Systematics and Biodiversity 5 (2): 187–198 doi:10.1017/S1477200006002283 Printed in the United Kingdom Klaas-Douwe B. Dijkstra1∗ , Viola Clausnitzer2 & Andreas Martens3 1 National Museum of Natural History Naturalis, P.O.Box 9517, NL-2300 RA Leiden, Netherlands 2 Graefestraße 17, D-06110 Halle, Germany 3 University of Education, Bismarckstraße 10, D-76133 Karlsruhe, Germany submitted June 2005 accepted July 2006 Issued 25 May 2007
C The Natural History Museum Tropical African Platycnemis damselflies (Odonata: Platycnemididae) and the biogeographical significance of a new species from Pemba Island, Tanzania Abstract The damselfly, Platycnemis pembipes sp. nov., is described from Pemba Island (Ngezi Forest, Tanzania) and its affinities with Guineo-Congolian and Malagasy congeners are examined. For this purpose the identity and distribution of Afrotropical Platycnemis is reviewed, especially the taxonomically confused continental species. The Pemba species is nearly identical to some species of the Malagasy radiation of Platycnemis, but distant from the Guineo-Congolian species that have tropical Asian affinities. It is argued that the species is a long-distance wind-borne arrival from Madagascar, which survived due to favourable climatic conditions on Pemba. Habitats on the mainland, only 50 km further, are or have been drier and therefore seem unsuitable. The new species, living proof of a remarkable colonisation event, is under immediate threat, confined to a single stream in an imperilled forest, over 1000 km from its nearest relatives. The holotype of the enigmatic P. mauriciana, not recorded on Mauritius after its description, cannot originate from the island as it pertains to the European P. latipes. Five species recalling the Asian genus Copera are known in the male sex from central and western Africa; all were confused to some degree with P. congolensis and a key is given. The lectotype of P. congolensis is designated and its identity is clarified. Platycnemis flavipes and P. xanthopus are junior synonyms of P. nyansana. Discovery of the P. rufipes female showed that P. escherichi, known only from the female holotype, is a junior synonym of it. The generic classification of Platycnemis and Copera is not resolved, but data and hypotheses that should aid future analysis are provided. Key words damselflies, Zygoptera, Platycnemis, Africa, Pemba, biogeography, taxonomy, synonyms Introduction A damselfly of the genus Platycnemis Burmeister, 1839 was recently discovered on Pemba Island, 50 km off the Tanzanian coast; separated by 1000 km of dry land and a similar distance of sea from its most proximal relatives in Uganda and the Comoros (Figs 1a, 2). Platycnemidine damselflies are represented by at least 35 species in Africa, Europe and Asia. Geographically four disjunct groups can be recognised: (1) six Western Palaearctic species, including the type species of Platycnemis; (2) at least five Guineo-Congolian species; (3) at least 10 species on Madagascar and associated islands, including the type species of Proplatycnemis Kennedy, 1920; (4) between 13 and 16 species in eastern Asia, including the type species of Copera Kirby, 1890 and Pseudocopera Fraser, 1922. Although highly ∗ Corresponding author. Email: dijkstra@nnm.nl distinctive as a group (Fig. 1), platycnemidine taxonomy is unresolved. Most species are classified as Platycnemis, while part of the Oriental species have been placed in Copera, but this generic classification is poorly supported (Ris, 1915; Schmidt, 1951b; Martens, 1996; Hämäläinen, 2003). To understand the Pemba species’ affinities and biogeography, the taxonomy, range and ecology of the Afrotropical Platycnemis (groups 2 and 3) must be reviewed (Tables 1–2, Fig. 2). In particular the taxonomy of the continental species is notoriously confused (Martens, 1996; Dijkstra, 2003). That of the insular species is quite well resolved (Schmidt, 1951b; Lieftinck, 1965), although their ecology is poorly known. Better knowledge of the group can only be obtained once the study of Madagascar’s Odonata is intensified (Dijkstra & Clausnitzer, 2004). Besides the continental species we examined the enigmatic P. mauriciana of Mauritius. It is of particular interest because, like the Pemba species, it is highly isolated from its nearest 187 188 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens Figure 1 Platycnemis males. (a) P. pembipes sp. nov., Ngezi Forest, Pemba Island, 1 October 2001; (b) P. nyansana, Mpanga Forest, Uganda, 14 May 2003. Photographs by V. Clausnitzer (a) and K.-D.B. Dijkstra (b). relatives. We do not attempt to resolve problems of the generic classification of Platycnemis and Copera, but provide data and hypotheses that should aid future analysis. Materials and methods We reviewed relevant Afrotropical Platycnemis (type) material in the Natural History Museum, London (BMNH), Institut Royal des Sciences Naturelles de Belgique, Brussels (ISNB), Museo Civico di Storia Naturale “Giacomo Doria”, Genova (MCSN), Muséum National d’Histoire Naturelle, Paris (MNHN), Musee Royal de l’Afrique Centrale, Tervuren (MRAC), Naturhistoriska Riksmuseet, Stockholm (NHRS), National Museums of Kenya, Nairobi (NMKE), Nationaal Natuurhistorisch Museum Naturalis, Leiden (RMNH), University of Michigan Museum of Zoology, Ann Arbor (UMMZ) and Museum für Naturkunde der Humboldt-Universität, Berlin (ZMHB). The Pemba species was compared with all continental African species and the insular P. alatipes, P. hova, P. malgassicum, P. pseudalatipes (all BMNH), P. agrioides and P. aurantipes (RMNH). Tropical African Platycnemis damselflies Range (Fig. 2) Thorax Dilations mid and hind tibiae Leg colour Penis Cerci /paraprocts /segment 10 Table 1 Continental group Insular group Uganda to The Gambia Often with distinct pale speckling, like spatter of bleach (Figs 3b–e) At most 2× as wide as shaft (Fig. 4b) Yellow to rufous, but never white and at most joints and tarsi darkened Madagascar, Comoros and Pemba Island Usually without pale speckling (Fig. 3a) Apex rounded; lateral branches absent (Fig. 6b) Long, often with prominent internal branch (Figs 5b–f) Two-thirds or more Equal or more Mayotte, Comoros Madagascar Madagascar Madagascar Madagascar Madagascar Nzwani, Comoros Pemba, Tanzania Madagascar Madagascar Madagascar Continental congolensis Martin, 1908 guttifera Fraser, 1950 Western Congo and Gabon Liberia to Nigeria, possibly Cameroon nyansana Förster, 1916 Congo Basin to Uganda rufipes (Selys, 1886) Cameroon, possibly Gabon sikassoensis (Martin, 1912) Uganda to The Gambia ∗ Up to 5× as wide as shaft (Fig. 4a) Often white, but sometimes a reddish or bluish colour, can be extensively marked with black Apex deeply incised; long lateral branches probably always present (Fig. 6a) Short, with small (often poorly visible) branch (Fig. 5a) Half or less Equal or less Comparison of Afrotropical Platycnemis groups. Insular agrioides Ris, 1915 alatipes (McLachlan, 1872) aurantipes Lieftinck, 1965 hova Martin, 1908 longiventris Schmidt, 1951 malgassica Schmidt, 1951 melana∗ Aguesse, 1968 pembipes sp. nov. protostictoides Fraser, 1953 pseudalatipes Schmidt, 1951 sanguinipes Schmidt, 1951 Table 2 189 Distribution of continental and insular Afrotropical Platycnemis species. corrected spelling Systematics Key to Afrotropical Platycnemis males The differences between the continental and insular species are given in Table 1. The key below is only applicable for mature Platycnemis males west of the Mozambique Channel. The species are difficult to key as they demonstrate extreme colour change with age; the characters of colour and markings refer to fully mature specimens and are in- cluded with some caution. A key to the Malagasy species is provided by Schmidt (1951b), with additions by Lieftinck (1965). 1 Dilations of mid and hind tibiae more than 3× as wide as shaft (Fig. 4a). Legs white. Cerci less than half length of paraprocts (Fig. 5a). Pemba Island, Tanzania . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pembipes sp. nov. Dilations of mid and hind tibiae at most 2× as wide as shaft (Fig. 4b). Legs yellow to rufous. Cerci at least twothirds length of paraprocts (Figs 5b–f). Uganda to The Gambia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 (1) Branches of cerci short, usually only visible in caudal view (Fig. 5f). Subapical pale ring of segments 3–7 reduced, abdomen appearing dark. Legs uniformly orange. Postclypeus with narrow pale border; synthorax usually with narrow complete antehumeral (pale) and humeral (black) stripes (Fig. 3e). Hindwing length 14–17 mm. Sunny rivers and streams . . . . . . . . . . . . . . . . . . sikassoensis Branches of cerci long, usually well visible in lateral and dorsal view (Figs 5b–e). Segments 3–7 with pale basal and subapical rings (may be incomplete dorsally), giving abdomen annulated effect. Legs yellow to rufous, often darkened at joints (‘ringed knees’). Postclypeus black at border; antehumeral stripe often broad and not bounded by humeral stripe, or mesepisternum largely dark (Figs 3b–d). Hindwing length 17–21 mm. Shaded forest pools, streams and rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3 (2) Cerci distinctly shorter than paraprocts; branches of cerci diverge from and usually fall (just) short of cerci tips (Figs 5b, d). Mid- and hind tibiae distinctly widened. Legs yellow and darkened at joints (‘ringed knees’). Synthorax with pale speckles and complete antehumeral stripe (Fig. 3c). Uganda to Gabon . . . . . . . . . . . . . . . . . . . . . . . . . 4 Cerci almost as long as paraprocts; branches of cerci lie close to and often surpass cerci tips (Figs 5c, e). Tibiae 190 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens sikassoensis sikassoensis sikassoensis guttifera rufipes ? congolensis ? ?? nyansana ? ? pembipes melana agrioides other insular species Figure 2 Approximate distributions of Afrotropical Platycnemis species; insular (black) and continental (grey). Especially the general southern limits and contact zones of continental species are unclear; range of P. sikassoensis (at least partly) incorporates ranges of other continental species. See Table 2 for further details. Figure 3 Platycnemis mature male head in dorsal and thorax in lateral view. (a) P. pembipes sp. nov.; (b) P. guttifera; (c) P. nyansana; (d) P. rufipes; (e) P. sikassoensis. Individual variation is great; dark markings are strongly reduced in younger specimens. P. congolensis is similar to P. nyansana. not widened, or dilations are mere ridges. Legs uniformly rufous or yellow with darkened joints. Synthorax often either with only speckles or antehumeral stripe (Figs 3b, d). Liberia to Cameroon, possibly Gabon . . . . . . . . . . . . . 5 4 (3) Posterior border of pronotum drawn out into two slender vertical processes. Branches of cerci reach almost to tips of cerci. Western Congo and Gabon . . . . congolensis Pronotum without vertical processes. Branches of cerci fall well short of tips. Congo Basin to Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . nyansana 5 (3) Branches of cerci with expanded tips (Fig. 5e). Legs rufous. Synthorax with smooth-edged black and pale stripes, including complete antehumeral stripe, at most speckled with a dot or two (Fig. 3d). Cameroon, possibly Gabon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . rufipes Branches of cerci with slender pointed tips (Fig. 5c). Legs yellow with darkened joints. Synthorax with irregular speckled markings, antehumeral stripe often reduced to some pale speckles on an otherwise all black mesepisternum (Fig. 3b). Liberia to Nigeria . . . . guttifera Tropical African Platycnemis damselflies Figure 6 Figure 4 Platycnemis male hind leg in lateral view. (a) P. pembipes sp. nov.; (b) P. nyansana. Platycnemis pembipes sp. nov. Figs 1a, 3a, 4a, 5a, 6a, 7 Material examined Holotype ♂, paratypes (7 ♂, 2 ♀): Tanzania, Pemba Island, Ngezi Forest (04◦ 56′ 16.6′′ S 39◦ 42′ 38.6′′ E), 38 m a.s.l., 6– 8 October 2001, leg. V. Clausnitzer. The holotype (RMNH INS 228162) and 7 paratypes (RMNH INS 228163 through 228169; 5♂, 2♀) are deposited in RMNH; 2 ♂ paratypes will be deposited in NMKE. Description Holotype male Black-and-white Platycnemis with strongly dilated white tibiae (Fig. 1a). Labium and genae whitish; labrum cream with a small blackish central spot at base; postclypeus black, narrowly pale at base; antennae black with second segment pale distally; dorsum of head black with restricted pale markings (tiny indistinct spot beside each lateral ocellus, narrow dot and line on eye border, long postocular streaks); underside largely white (Fig. 3a). Prothorax brownish black dorsally, with pale sublateral band and narrow lower border. Synthorax Figure 5 191 Platycnemis male penis (genital ligula) in lateral view. (a) P. pembipes sp. nov.; (b) P. nyansana. black dorsally down to interpleural suture, with complete creamy white antehumeral stripe less than one-fifth as wide as mesepisternum and as wide as humeral black stripe below it (Fig. 3a). Humeral stripe separates it from an equally wide pale stripe posterior of humeral suture, this posthumeral stripe interrupted along anterior two-fifths of suture but continued by pale stripe on mesokatepisternum. Synthorax sides and venter uniformly cream with a broad brownish black stripe on metapleural suture. Legs creamy white except for dark brown tarsi, blackish joints (especially femoral apices) and outer face of fore femora and tibiae; bristles brown. Fore tibiae with whitish expansions barely differentiated; mid and hind tibiae strongly expanded (about 3.5× as long as wide), all white (Fig. 4a). Abdomen brownish black; tergites 1 and 2 whitish ventrally, 3 to 6 with narrow pale ventral borders reduced apically but extended basally into narrow incomplete rings, giving abdomen an annulated impression; 7 to 9 dark; 9 with small apical white spot on each side; 10 entirely creamy white with apical rim narrowly and partially black. Appendages creamy white, cerci darkened ventrally, paraprocts black dorso-basally. Cerci roundly triangular, with black internal subbasal tooth, which is almost invisible without dissection. Paraprocts over twice as long as cerci and tergite 10, strongly curved in and down, with small dorsal subbasal knob (Fig. 5a). Penis (genital ligula) with deeply incised apex, making it strongly bilobed, and long and slender lateral branches (Fig. 6a). Wings clear; pterostigma blackish and about 1 cell long; its anterior and posterior borders equal in length, about 1.5× as long as distal and Platycnemis male appendages in dorsal (above) and lateral view (below). (a) P. pembipes sp. nov.; (b) P. congolensis; (c) P. guttifera; (d) P. nyansana; (e) P. rufipes (possible outline of damaged right cercus indicated); (f) P. sikassoensis. 192 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens Figure 7 Platycnemis pembipes sp. nov. female pronotum in lateral view. proximal borders. 15 postnodal veins in both fore wings. Measurements: abdomen (excluding appendages): 33.1 mm; hindwing: 18.5 mm; mid tibiae length × width: 3.35 × 0.95 mm; hind tibiae length × width: 3.80 × 1.12 mm. Paratype female Similar to holotype, but more faded, with dark markings reduced, indistinct and browner overall. Head pale brown overall with mere hints of darker markings, most clearly a dark line between eyes through bases of antennae. The hind lobe of the pronotum is slightly raised, forming a fairly low ridge that is shallowly incised at two points and therefore tripartite (Fig. 7). Pale brown synthorax with only dark brown dorsal half of mesepisterna and dark spot in humeral fossa distinct, latter extended into a very fine dark humeral stripe; a similar but weaker dark marking on metapleural suture. Legs evenly beige, very faintly darkened at femoral apices; tibiae not dilated. Tergites 1 and 2 pale brown, darker dorsally; 3 to 5 pale brown, apical sixth dark brown, slightly paler at base and basally of dark apical ring; 6 to 9 dark brown; apical border of 9 and entire tergite 10 dark cream. Ovipositor projects slightly beyond cerci. Pterostigma with pale brown centre, surrounded by a narrow cream border. 15 postnodal veins in fore wings. Measurements: abdomen: 32.0 mm; hindwing: 20.3 mm. Variation Paler, probably younger, males have all white cerci, facial markings more extensive and contrasting; dark thoracic markings reduced and slightly fragmented. 14–16 postnodal veins in fore wings. Habitat, behaviour and accompanying species Ngezi Forest is the only remaining large patch of tropical moist forest that once covered most of Pemba. It attained Forest Reserve status in 1959; today’s gazetted area is about 1500 ha. The reserve is situated in the northwestern corner of the island and is bordered on most sides by the Indian Ocean, as well as by agricultural land. The climate is hot and humid with a mean annual rainfall of 1860 mm, with downpours almost every week of the year. Pemba has the highest rainfall in the coastal region, where the mean annual rainfall is below 1000 mm (Clarke, 2000). The monthly temperature variation of 3.8 ◦ C is less than on the nearby mainland. The underlying bedrock is Pleistocene Azanian limestone. Most of the soils are alluvial sands; clayey soils are found in waterlogged areas by streams and ponds. Ngezi is coastal swamp forest; a rare forest type in the coastal forest belt that is very distinct from other vegetation types (Clarke & Robertson, 2000): the dominant trees are often monocots and some parts are at least seasonally flooded. The small sluggish stream that supported P. pembipes is the only permanent one. It runs through the forest’s centre towards mangroves in the north, bordered by a 30–50 m wide belt of riverine swamp forest dominated by Barringtonia racemosa and Raphia farinifera. The surrounding forest is defined as moist forest; dominated by the tree Odyendea zimmermannii (Beentje, 1990). No sunshine penetrates the dense canopy and the stream has cut a 1 m deep gully. The amount of light reaching the forest floor of moist coastal closed-canopy forest is only 0.2% of the sunlight (Moreau, 1935) and the stream’s entrenchment increases its shadiness further. Both sexes, including freshly emerged specimens, were found close to the stream. Males settled on twigs and leaves closely above or along it, facing the water. They approached females carefully, taking them into tandem position. Oviposition (seen only once) took place in tandem into dead plant material lying in the stream. The slight annual climatic variation on Pemba suggests the absence of seasonality, but more observations are needed to confirm this. Ngezi Forest’s odonates represent a typical coastal forest assemblage: Teinobasis alluaudi (Martin, 1896), Gynacantha usambarica Sjöstedt, 1909 and Thermochoria jeanneli Martin, 1915 were found in the flooded swamp forest, where they most likely reproduce, while Tetrathemis polleni (Selys, 1869) oviposited in puddles along the stream. Reproductively inactive Aciagrion cf. zambiense Pinhey, 1972 were found in the dense forest understorey. The Aciagrion may be a distinct East Coast species (own observations); more on T. alluaudi is provided in the discussion. Etymology The name, meaning ‘Pemban foot’, is consistent with names with the suffix pes that are prevalent in the genus. Platycnemis congolensis Martin, 1908 Fig. 5b Platycnemis congolensis Martin, 1908: 664 [lectotype ♂: Congo; MNHN (here designated)]. The identity of P. congolensis has long been unclear: records of that species appear to pertain to three to five species, although mostly the diagnosis by Schmidt (1951a) has been followed. MNHN possesses six and MCSN two specimens probably considered ‘Platycnemis congolensis’ by Martin (1908, 1912), which are of diverse origin, specific identity, sex and completeness (Table 3). Three specimens have the indication ‘type’, although a primary type has not been designated. Only one of these is literally labelled ‘Congo’ and fortunately this is one of only four complete males. Because this is the only specimen with labels identifying it as a type of P. congolensis from the Congo, we designate it as the lectotype in order to end further confusion regarding the species’ identity. It is not conspecific with Schmidt’s species, differing in having the posterior border of the prothoracic hindlobe drawn Tropical African Platycnemis damselflies 1♂ 1♂ 2♀ 1♂ 1♂ 1♂ 1♂ 1♀ Table 3 ∗ 193 Collection Indicated locality Labelled ‘P. congolensis’ Labelled ‘Type’ Damage Identity MNHN MNHN MNHN MNHN MNHN MNHN MCSN MCSN ‘Congo’ ‘Congo’ ‘Congo’ ‘Congo’ ‘Côte d’Ivoire’ ‘Sikasso’ (Mali) ‘F. Vaz’∗ ‘Guinée française’ yes yes yes no yes yes yes yes yes no no no yes yes no no none none none S8–10 absent S8–10 absent none none none P. congolensis LECTOTYPE P. congolensis unknown probably P. nyansana probably P. guttifera P. sikassoensis P. congolensis possibly P. congolensis Details of ‘Congolese’ Platycnemis specimens labelled by René Martin in MNHN and MCSN. Fernand-Vaz, present-day Omboué, on the Gabon coast south of Port-Gentil. out into two slender vertical processes and the longer branch of the cerci. Schmidt’s species is prevalent in Democratic Republic of Congo (DRC) and Uganda and should be called P. nyansana (see below). The distribution of P. congolensis is poorly known: The MCSN specimens are from coastal Gabon; NHRS possesses three of each sex from Mukimbungu in far western DRC on the Congo River just downstream of Luozi (Sjöstedt, 1917). Platycnemis guttifera Fraser, 1950 Figs 3b, 5c Platycnemis guttifera Fraser, 1950: 615 [holotype ♂: Ziabli, Liberia; MNHN]. Described from Liberia and since reported from Côte d’Ivoire (Legrand, 1982; Legrand & Couturier, 1985), Ghana (O’Neill & Paulson, 2001), Liberia (Lempert, 1988) and Togo (T. Lieckweg, pers. comm.). BMNH possesses material from Nigeria, identified by R.M. Gambles as P. rufipes. Possibly ranges to Guinea and Cameroon, but no specified records have been published (Legrand & Couturier, 1985). Reported from deeply shaded calm and stagnant sections of larger streams and small rivers in rainforest (Legrand & Couturier, 1985; Lempert, 1988; Dijkstra & Lempert, 2003). Platycnemis latipes Rambur, 1842 Platycnemis latipes Rambur, 1842: 242 [type not designated: Montpellier, France; MNHN (not seen)]. Platycnemis mauriciana Selys, 1862: 34; nomen nudum (no description). Platycnemis mauriciana Selys, 1863: 167 [holotype ♂: Ile Maurice; ISNB (seen)]; new synonymy. Platycnemis mauritiana Selys, 1863 – Selys (1869: 24); misspelling. The Odonata of Mauritius have been well studied, but P. mauriciana was not recorded after its description (Clausnitzer & Martens, 2004). A single male in Selys’s collection labelled ‘(Ile de France)’ and ‘Platycn. mauriciana DS.’ must represent the holotype, but lacks the terminal abdomen half, secondary genitalia, lower synthorax and three wings. Selys (1863, 1869) already doubted the specimen’s origin, regarding it a possible variety of the Western Palaearctic P. latipes. Fraser (1949) suggested that it could be close to a Malagasy species, possibly P. hova. Nonetheless the markings and broad tibiae clearly show that the male represents a Western Palaearctic species, either P. latipes or P. dealbata, which can best be separated by the appendages (lost in holotype). It is remarkable that Selys interpreted ‘Ile de France’ as Mauritius, as the island only bore that name officially from 1715 to 1810, while the part of France around Paris has been known by it since the late 14th century. The Ile de France region lies at the extreme northern limit of the recent range of P. latipes (Martens, 1996), of which we consider P. mauriciana a synonym. Platycnemis nyansana Förster, 1916 Figs 1b, 3c, 4b, 5d, 6b Platycnemis nyansana Förster, 1916: 25 [type ♂: Entebbe, Uganda; lost]. Platycnemis flavipes Navás, 1924: 12 [holotype ♂: Kibwezi, eastern Africa; MNHN (seen)]; new synonymy. Platycnemis xanthopus Navás, 1924: 13 [holotype ♂: Kibwezi, eastern Africa; MNHN (seen)]; new synonymy. Platycnemis congolensis nec Martin, 1908 – Schmidt (1951a: 223); Pinhey (1961: 19); Miller (1995: 4); Miller and Miller (2003: 121). P. nyansana was described from Entebbe, but the type is not in UMMZ (where types from the same publication have been found) and thus appears to be lost (Garrison, von Ellenrieder & O’Brien, 2003). Two species occur in Uganda, including the Entebbe area. Fraser (1928) described Copera subaequistyla, a synonym of P. sikassoensis (see below), from Entebbe. Pinhey (1961) supposed that specimens from Kampala, Bwamba Forest (= Semliki NP) and nearby Congolese Mutwanga could pertain to P. flavipes, but listed them as P. congolensis. Indeed Navás’s (1924) drawings of the P. flavipes appendages are similar to those by Schmidt (1951a) of so-called P. congolensis from Akula, Democratic Republic of Congo. Miller (1995) also reported P. congolensis from near Kampala. This second species differs from the P. congolensis lectotype in details of the male cerci and prothoracic hindlobe (see above) but matches the description of P. nyansana, being 194 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens larger than P. sikassoensis with darkened leg joints, doubleringed tergites 3 to 5 and shorter cerci. Navás (1924) described P. flavipes and P. xanthopus from ‘Africa oriental inglesa: Kibwezi’. We only know of a place by that name halfway between Nairobi and Mombasa in the dry thornbush. This is an unlikely locality for the genus; moreover there are no other Platycnemis records from Kenya. The specimens could have come from Uganda and we do not regard Platycnemis as part of the Kenyan fauna. Contrary to Pinhey (1962), the holotypes of P. flavipes and P. xanthopus are not lost (Legrand & Lachaise, 1993). Their labels add no information about their origin, reading ‘B.E. Africa, Kibwezi’. The two are alike; differences described may follow from the P. xanthopus appendages being partly covered with glue. Both agree with P. nyansana, which is the oldest and thus valid name. P. nyansana favours stagnant pools in the deep shade of rainforest (Miller, 1995; Miller & Miller, 2003; own observations). Its range encompasses Uganda and most of the Congo Basin (Schmidt, 1951a; MRAC; own observations). Platycnemis rufipes (Selys, 1886) Figs 3d, 5e Metacnemis rufipes Selys, 1886: 139 [holotype ♂: Cameroon; BMNH (seen)]. Allocnemis rufipes (Selys, 1886) – Kirby (1890: 131). Platycnemis escherichi Schmidt, 1951a: 224 [holotype ♀: Ekododo, Cameroon; ZMHB (seen)]; new synonymy. Platycnemis rufipes (Selys, 1886) – Pinhey (1962: 110); Kimmins (1970: 181). Platycnemis congolensis nec Martin, 1908 – Vick (1999: 246). Platycnemis camerunica Fraser, in litteris [holotype ♂: Cameroon; BMNH (seen)]; nomen nudum. The identity and generic affiliation of this species has been uncertain. Pinhey (1962) stated that ‘whilst this insect does not appear to be considered a true Platycnemis it does not seem to belong to Allocnemis nor perhaps to the related Stenocnemis. Gambles who has taken the species in Nigeria thinks it is a Stenocnemis.’ Kimmins (1970) added ‘currently placed (with doubt) in Platycnemis’. Neither author explained their doubts. The holotype perfectly matches Platycnemis by build, markings and appendages; moreover BMNH possesses conspecific material labelled as Platycnemis camerunica by Fraser (unpublished name). Platycnemis escherichi is known only with certainty from the female holotype, which has two diagnostic spikes on the middle of the pronotum. Legrand (1975) reported males from NE Gabon, without diagnosing the male or explaining his identification. Considering the P. escherichi holotype’s origin and Schmidt’s (1951a) remark ‘Allocnemis rufipes Selys [. . .] hat uns nicht vorgelegen’ it stands to reason that both are conspecific. The P. rufipes holotype perfectly matches the markings illustrated by Schmidt (1951a), including the dark wedges beside the eyes and the wide tapering antehumeral stripe that are not seen in other species (Fig. 3d). Material from SW Cameroon published as P. congolensis by Vick (1999) includes both sexes; the males agree with P. rufipes and the females with P. escherichi, although ‘the size of the spikes is a little variable and mostly a bit smaller than Schmidt’s drawing, but I do not think this significant’ (Vick, in litt.). Both sexes agree in markings and were collected in the same general area and once at the same site, substantiating the synonymy. Platycnemis sikassoensis (Martin, 1912) Figs 3e, 5f Psilocnemis sikassoensis Martin, 1912: 98 [holotype ♂: Sikasso, Mali; MNHN (seen)]. Copera subaequistyla Fraser, 1928: 127 [lectotype ♂: Entebbe, Uganda; BMNH (seen; designated by Kimmins, 1966: 214)]; junior synonym – Legrand (1982: 9). ? Copera congolensis (Martin, 1908) – Nielsen (1934: 180). Platycnemis subaequistyla (Fraser, 1928) – Pinhey (1961: 19). Platycnemis sikassoensis (Martin, 1912) – Pinhey (1962: 110). Unlike its continental Afrotropical congeners, P. sikassoensis favours exposed habitats and is wide-ranging, inhabiting rather open flowing waters like large rivers and sunny streams, both in forest and savanna (Legrand & Couturier, 1985; Dijkstra & Lempert, 2003). The distribution of this small species incorporates the combined ranges of the larger ones, which appear to be allopatric and restricted to forest shade (Fig. 2). It has been recorded from Benin (Tchibozo & Dijkstra, 2004), Central African Republic (Pinhey, 1971), Côte d’Ivoire (Lindley, 1974; Legrand, 1982; Legrand & Couturier, 1985), Gambia (Gambles et al., 1998), Ghana (Marshall & Gambles, 1977; O’Neill & Paulson, 2001), Guinea (Legrand & Girard, 1992; Legrand, 2003), Liberia (Lempert, 1988), Mali (Martin 1912; Dumont, 1977), Sierra Leone (Aguesse, 1968; Carfı̀ & D’Andrea, 1994), Togo (T. Lieckweg, pers. comm.) and Uganda (Fraser, 1928; Pinhey, 1961). The species illustrated as Copera congolensis from Mobeka, Democratic Republic of Congo by Nielsen (1934) may also pertain to this species. Discussion Relationships The two Afrotropical Platycnemis groups can be clearly separated by morphology and coloration (Table 1). The continental species are diagnosed and compared with P. pembipes in the key. The Pemba species perfectly fits the diagnosis of the insular group (Table 1) and is nearly identical to the Malagasy P. aurantipes and P. hova. The former is more colourful, with a more contrasting head pattern and orange-yellow legs. The male tibiae are shaped similarly, but are slightly narrower. Morphological differences are especially slight; the penises appear identical. The male cerci of P. pembipes are a bit more pointed; both species have a pale bulge on the inner border that lies more dorsal in P. aurantipes, making the internal profile more angular (dorsal view). The internal tooth of the cerci is well visible and pale with only a black tip in P. aurantipes, but entirely black and hard to see in P. pembipes. The female pronotal hindlobe of P. aurantipes is slightly broader and more deeply incised; the gap between the mesostigmal plate and mesepisternum is larger. Platycnemis hova also has more extensive and contrasting Tropical African Platycnemis damselflies pale facial markings; the penis apex is less deeply incised, the tibiae narrower and the appendages entirely white. The internal cercal tooth is black, but more prominent than in P. pembipes. Width and colour of the P. pembipes legs recall the Malagasy P. alatipes and P. pseudalatipes, whose body markings differ. Platycnemis agrioides and Platycnemis melana from the Comoros, geographically between P. pembipes and the Malagasy species (Fig. 2), are relatively dissimilar with their narrow tibiae and dark coloration. The insular species are very uniform in their contrasting coloration and penis and appendage morphology (Schmidt, 1951b; Lieftinck, 1965), and are probably monophyletic. Similar traits are seen in Eurasian species, but not elsewhere in the Afrotropics. Several authors noted the similarity between the southern Asian Copera s.s. – as defined by Selys (1863) under the homonym Psilocnemis with type species C. marginipes (Rambur, 1842) designated by Kirby (1890) – and the Guineo-Congolian (continental) platycnemidines: Nielsen (1934) transferred P. congolensis to Copera and Schmidt (1951a) placed P. nyansana close to that genus by penis characters; at least P. guttifera and P. sikassoensis are similar in this regard (own observations). Fraser (1928, 1950) described Copera subaequistyla, now a synonym of P. sikassoensis, and P. guttifera emphasising their similarity to C. marginipes and C. vittata (Selys, 1863). Recently Donnelly (2002) listed the species as Copera sikassoensis for that reason, despite the status quo to place all African platycnemidines in Platycnemis. Indeed the two most common species in their respective continents, P. sikassoensis and C. vittata, are virtually identical except for differences in the male appendages and female pronotum. Although it can be concluded that the Pemba species’ nearest relatives are found in Madagascar and those of the continental species in tropical Asia, only worldwide phylogenetic research of the group can elucidate the exact relationships. The platycnemidine damselflies are probably a monophyletic group, but the prevailing separation into two genera almost undoubtedly makes Platycnemis paraphyletic. Of the four geographically disjunct platycnemidine groups, three are morphologically uniform and probably monophyletic. Only the large eastern Asian group is heterogeneous; besides species with Guineo-Congolian affinities (Copera s.s.), it includes species ranging into the Eastern Palaearctic that are more similar to the Western Palaearctic and Madagascar groups. Interesting in this regard is that Fraser (1962) illustrated a Platycnemis larva from Madagascar with frilled caudal lamellae. This is like those described for C. marginipes and C. vittata by Lieftinck (1940), while Western Palaearctic and Japanese Copera and Platycnemis (including the latter’s type species) have unfrilled lamellae (Kawai, 1985; Martens, 1996). Discovery of the larva of African Platycnemis will help enlighten their position; it is assumed to have frilled lamellae. Another overlooked but possibly informative feature is eye-coloration, which is lost in preservation. Examined species with (assumed) frilled lamellae have a darker dorsum of the eye plus a distinct horizontal band below it, while unfrilled species only have a dark dorsal half (Eastern Palaearctic) or very faint markings altogether (Western Palaearctic). 195 Biogeography The forests of the eastern African coast and adjacent mountains are recognized as core areas of biodiversity and endemism in Africa (Kingdon, 1989; Fjeldså & Lovett, 1997; Burgess, 2000) and are listed as important conservation areas (Stuart et al., 1990). This richness, especially of relict species, has been explained by the area’s relative climate and habitat stability (Fjeldså et al., 1997). Most island biota are similar to those on the mainland (Burgess, 2000). Pemba, however, is recognized for its great distinctiveness with a high degree of endemism despite its size (about 1000 km2 ) and proximity to the mainland (50 km, similar to the distance between Zanzibar and Tanzania). The moist forest of Ngezi, for instance, is peculiar in its floristic composition, combining tree species restricted to coastal East African forests, Madagascan links, eastern Indian species and Afromontane elements (Beentje, 1990). Despite the nearness of its mainland relatives, the owl Otus pembaensis is one of the most distinctive Otus species in and around the western Indian Ocean (Rasmussen et al., 2001). The comparatively early separation from the mainland by faulting that produced the Pemba Channel, possibly 6–10 million years ago, may explain Pemba’s unique position (Richmond, 1997; Clarke & Burgess, 2000). The Malagasy connection of P. pembipes has parallels in other Pemban taxa. The tree Chrysaliduocarpus pembanus and palm Dypsis pembanus are endemics belonging to otherwise Madagascan genera (Beentje, 1990; Krain et al., 1994). The tree genus Typhonodorum and the fruit bat Pteropus voeltzkowi also have their closest relatives in Madagascar (Beentje, 1990; Entwistle & Corp, 1997). Similarly, the latter’s relative Pteropus comorensis inhabits Mafia off the Tanzanian coast and the Comoros, and is closely related to Pteropus seychellensis of the Seychelles. Other Pemban taxa show a stronger affinity to mainland Africa than to Madagascar: the milkweed genus Secamone has its greatest diversity in Madagascar (62 species) with only 16 species in continental Africa, but is represented on Pemba only by a mainland species (Goyder, 1991). Keeping these affinities in mind, the questions arise (1) what the origin of Platycnemis on Pemba is and (2) why populations are absent from the mainland. The origin of P. pembipes is most likely from Madagascar by wind-aided dispersal across the Mozambique Channel. A strong monsoon follows the East African coast north(west-)wards from June to September (Richmond, 1997) and could carry damselflies over long distances. The American damselfly Ischnura hastata (Say, 1839) that has colonised the Azores and Galapagos Islands has been collected with nets fixed to aeroplanes at 300 m altitude (Cordero Rivera et al., 2005). It is unlikely that P. pembipes is an ancient relict: Madagascar separated from the African mainland many millions of years earlier than Pemba, although it ‘. . . apparently slid south along the east African coast for most of the Cretaceous rather freely interchanging plant and animal taxa with the mainland at least until 90 Ma and perhaps intermittently thereafter’ (Gentry, 1993). Such an ancient split would predict a much greater character divergence between species on Madagascar and Pemba, moreover Platycnemis species on the volcanic and relatively young 196 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens Comoros (at most 8 million years old) could also only reach these islands airborne. The Comoran species pair and P. pembipes both have close relatives on Madagascar, but are rather dissimilar to each other, suggesting the Comoros did not act as a stepping-stone between Madagascar and Pemba, but were colonised in a separate event. The second question is more difficult to answer. Undiscovered populations may be present in unsampled coastal forests in Mozambique, southern Tanzania and Mafia Island. Alternatively insular species may have never reached the mainland or become extinct there. Considering the proximity of Pemba and the suspected mode of dispersal it is unlikely that Platycnemis never arrived on continental shores (see below). Subsequent extinction, on the other hand, seems probable. Tropical Africa has undergone marked climatic changes, for instance with a relatively wet period 12–10 000 years ago and a dry one in the few thousand preceding years (Hamilton, 1981). The impoverishment of the forest flora and fauna of Africa is due to extinctions during dry spells, glacial advances and the lack of refugia during the Pleistocene. Extinctions would have mainly hit the moist and hot lowland forests, i.e. cooling below the tolerance of tropical stenothermic species and altering their habitats. A number of plant taxa shared by the lowland forests of Madagascar and South America, for instance, are known as fossils but absent in mainland Africa today (Gentry, 1988, 1993). Examples of such random extinctions in the African rain forests are discussed by Colinvaux (1993) and Clausnitzer & Lindeboom (2002). Although the coastal forests of eastern Africa are considered relicts of a former pan-African tropical forest and are believed to have been more stable during cooler and drier periods than other African forests (Fjeldså et al., 1997; Fjeldså & Lovett, 1997), small changes in temperature or humidity may have caused the extinction of forest species. The mainland coast is much drier than Pemba and is even drier now than in the past (Clarke, 2000): suitable habitats for P. pembipes or a related species may no longer be present. The insular Platycnemis scenario has parallels in other Odonata, be it on a larger scale. The genera Teinobasis Kirby, 1890 and Hemicordulia Selys, 1870 have their greatest diversity in Australasia and the adjacent Pacific. They are poorly represented on the Asian mainland, but occur on several Indian Ocean islands, Madagascar and eastern Africa, with only a single species in each area occupied. This distribution and the restricted taxonomic differentiation of the scattered populations, suggest a recent wind-borne transoceanic colonisation. The damselfly T. alluaudi inhabits Madagascar, the Seychelles, coastal Kenya and Tanzania, Zanzibar and north Malawi, but populations are widely separated due to the scarcity of appropriate swamp forest habitat (Clausnitzer, 2003b). A complex of very similar Hemicordulia species inhabits Madagascar, the Seychelles and Mascarenes, and eastern Africa west to the Albertine Rift and south to KwaZulu-Natal, but records and habitat information are more fragmentary than in Teinobasis (K.-D.B. Dijkstra, unpublished). These cases show that the dispersal of Malagasy Platycnemis to the African coast is feasible, but restricted and probably reduced availability of habitat may have precluded their survival in large parts of this region; perhaps conditions were only favourable on Pemba. Conservation Owing to its distinctive fauna and flora, Pemba is recognised as a high priority site for conservation (Rodgers & Burgess, 2000): the Ngezi Forest is significant for biodiversity conservation on the regional and global level. Until recently Pemba was entirely covered with forest; clearing for cash crop plantations (cloves, cardamom) began in 1830 (Clarke & Karoma, 2000). Now just a few square kilometres of forest are left and endemic forest species are already thought to be extinct (Clarke et al., 2000). This fate definitely awaits P. pembipes, living proof of a unique colonisation event, if the remaining forest on Pemba is not conserved. The species is one of several odonates that are almost certainly confined to a single stream and thus under imminent threat (IUCN, 2005). Others are Amanipodagrion gilliesi Pinhey, 1962 from the Usambaras on the nearby Tanzanian mainland (Clausnitzer, 2003a, 2004) and Risiocnemis seidenschwarzi Hämäläinen, 2000 from Cebu in the Philippines (Hämäläinen, 2004). Platycnemis pembipes has been submitted for inclusion in the global Red List as Critically Endangered. Acknowledgements We thank Jan van Tol (RMNH) for the loan of P. aurantipes. Dr Roberto Poggi (MCSN), Rasmus Hovmöller (NHRS), Tammo Lieckweg and Graham Vick kindly provided information. Heinrich Fliedner was consulted on the spelling of P. melana. Dave Goodger, Kjell Arne Johanson and Jean Legrand were helpful during visits to BMNH, MNHN and NHRS. Thanks are also due to the following institutions for assistance: International Centre for Insect Physiology and Ecology (ICIPE), Kenya; NMKE; Tanzanian Commission for Science and Technology (COSTECH). The studies were supported by grants from the German Federal Ministry of Science (BMBF, BIOLOG Programme, 01LC0024, 01LC0025 and 01LC0404). The first author’s work in BMNH, MNHN and NHRS was supported by the SYS-RESOURCE, COLPARSYST and HIGH LAT infrastructures of the European Union IHP Programme; RMNH provided working facilities. References AGUESSE, P. 1968. Quelques Odonates récoltés en Sierra Leone. Bulletin de l’Institut fondamental d’Afrique noire (A) 30, 518–534. BEENTJE, H. 1990. Botanical assessment of Ngezi Forest, Pemba. Zanzibar Forestry development Project Technical Paper 1. FINNIDA. BURGESS, N.D. 2000. Global importance and patterns in the distribution of Coastal Forest species. In: BURGESS N.D. & CLARKE, G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 235–248. CARFÌ, S. & D’ANDREA, M. 1994. Contribution to the knowledge of odonatological fauna in Sierra Leone, West Africa. Problemi Attuali di Scienza e di Cultura 267, 111–191. CLARKE, G.P. 2000. Climate and climatic history. In: BURGESS, N.D. & CLARKE, G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 47–67. CLARKE, G.P. & BURGESS, N.D. 2000. Geology and geomorphology. In: BURGESS, N.D. & CLARKE, G.P., Ed., Coastal Forests of Tropical African Platycnemis damselflies Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 29–39. CLARKE, G.P. & KAROMA, N.J. 2000. History of anthropic disturbance. In: BURGESS N.D. & CLARKE G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 251–261. CLARKE, G.P. & ROBERTSON, H.M. 2000. Vegetation communities. In: BURGESS, N.D. & CLARKE, G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 83– 102. CLARKE, G.P., VOLLESEN, K. & MWASUMBI, L.B. 2000. Vascular plants. In: BURGESS N.D. & CLARKE G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK, pp. 129–147. CLAUSNITZER, V. 2003a. Rediscovery of Amanipodagrion gilliesi, with notes on habitat, behaviour and conservation (Odonata: Megapodagrionidae). International Journal of Odonatology 6, 1–8. CLAUSNITZER, V. 2003b. Teinobasis alluaudi Martin, 1896 from mainland Africa: notes on ecology and biogeography (Zygoptera: Coenagrionidae). Odonatologica 32, 321–334. CLAUSNITZER, V. 2004. Critical species of Odonata in Eastern Africa. In: CLAUSNITZER, V. & JÖDICKE, R., Ed., Guardians of the Watershed. Global status of dragonflies: critical species, threat and conservation. International Journal of Odonatology 7, 189– 206. CLAUSNITZER, V. & LINDEBOOM, M. 2002. Natural history and description of the dendrolimnetic larvae of Coryphagrion grandis (Odonata). International Journal of Odonatology 5, 29–44. CLAUSNITZER, V. & MARTENS, A. 2004. Critical species of Odonata in the Comoros, Seychelles, Mascarenes and other small western Indian Ocean islands. In: CLAUSNITZER, V. & JÖDICKE, R., Ed., Guardians of the Watershed. Global status of dragonflies: critical species, threat and conservation. International Journal of Odonatology 7, 207–218. COLINVAUX, P. 1993. Pleistocene Biogeography and Diversity in Tropical Forests of South America. In: GOLDBLATT P., Ed., Biological Relationships between Africa and South America. Yale University, Dexter, Michigan, pp. 473–500. CORDERO RIVERA, A., LORENZO CARBALLA, M.O., UTZERI, C. & VIEIRA, V. 2005. Parthenogenetic Ischnura hastata (Say), widespread in the Azores (Zygoptera: Coenagrionidae). Odonatologica 34, 1–9. DIJKSTRA, K.-D.B. 2003. A review of the taxonomy of African Odonata: finding ways to better identification and biogeographic insight. Cimbebasia 18, 191–206. DIJKSTRA, K.-D.B. & CLAUSNITZER, V. 2004. Critical species of Odonata in Madagascar. In: CLAUSNITZER, V. & JÖDICKE, R., Ed., Guardians of the Watershed. Global status of dragonflies: critical species, threat and conservation. International Journal of Odonatology 7, 219–228. DIJKSTRA, K.-D.B. & LEMPERT, J. 2003. Odonate assemblages of running waters in the Upper Guinean forest. Archiv für Hydrobiologie 157, 397–412. DONNELLY, T.W. 2002. Odonata in and around Murchison Falls National Park, Uganda. Notulae Odonatologicae 5, 114–115. DUMONT, H.J. 1977. Odonata from Mali, West Africa. Revue de Zoologie africaines 91, 573–586. ENTWISTLE, A.C. & CORP, N. 1997. Status and distribution of the Pemba flying fox Pteropus voeltzkowi. Oryx 31, 135–142. FJELDSÅ, J. & LOVETT, J.C. 1997. Geographical patterns of old and young species in African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6, 325–346. FJELDSÅ, J., EHRLICH, D., LAMBIN, E. & PRINS, E. 1997. Are biodiversity ‘hotspots’ correlated with current ecoclimatic stability? A pilot study using the NOAA-AVHRR remote sensing data. Biodiversity and Conservation 6, 401–422. 197 FÖRSTER, F. 1916. Beiträge zu den Gattungen und Arten der Libellen IV. Neue Beiträge zur systematischen Insektenkunde 1, 23–25. FRASER, F.C. 1928. Odonata of the African continent. Transactions of the Royal Entomological Society of London 76, 123–138. FRASER, F.C. 1949. The Zygoptera of Mauritius (Order Odonata). Transactions of the Royal Entomological Society of London 100, 135–146. FRASER, F.C. 1950. Report on a collection of Odonata from Liberia. Bulletin de l’Institut française d’Afrique noire (A) 12, 611–624. FRASER, F.C. 1962. Report of the Austrian expedition to Madagascar, 1958 Order Odonata. Mémoires de l’Institut scientifique de Madagascar (E) 13, 147–159. GAMBLES, R.M., MOORE, N.W., HÄMÄLÄINEN, M. & PRENDERGAST, E.D.V. 1998. Dragonflies from The Gambia: an annotated list of records up to the end of 1980. Odonatologica 27, 25–44. GARRISON, R.W., VON ELLENRIEDER, N. & O’BRIEN, M.F. 2003. An annotated list of the name-bearing types of species-group names in Odonata preserved in the University of Michigan Museum of Zoology. Occasional Papers of the Museum of Zoology University of Michigan 736, 1–73. GENTRY, A.H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annals of Missouri Botanical Gardens 75, 1–34. GENTRY, A.H. 1993. Diversity and floristic composition of lowland tropical forest in Africa and South America. In: GOLDBLATT, P., Ed., Biological Relationships between Africa and South America. Yale University, Dexter, Michigan, pp. 500–547. GOYDER, D.J. 1991. Secamone (Asclepiadaceae subfam. Secamonoideae) in Africa. Kew Bulletin 47, 437–474. HÄMÄLÄINEN, M. 2003. Platycnemis phasmovolans sp. nov. – an extraordinary damselfly from Laos with notes on its East Asian congeners (Odonata: Platycnemididae). Tombo 46, 1–7. HÄMÄLÄINEN, M. 2004. Critical species of Odonata in the Philippines. In: CLAUSNITZER V. & JÖDICKE R., Ed., Guardians of the Watershed. Global status of dragonflies: critical species, threat and conservation. International Journal of Odonatology 7, 305– 310. HAMILTON, A.C. 1981. The quaternary history of African forests: its relevance to conservation. African Journal of Ecology 19, 1–6. IUCN 2005. 2005 IUCN Red List of globally threatened species. <www.redlist.org/>. KAWAI, T. 1985. An illustrated book of aquatic insects of Japan. Tokai University Press, Tokyo. KIMMINS, D.E. 1966. A list of the Odonata types described by F.C. Fraser, now in the British Museum (Natural History). Bulletin of the British Museum (Natural History), Entomology 18, 175–227. KIMMINS, D.E. 1970. A list of the type-specimens of Odonata in the British Museum (Natural History). Part 3. Bulletin of the British Museum (Natural History) Entomology 24, 171–205. KINGDON, J. 1989. Island Africa. Princeton University Press, Princeton. KIRBY, W.F. 1890. A Synonymic Catalogue of Neuroptera Odonata, or Dragonflies. Gurney and Jackson, London. KRAIN, E., ISSA, J.A., KULLAYA, A. & HARRIES, H.C. 1994. The Coconut Palm in East Africa 2. The Pemba dwarf in Zanzibar. Principes 38, 138–141. LEGRAND, J. 1975. Contribution a la faune du Gabon, Odonates: 1re Note. Annales de la société entomologique de France (N.S.) 11, 679–696. LEGRAND, J. 1982. Contribution a l’étude des Odonates de Lamto, Côte d’Ivoire. Revue française d’Entomologie (N.S.) 4, 7–17. LEGRAND, J. 2003. Les Odonates du Nimba et de sa région. In: LAMOTTE, M. & ROY, R., Ed., Le peuplement animal du mont Nimba (Guinée, Côte d’Ivoire, Liberia). Mémoires du Muséum national d’Histoire naturelle 190, 231–310. LEGRAND, J. & COUTURIER, G. 1985. Les Odonates de la forêt de Taı̈ (Côte d’Ivoire). Premières approches: faunistique, répartition écologique et association d’espèces. Revue d’Hydrobiologie tropicale 18, 133–158. 198 Klaas-Douwe B. Dijkstra, Viola Clausnitzer & Andreas Martens LEGRAND, J. & GIRARD, C. 1992. Biodiversité des odonates du Simandou, recensement des espèces de Guinée, Afrique Occidentale (Odonata). Opuscula zoologica fluminensia 92, 1–23. LEGRAND, J. & LACHAISE, M. 1993. Les types de trois espèces de Platycnemididae des collections du Muséum (Odonata, Zygoptera). Revue française d’Entomologie (N.S.) 15, 156. LEMPERT, J. 1988. Untersuchungen zur Fauna, Ökologie und zum Fortpflanzungsverhalten von Libellen (Odonata) an Gewässern des tropischen Regenwaldes in Liberia, Westafrika. Diplomarbeit, Friedrich-Wilhelms Universität, Bonn. LIEFTINCK, M.A. 1940. Revisional notes on some species of Copera Kirby, with notes on habits and larvae (Odon., Platycneminidae). Treubia 17, 281–306. LIEFTINCK, M.A. 1965. Notes on Odonata of Madagascar, with special reference to the Zygoptera and with comparative notes on other faunal regions. Verhandlungen der naturforschenden Gesellschaft in Basel 76, 229–256. LINDLEY, P. 1974. The dragonflies of Korhogo, Ivory Coast. Bulletin de l’Institut fondamental d’Afrique noire (A) 36, 682–698. MARSHALL, A.G. & GAMBLES, R.M. 1977. Odonata from the Guinea Savanna zone in Ghana. Journal of Zoology, London 183, 177–187. MARTENS, A. 1996. Die Federlibellen Europas. Westarp Wissenschaften, Madgeburg and Spektrum Akademischer Verlag, Heidelberg. MARTIN, R. 1908. Voyages de feu Leonardo Fea dans l’Afrique Occidentale: Odonates. Annali del Museo civico di Storia naturale di Genova 43, 649–667. MARTIN, R. 1912. Les libellules du cercle de Sikasso. Feuilles des jeunes Naturalistes 5, 92–99. MILLER, P. 1995. Some dragonflies of forests near Kampala, Uganda, with notes on their ecology and behaviour (Odonata). Opuscula zoologica fluminensia 136, 1–19. MILLER, P. & MILLER, K. 2003. East African Dragonflies. Nature Kenya, Nairobi. MOREAU, R.E. 1935. Some eco-climatic data for closed evergreen forest in Tropical Africa. Zoological Journal of the Linnean Society 39, 285–293. NAVÁS, R.P.L. 1924. Odonatos nuevos o interesantes. Memorias de la real Academia de Ciencias y Artes de Barcelona 18, 315–332. NIELSEN, C. 1934. Odonati di Mobeka (Congo belga). Bollettino del Laboratorio di Entomologica del R. Istituto Superiore Agrario di Bologna 7, 164–185. O’NEILL, G. & PAULSON, D.R. 2001. An annotated list of Odonata collected in Ghana in 1997, a checklist of Ghana Odonata, and comments on West African odonate biodiversity and biogeography. Odonatologica 30, 67–86. PINHEY, E. 1961. A survey of the dragonflies (order Odonata) of Eastern Africa. British Museum (Natural History), London. PINHEY, E. 1962. A descriptive catalogue of the Odonata of the African continent. Publicações Culturais Companhia de Diamantes de Angola 59, 11–322. PINHEY, E. 1971. Odonata collected in Republique Centre-Africaine by R. Pujol. Arnoldia Rhodesia 5(18), 1–16. RAMBUR, J.P. 1842. Histoire naturelle des insectes névroptères. Roret’s Suite à Buffon, Paris. RASMUSSEN, P.C., SCHULENBERG, T.S., HAWKINS, F. & VONINAVOKO, R. 2001. Geographic variation in the Malagasy Scops-Owl (Otus rutilus auct.): the existence of an unrecognized species on Madagascar and the taxonomy of other Indian Ocean taxa. Bulletin of the British Ornithologists’ Club 120, 75–102. RICHMOND, M.D. 1997. The seashores of Eastern Africa and the western Indian Ocean islands. Sida, Department for Research Cooperation, SAREC. RIS, F. 1915. Eine kleine Sammlung Libellen von den Comorischen Inseln (Odonata). Entomologische Mitteilungen 4, 137–146. RODGERS, W.A. & BURGESS, N.D. 2000. Taking conservation action. In: BURGESS, N.D. & CLARKE, G.P., Ed., Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Camebridge, UK, pp. 317–334. SCHMIDT, E. 1951a. Über neue und weniger bekannte afrikanische Platycnemididen (Odon.). Mitteilungen der münchner entomologischen Gesellschaft 41, 217–240. SCHMIDT, E. 1951b. The Odonata of Madagascar, Zygoptera. Memoires de l’Institute Scientifique de Madagascar (A) 6, 116–283. SELYS-LONGCHAMPS, E. DE 1862. Annexe K. Névroptères. In: MAILLARD, L., Ed., Notes sur l’ile de La Réunion (Bourbon). Dentu, Paris, pp. 32–35. SELYS-LONGCHAMPS, E. DE 1863. Synopsis des Agrionines, 4me légion: Platycnemis. Bulletins de l’Académie Royale des Sciences, des Lettres et des Beaux-Arts de Belgique (2) 16, 147–176. SELYS-LONGCHAMPS, E. DE 1869. Énumeration des odonates de Madagascar et des iles Comores et Mascareignes. Recherches sur la faune de Madagascar et des ses dépendances, d’après les découvertes de Francois P.L. Pollen et D.C. van Dam. 5me partie, 1re livraison. J.K. Steenhof, Leiden, pp. 21–25. SELYS-LONGCHAMPS, E. DE 1886. Revision du synopsis des Agrionines. Première partie comprenant des légions Pseudostigma – Podagrion – Platycnemis et Protonevra. Mémoires couronnés et autres Mémoires publies par l’Académie royale des Sciènces, des Lettres et des Beaux-Arts de Belgique 38(4), [1]+iv+ 1–233. SJÖSTEDT, Y. 1917. Odonaten aus Abessinien, Ost- und Westafrika. Arkiv för Zoologi 11(14), 1–27+ 5 pls. STUART, S.N., ADAMS, R.J. & JENKINS, M.S. 1990. Biodiversity in sub-saharan Africa and its islands – conservation, management and sustainable use. IUCN, Gland and Cambridge. TCHIBOZO, S. & DIJKSTRA, K.-D.B. 2004. Rapport d’inventaire préliminaire des libellules des zones humides du Sud-Bénin. IDFReport 6, 1–6. VICK, G.S. 1999. A checklist of the Odonata of the South-West province of Cameroon, with the description of Phyllogomphus corbetae spec. nov. (Anisoptera: Gomphidae). Odonatologica 28, 219–256.