanges thin; anal sclerites large, separated and without root; apical band of segment X very short, separated basally and articulating only with paired tendons of phallotheca; tergal strap absent; dorsal process of phallotheca produced distally in broad, freely protruding process; coxopodite of inferior appendages with slightly concave ventral margin, a median flat lobe and with slender basal apodemes articulating with lateral suture of segment IX; lobes of harpago rounded, medial sides each with a patch of short, black spines; filiform appendage of harpago originating from medial base and running in medio-ventral direction, its apex with 2 short spines; phallic apparatus short, aedeagus a simple, slender tube without dorsal or ventral plates but with broader base; parameres short, bearing 6-8 black bristles.

Remarks:
R. cordilliera sp.n. belongs to the R. cordilliera sp.n. belongs to the R. cordilliera spinosellata group.It differs from all other members of the group spinosellata group.It differs from all other members of the group spinosellata by the large and compact harpago in the male genitalia with two patches of strong, short spines on the inner side.It is related to the preceding species by the common presence of harpagal appendages on the ventral base.

2. Phylogeny
The Philippines are a very diverse archipelago in terms of vegetation, climate, island sizes and topograpy.The great variety of available habitats is an ideal situation for arriving immigrants to survive and to settle permanently.The climatic cycles of the last 2 Mio years resulted in a repeated change of vegetation, shifting of vegetation zones and oscillations of island sizes.With each cycle species dispersed, extended and reduced their ranges or became isolated.These processes triggered speciation events and resulted in the formation of a wide spectrum of descendants ranging from simple species pairs to a multiplication of species.In Trichoptera the hamifera -group of hamifera -group of hamifera the genus Hydropsyche PICTET is an example for the evolution of a large species swarm with more than 20 species distributed over the entire archipelago (MEY 2003).In contrast to MEY 2003).In contrast to MEY Hydropsyche, species of Rhyacophila are ecologically more restricted, more localised and usually occur in smaller Rhyacophila are ecologically more restricted, more localised and usually occur in smaller Rhyacophila numbers.Although more species doubtlessly occur in the Philippines, but escaped collecting so far, the genus has not developed an adaptive radiation like Hydropsyche, with many species occurring in the same river system.Small headwater streams and spring areas remain the habitats of Rhyacophila species as on Borneo.Nevertheless, Rhyacophila species as on Borneo.Nevertheless, Rhyacophila Rhyacophila exhibits a radiation, but of Rhyacophila exhibits a radiation, but of Rhyacophila a different type.This radiation type is characterised by a pool of related species, each of them restricted or endemic to a single island.Congeneric species may occur sympatrically but in these cases they belong to different species groups.It is an open question, whether the species live only on the larger islands or also on smaller ones This would determine the size of the radiation.The Philippine species are distributed between two species groups.These are not closely related (ROSS 1956, SCHMID 1970)and are treated here separately: The castanea-group contains four species which occur on three islands.The morphological analysis revealed that they are descendants of a single ancestor (Tab. 1, Fig. 25) that obviously immigrated from Borneo.The spinosellata-group, established as new earlier in this paper, contains six species on five islands.In contrast to the castanea group it can be divided into three clades, which are morphologically castanea it can be divided into three clades, which are morphologically castanea well differentiated (Tab.2, Fig. 26).Their ranges in the Philippines do not overlap.This supports the idea that the origin of these species is based on an old, perhaps Miocene colonisation and subsequent diversification.Alternatively, two or three independent colonisation processes could have caused the observed pattern.According to male genital characters the group is related to the yosiiana-and betteni-groups (Tab.3, Fig. 27).The distribution of the former encompasses the Japanese Islands and the adjacent continental area (Fig. 31).The range of the latter group is confined to western North America.

Biogeography
Biogeographical research in the Philippines has to consider the biology and the ecological requirements of the studied taxa.On this basis two questions have to be answered: Where did the species come from or where did the group originate, and how did they manage to cross the sea?
Rhyacophila is a large genus encompassing today more than 550 species.It has a predominantly Rhyacophila is a large genus encompassing today more than 550 species.It has a predominantly Rhyacophila Holarctic distribution.The Himalayan Region is home of a large number of species and species groups and can be considered their diversity centre (SCHMID 1989).From this region the numbers are declining towards South East Asia and Sundaland.Beyond Sulawesi, the genus is not known to occur (MEY 1999).
MEY 1999).MEY The aquatic larvae of nearly all known Rhyacophila species are free-living predators in running Rhyacophila species are free-living predators in running Rhyacophila waters.They have a clear preference for temperate and cold streams.Accordingly, the majority of species dwell in the headwaters of rivers in mountain ranges from low to highest elevations.A few species, those of the Philippine included, have adapted to live in springs or spring areas.In the Philippines this habitat-type is not restricted to mountains but occurs in lowlands too.personal observations the adults are weak flyers and only fly for short distances.The body is soft and not as well sclerotised as in temperate species.Thus, the imagines are prone of desiccation.When stored in glass vials they can survive half a day.The dispersal abilities of the adults appear to be poor.They cannot survive in open and sunny places and remain close to the waters where their larvae live.Within the forest, dispersal flights may happen along the streams and its ramifications, and at higher elevations between different water catchment areas.Mountain ridges do not represent serious barriers.The probability of successful migration between isolated forest patches separated by cultivated land, grassland or savanna is expected to be rather low.Adults of aquatic insects have different flight behaviours and dispersal abilities, but it is generally limited (KOVATS et al. 1996).Active dispersal of flying Rhyacophila adults over sea straits to reach distant islands can be largely excluded.
Rhyacophila adults over sea straits to reach distant islands can be largely excluded.Rhyacophila It is still a matter of debate, whether the small continental terranes (e. g. the Zamboanga Peninsula, Mindoro Island) were permanently sub-aerial or not (TAN et al. 1988, DE JONG & TREADAWAY 1993, TREADAWAY 1993, TREADAWAY HALL 1996).They are considered here as oceanic islands.All colonisation of the Philippine Islands HALL 1996).They are considered here as oceanic islands.All colonisation of the Philippine Islands HALL except the Palawan and Busuanga Islands, which lies on the Sunda-Shelf, must have occurred by over water dispersal in a kind of passive transport.The often discussed rafts of tangled vegetation, carried to the sea by rivers in flood and drifting away to other islands, was a mechanism that could also have carried caddisflies.Floating islands of vegetation are often invoked as mechanism that permitted the dispersal of larger animals including mammals (WALLACE 1880, SMITH 2001).During the Quarternary the climate was characterised by cycles of humid and dry periods.Humid periods with more extensive rainforest correlated with periods of high sea level and wider sea gaps.In dry periods, when sea gaps were at their narrowest, rainforests shrank to smaller patches and vegetation of the exposed Sunda shelf was dominated by savanna and grassland (OOSTERZEE 1997, MORLEY 2000).The existing land bridges were of great significance for dispersal for MORLEY 2000).The then existing land bridges were of great significance for dispersal for MORLEY savanna rather than for rainforest species.However, dense forest must have been present along streams and could have served as dispersal corridors.Rhyacophila species of lowland streams and Rhyacophila species of lowland streams and Rhyacophila springs could surely survive and disperse in this ecosystem.The expanded mega-islands, which appeared at times of low sea level, allowed dispersion to new areas and mountain ranges that later on became islands again.They also shortened the gaps between islands considerably and formed landbridge-like corridors along the modern Sulu Archipelago, bridging Borneo and Mindanao (HEANEY 1985).But even at times of lowest sea level it remained a chain of neighbouring islands.HEANEY 1985).But even at times of lowest sea level it remained a chain of neighbouring islands.HEANEY The final crossing of narrow sea straits or sea surfaces probably happened by passive air transport.In the course of tropical thunderstorms and typhoons insects can be lifted to high altitudes, and are carried away by air currents.They are capable of transporting insect specimens over long distances.Each year in the summer months typhoons coming from the Caroline Basin of the West Pacific and running along the west coast of Mindanao and Luzon towards Taiwan or the Asian continent are recurrent events that affect, more or less strongly, the entire Philippine Islands.They have the potential to be important drivers of insect dispersal by passive transport and probably responsible for most of the colonisation processes that happen to occur in all groups of insects, including Trichoptera and Rhyacophila species.
Rhyacophila species.Rhyacophila Rhyacophila is represented in the Philippines by only two species groups, the Rhyacophila is represented in the Philippines by only two species groups, the Rhyacophila castanea -and castanea -and castanea spinosellata groups, whereas four and five species groups including the spinosellata groups, whereas four and five species groups including the spinosellata castanea-group occur on Borneo and Sundaland respectively (MEY 1999(MEY , 2010)).With nine species the MEY 1999MEY , 2010)).With nine species the MEY curvata-group is the largest group in Borneo.Interestingly, the group has not been reported from the Philippines.Although there are some curvata -group species, dwelling in lowland streams in northern Borneo curvata -group species, dwelling in lowland streams in northern Borneo curvata (e.g.R. tawauensis MEY, 2010) and therefore capable of dispersing in lowland areas, no species of the group has apparently arrived in the Philippines.It would not come as a surprise, were a member of the group to be eventually found in Mindanao.
The castanea group has an extended range in South East Asia (Fig. 30) with many species but only castanea group has an extended range in South East Asia (Fig. 30) with many species but only castanea one species, R. argentipunctella KIMMINS, 1955, is known so far in Borneo.By contrast four species are described from the Philippines.They are closely related but live on different islands (Fig. 28).The ancestor of the Philippine species was a close relative of a group with R. argentipunctella as the only representative in Borneo.In the Pleistocene land-bridges developed and connected Borneo with Mindanao (HEANEY 1985).Having arrived in Mindanao, probably along the Sulu HEANEY 1985).Having arrived in Mindanao, probably along the Sulu HEANEY Archipelago, the species was able to colonise southern Luzon and spread to northern Luzon.Even the Visayan Islands were reached with Negros as the hitherto only with a species of this group (R. negrosana MEY, 1998).Occurrence on other Visayan Islands like Panay or Cebu is very probable (Fig. 28).

R. bettenei
The spinosellata group consists of six spinosellata group consists of six spinosellata species.The first described species of the group were assigned to the yosiiana -group.Interestingly, the group iana -group.Interestingly, the group iana has no close relatives on Borneo, Sundaland and in South East Asia (cf.MALICKY 2010) (Fig. 31).Rhyacophila brechlini MEY, 1998, a Vietna mese species, was originally placed in the yosiiana group.yosiiana group.Re-yosiiana examination of the holotype revealed that it was misplaced in the group.In fact the species exhibit the primitive coupling of coxopodite and phallotheca by apodemes and the presence of a long tergal strap.The phallotheca is not firmly connected with the ventral side of segment X -the synapomorphy of the spinosellata-, yosiiana-and betteni groups.
betteni groups.betteni The species is probably better placed in the scissa group.The origin of scissa group.The origin of scissa the spinosellata group appears to spinosellata group appears to spinosellata be somewhat enigmatic.A closer examination of the genital morphology revealed that the oldest evolutionary line is represented by R. amoyanensis sp.n., discovered on Palawan (Fig. 29).From a geological viewpoint Palawan does R. amoyanensis sp.n., discovered on Palawan (Fig. 29).From a geological viewpoint Palawan does R. amoyanensis not belong to the Philippines, but is situated on the Sunda Shelf close to Borneo (HALL 1996).At the HALL 1996).At the HALL height of the Pleistocene Palawan, including the Busuanga Islands further north, formed an elongate extension on the north-east margin of the fully exposed Sundaland shelf.Though in a peripheral position, it received the full spectrum of the Sundaland fauna (OOSTERZEE 1997, HEANEY et al. 1998). HEANEY et al. 1998).HEANEY R. amoyanensis sp.n. can be regarded as a Sundaland species.It is perhaps the surviving descendant R. amoyanensis sp.n. can be regarded as a Sundaland species.It is perhaps the surviving descendant R. amoyanensis of a more widespread species whose range extended through Borneo.Future fieldwork should clarify whether or not it is the only survivor.This ancestral Bornean species was able to spread over the exposed Sulu Archipelago and immigrated into Mindanao in accordance with the scenario described for the castanea group.Subsequent dispersion and speciation gave raise to different clades on the castanea group.Subsequent dispersion and speciation gave raise to different clades on the castanea northern islands (Luzon, Mindoro) and on the southern islands (Negros, Mindanao) (Fig. 29).Alternative scenarios of immigration pathways into the Philippines were discussed by MEY (1999).MEY (1999).MEY They are based on a presumed, extended range of the yosiiana group towards southern China.From yosiiana group towards southern China.From yosiiana there, the route of a pre-Pleistocene ancestor into the Philippines should have run via Taiwan or was mediated by Mindoro, a continental terrane shifted with the opening South China Sea to its present position on the west side of the Philippines.Supporting arguments for these hypotheses might be provided by the discovery of hitherto undetected ancestral species still occurring in southern China and Taiwan.Up to now, such species have not been found.The discovery of three new species in the Philippines does not render these hypotheses less plausible.In fact, all proposed hypothesis try to explain an apparently extinction-mediated distribution pattern.Support for any scenario can be provided only if the extinction was not complete and surviving species still can be found.In any case, the range of the spinosellata group is that of a relict group separated by wide disjunc-spinosellata group is that of a relict group separated by wide disjunc-spinosellata tions from related groups.It is restricted to the Philippine Islands, and this makes the Philippines  3).spinosellata group (numbers refer to characters in table 3).spinosellata  an archipelago of refuge.There are probably additional groups in Trichoptera, and in other insect orders, for which the Philippines play the same role as a survival area.The search for additional relict groups is an exciting venture.The study of their phylogeny and biogeographic history might result in new arguments that can new light on the radiation of Rhyacophila and the herein presented Rhyacophila and the herein presented Rhyacophila attempt to find a plausible explanation.

Fig. 26 :
Fig. 26: Cladogram of presumed phylogenetic relationships of the spinosellata group (numbers refer to characters in table3).spinosellata group (numbers refer to characters in table3).spinosellata

Fig. 25 :
Fig. 25: Cladogram of phylogenetic relationships within the spinosellata group in the Philippines (numbers refer to characters spinosellata group in the Philippines (numbers refer to characters spinosellata in table 2).

Fig. 24 :
Fig. 24: Cladogram of phylogenetic relationships within the castanea group in the Philippines (numbers refer to characters in table 1).
Short description of relevant, male genital character states of the R. castanea group in the castanea group in the castanea Philippines.