Male genitalia of 25 spp. are studied using scanning electron microscopy, and the structure of the fourth penile segment is described. Remarkable diversity exists among spp., particularly in the size and shape of the lateral lobes, and the morphology of the cornua. There are also differences in the number of cornua among the taxa surveyed. The surface of the lobes of many spp. is covered with spines which anchor the penis during copulation, and may trap and remove sperm when the penis is collapsed and withdrawn following copulation. Spp. are categorized according to the morphology of the penis and inferred patterns of sperm removal. Type 1 taxa possess relatively large, broad, flat lateral lobes, and lack cornua, or possess cornua that are greatly reduced in size. These spp. are believed to displace sperm in the bursa copulatrix before depositing their own sperm, thereby gaining positional priority during oviposition. Type 2 spp. possess elongated lateral lobes and/or cornua. These taxa are believed to engage in a mixed strategy of sperm displacement and sperm removal. Optimization of these characters on a phylogeny of the 3 genera indicates that the Type 1 sperm displacement strategy is ancestral, and that the Type 2 strategy was subsequently derived within the majority of the Libellula s.s. taxa.

A checklist of 48 spp. recorded from the island of Bioko (3degree30’N 8degree40’E), based on historical records compiled by Dr Elliot Pinhey and augmented by a collection made in March-April 1999, is presented, together with notes on the distribution of the spp. on the island. The Odonata fauna apparently comprises 2 elements, an upland fauna which includes 6 spp. that only occur above 500 m, and a lowland fauna of 17 spp., none of which occurs above 500 m. A further 13 spp. have been found in both lowland and upland sites. There are insufficient published data on the remaining 13 spp. to assign them to either of these faunistic elements. In addition, unique fan-shaped setae on the dorsum of the head, thorax and abdomen of early instar libellulid larvae are described. These setae are thought to be mechano-receptors that are held above the sediment to detect the presence of prey and current flow while the larva remains buried.

9 spp. have been recorded from sub-Saharan Africa, of which only 5 spp. have been described as larvae. A more complete larval description of A. scotias Pinhey from Cameroon and a review of the current level of information on the larvae of the genus in the region, are provided here, and the known larvae are keyed.

Preliminary considerations concerning the species diversity of East African dragonflies and the problems of identifying and using such diversity figures are given. For a detailed approach the basic problem is lack of sufficient data in that area. A checklist of species recorded so far for East Africa is given. Looking at pure species number in relation to area, Uganda is definitely more important for dragonfly diversity than its eastern neighbouring countries. If taking endemism and taxonomic singularity into account, the coastal forests of Tanzania and Kenya are very important too.

Collections were made at 8 localities in southern Ghana during the summer of 1997. Three regions were sampled: coastal savanna, wooded savanna, and rainforest. 71 spp. were collected, 24 of which are new for the country, bringing the Ghana list to 123 spp. A list of spp. known from the country is included.Trithemis dejouxi Pinhey, 1978, is raised to specific rank. Individual variation in Phaon iridipennis and Palpopleura lucia is quantified. West African Odonata biodiversity and biogeography are discussed.

The information on taxonomic disjunction at the gap between the Paluma Range (ca 19degreeS) and Eugella (ca 21degreeS) of mostly species pairs of Odonata is updated and discussed. The holotype male of A. u. pinheyi ssp. n. comes from Queensland, Carnarvon Gorge (18/21-II-1990), deposited in ANIC, Canberra.A. u. coloola Theischinger and A. u. speciosa Sjostedt are appraised, respectively re-appraised, as distinct species.

The morphology of this African sp. is illustrated and described. A comparison is made with other known Brachydiplactinae, the South American Elga leptostylaand Nephepeltia phryne and the African Hemistigma albipuncta. The group is a very heterogenous one in terms of larval morphology.

The sp. inhabits streams and rivers in the mountain forests of the Indian Ocean of La Reunion. Because of heavy seasonal rainfalls and steep inclines, there are drastic changes of water level. As a consequence, no water plants, roots or wet driftwood at all are available for oviposition. The sp. seems to be well adapted to this habitat: females oviposit into wet soft and porous lava stones. The closely related C. insularis and C. rufipes inhabit streams and rivers on Mauritius, where there are less extreme conditions. Both spp. deposit eggs in plant tissue.

Aciagrion had until now not been recorded in South Africa. The new sp. from a pan in thick subtropical savanna is described, illustrated and compared with similar congeners. Holotype male, allotype female (in copula): South Africa, KwaZulu-Natal prov., Ndumo Game Reserve, 21-I-2000; deposited at SAM, Cape Town.

---

The new genus is established to receive the 2 new spp., P. kiautai sp. n. (type sp.; holotype male, paratypes of both sexes: China, Guizhou, Chishui Alsophila Nature Reserve, 18-V-2000; deposited at the Zheijang Mus. Nat. Hist., Hangzhou) and P. pinheyi sp. n. (holotype male: China, Guangxi, Damingshan, 13-V-1997, paratypes of both sexes; same locality, 14-V-1997; holotype and 1 female to be deposited at the Tai Lung Experimental Stn, Agriculture and Fisheries Dept, Lin Tong Mei, Sheung Shi, NT, Hong Kong). Priscagrion gen. n. is compared with Arrhenocnemis Lieft. and Mesopodagrion McL.

Genetic methods are often utilised for the ecological study of odon. spp. In many instances, especially from a conservation standpoint, it is desirable to employ a method of extracting DNA that does not affect the subsequent survival of the animal under investigation. Removal of part of an odon. leg has been demonstrated not to affect the subsequent reproductive success of the animal. Thus for odonates, a simple and quick method of extracting DNA from a portion of an odon. leg is presented that provides high yields of DNA suitable for PCR.

The odon. fauna from Hainan is enumerated. Vestalis miao sp. n., Bayadera kirbyi sp. n. (Calopterygidae), Burmargiolestes xinglongensis sp. n., Rhinagrion hainanense sp. n., Podolestes pandanus sp. n. (Megapodagrionidae), Drepanosticta zhoui sp. n., D. elongata sp. n., Sinosticta hainanense sp. n. (Platystictidae),Chlorogomphus icarus sp. n., C. gracilis sp. n. (Chlorogomphidae), Planaeschna celia sp. n. and Oligoaeschna sabre sp. n. (Aeshnidae) are described.Lamelligomphus hongkongensis Wilson is assigned as a synonym of L. hainanensis Chao. 6 spp. of Macromia are treated. Macromia hamifera Lieftinck is synonymised with M. clio Ris and other potential synonymies of Vietnamese and South China macromias are discussed. Pseudagrion australasiae Lieftinck,Ceriagrion indochinense Asahina, Gomphidia a. abbotti Williamson and Macromia calliope Ris are recorded from Chinese territory for the first time. A checklist of dragonfly species from Hainan and details of 71 taxa not previously recorded from Hainan are provided.

Rhinocypha seducta sp. n. (holotype male: southern Vietnam, Lam Dong prov., nr Di Linh, 26-IV-1998) and Euphaea hirta sp. n. (holotype male: southern Vietnam, Lam Dong prov., Bao Loc, 14-VI-1996) are described and illustrated, and their taxonomic status is discussed. The latter sp. co-occurs with E. gueriniRamb. and E. masoni Sel.

The new sp. is described and illustrated from a single male (holotype): Taiwan, Hsinchu Hsien, Chienshih, nr Yuan Yang Lake, alt. 1670 m, 11-VIII-2000; to be deposited at Taiwan Forest Res. Inst., Taipei. It is compared with S. dido Needham.

The new sp. is described from the Sierra de Chiribiquete of Colombian Amazonia. Holotype male: Colombia, Sierra de Chiribiquete, Puerto Abeja, 5-VII-1996; to be deposited in USNM, Washington, DC. It belongs to the modesta group of G.H. Bick and J.C. Bick (1991, Odonatologica 20: 453-458), and can be distinguished from all other Polythoridae by transverse gold bands on the hind wings.

On several occasions aggregations were seen hovering high among trees beside streams in Mexico. Intermittent arrival of conspecific females at some aggregations and their departure, sometimes in tandem, and the apparent absence of foraging within aggregations supports the inference that such aggregations are mating swarms. It remains to be discovered whether trees over which swarms form are potential oviposition sites for this sp., or are places in which both males and females overnight. Flying aggregations of males in other spp. of Protoneuridae are described and discussed.

Long-term monitoring of a remnant population of M. xanthomelas, located at Tripler Army Medical Center (TAMC) began in May 1997 and continued to Febr. 2000 for the mitigation ponds and June 2000 for the TAMC stream. This sp. has been reduced to little more than 100 m of stream habitat on Oahu at the TAMC. Threats to M. xanthomelas include alien fish spp., stream dewatering, and habitat alteration. The TAMC stream now requires augmented water flow because construction of a facility up gradient of the TAMC stream disrupted the normal hydrology of the small stream. The Oahu race of M. xanthomelas will soon become extinct if the stream were allowed to become dry, as nearly happened in June 1997. The most cost-effective way to ensure the survival of this sp. on Oahu would be to continue some mitigation water flows to the TAMC stream. The next step would be the establishment of another wild population to a stream lacking alien fish spp. It is highly recommended that a cooperative association of biologists from the Bishop Museum, University of Hawaii, U.S. Fish and Wildlife Service, and U.S. Army environmental staff continue to monitor the population of M. xanthomelas, arguably the rarest insect population in the United States.

The final-stadium larva is redescribed from 24 authenticated larvae and exuviae. It is a medium-sized larva (total length 17.6-20.8 mm) with three ventral stripes and a dorsal pattern on the abdomen that includes large spots in the sublateral area on at least segments 4-7. The penultimate stadium is also briefly described and some natural history observations are reported. A key for the seven nearctic Leucorrhinia spp. is given.

The pattern of the number of spp. per pond changed completely between 1988 and 2000. The growth of scrub on the edges of the ponds caused a decline in the number of dragonfly spp. when more than 50% of pond edge was shaded by bushes. When the ponds were completely shaded they lost all their dragonfly spp. The decline in spp. appeared to be caused by shading rather than changes in the aquatic flora caused by shading. When the scrub was not controlled the ponds were inhabited by dragonflies from 26 to at least 39 yr. In ponds where no scrub developed, or where it was controlled, Coenagrion puella, Ischnura elegans,Aeshna cyanea, Libellula quadrimaculata, Sympetrum striolatum and Sympetrum sanguineum were still continuing to inhabit the ponds after 39 yr. Aeshna grandis, which had occurred most yr. in the 1962-1988 period, was still present in 2000. Pyrrhosoma nymphula and Brachytron pratense, which had bred from 1964 to 1972 and 1968 to 1973 respectively, both reappeared after 20 yr. absence. Their return to the ponds appeared to be connected with their increase in adjacent habitats at Woodwalton Fen in the 1989-2000 period. Anax imperator and Libellula fulva, which had also increased at Woodwalton Fen, were seen for the first time at the ponds during this period. The reason for the recent disappearance of Lestes sponsa from the ponds and adjacent habitats is not known. This study yet further emphasises the need to conserve large core populations of dragonflies, and it reiterates the need for really long term monitoring.

A large number of fossil dragonfly larvae have been collected in Lushangfen Formation (Lower Cretaceous), SW of Beijing, China. All the well-preserved specimens are closely similar to the larvae attributed by L.N. PRITYKINA (1977, Trans. Soviet-Mongol paleontol. Exped. 4:81-96) to Hemeroscopus baissicusPritykina, 1977, particularly for the labial mask structures.

The sp., formally known from only Chapada, Brazil, has now been reported at other sites and 2 other countries from South America (Bolivia, Paraguay). Some additional characters including penis structure in the male and the prothoracic hind lobe, mesostigmal plates and caudal appendages of the female are described and illustrated.

In order to define the eastern limit of the distribution of C. s. servilia (2nmale = 25, XO) and the western limit of C. s. mariannae (2nmale = 24, neo-XY), the karyotypes were studied in specimens from Cheju island and the mainland of the Republic of Korea, and from the islands of North Kyushu, Japan. The dividing line between the 2 sspp. appears in the Korea Strait, between the Korean mainland and Tsushima island. The mechanism of maintaining a distinct karyomorph in each area of distribution is discussed in details.

Amphigomphus nakamurai sp. n. (holotype male: Vietnam, Binh Phuh prov., Mt Tamdao, 30-V-1993), Merogomphus tamdaoensis sp. n. (holotype male: Vietnam, Binh Phuh prov., Mt. Tamdao, 19/24-V-1993), and Leptogomphus baolocensis sp. n. (holotype male: Vietnam, 15 km from Baoloc to Ho-chi-minh, Lamdong prov., 16-VI-1996) are described, illustrated and their affinities are pointed out. Material is deposited in the author’s institution.

This synopsis includes diagnoses for adults, a key for males and females, illustrations of taxonomic characters and updated distribution data for each sp. The ventral terga contour is found to vary interspecifically allowing identification of all species. Other useful characters are the presence or absence of black stripes over frontoclypeal and fronto-ocular grooves, abdominal colour pattern and shape of the cerci. Some colour characters of A. variegata vary geographically and two forms are described: a dark (humid biomes) and light form (dry biomes). The synonymy of A. peralta and A. variegata is considered doubtful.

During copulation and before sperm transfer, odon. males are able to manipulate rival sperm stored in the female sperm storage organs (usually the bursa copulatrix and spermathecae). males of the territorial C. h. asturica use 2 mechanisms for this. Bursal sperm is removed physically whilst spermathecal sperm is displaced via aedeagal stimulation (through a series of abdominal flexions) of the female sensory system that controls spermathecal sperm ejection. Most bursal sperm is removed but there is individual variation in spermathecal sperm displacement. Previous results have found that this variation is related to aedeagal width. In this paper 4 variables that may also explain variation in spermathecal sperm displacement ability are investigated: male age and status (territorial and nonterritorial), duration of the sperm displacement stage and the number of aedeagal stimulatory flexions. Variation in the ability to displace spermathecal sperm, however, was not related to these variables. This suggests that variation in this ability is reliant only on male genitalic attributes, aedeagal width. These results are briefly discussed in terms of current theory of sexual selection as the process propelling genitalic evolution.

The female is described and illustrated, and illustrated keys to the known male male and female female of Leptagrion from Brazil are presented.

From May 1997 to April 1998 larvae were recorded at the Weidlingbach, a fourth order tributary of the Danube or Vienna, at 12 sampling stations from source to mouth. From the 14 larval instars reported for the genus, 5 (F to F-4; based on head width) could be identified in both spp.; head widths of female larvae were significantly larger than in male. – Both spp. were most abundant within medium sand sediments with a mean grain size (Q50) of 2.04 mm in C. bidentata and 2.79 mm in C. heros. Mean water depths and nose current speeds measured at larval microhabitats were 4.4 cm and 2.3 cm s-1 (C. bidentata) and 5.6 cm and 2.6 cm s-1 (C. heros). During the winter months larvae chose the water depths slightly deeper than during summer. Throughout the observation period, a high proportion of the larvae (C. bidentata: 70-100%; C. heros: 41-90%) were burrowed in sandy sediments, either totally or displaying the typical ambush posture with only head and anal pyramid visible. In winter, the proportion of burrowing larvae was insignificantly higher than in summer. – C. bidentata larvae were most abundant near the sources, preferring first order stream sections (discharge 0.1-3.21 s-1) with high hardness (up to 34 German degrees) and conductivity (up to 1100 muS cm-1) and a high proportion of fine sediments. Although C. heros larvae were also collected at such first order sites, they reached their highest abundance (larval density up to 7.84 specimens per 10 meter shore length) at second order stream sections (discharge 0.3-6.01 s-1) with lower hardness and conductivity and a higher proportion of coarse sediments.

According to their social status, the males are divided into 2 categories: territory holders (territorial males) and non-territory holders (wandering males). The duration of copulation was longer in wandering (461.5+-347.8 s, n=46) than in territorial pairs (201.3+-149.8 s, n=85). Oviposition modes are 3-fold: viz. (1) single, with non-contact guarding (territorial pairs: 115.1+-75.5 s, n=27; wandering pairs: 133.9+-45.5 s, n=14); – (2) tandem oviposition (territorial pairs: 214.6+-76.6 s, n=5; wandering pairs: 141.0+-76.2 s, n=7); – and (3) tandem oviposition+non-contact guarding (NCGO) (territorial pairs: 131.6+-93.8 s, n=5; wandering pairs: 157.5+-72.8 s, n=2). The first type was most common in territorial (75%) and in wandering pairs (62.1%). The second type was in wandering pairs (24.1%) twice as frequent as in territorial pairs (10.4%). The third mode was infrequent (territorial pairs: 14.6%, wandering pairs: 13.8%). The relationships between male social status and the interference of territorial male male on one hand, and the duration of copulation and the oviposition mode on the other, are discussed. The effects of vegetation and air temperature on the oviposition mode are briefly outlined.

The last larval instar of Aeshna absoluta Calvert 1952 and A. confusa Rambur 1842, as well as that of the light form of Aeshna variegata Fabricius, are described for the first time and compared with the other Patagonian larvae of this genus. They can be identified based on prementum width/length ratio, number of teeth on each side of the prementum median cleft, shape of prothoracic processes and relative length of the terminalia.

Five spindle-shaped sperm ‘pellets’ (bundles) were found in the vagina of copulation-interrupted androchrome female. The second and third pellets had a long thin tail while the fourth had partly and the fifth completely disintegrated. In an other female, the vagina was packed with sperm material even though the sperm storage organs (spermatheca and bursa copulatrix) were completely filled. It is proposed that this sp. exhibits a transitional reproductive behaviour with respect to sperm competition.

The new sp. is described, illustrated and compared with L. simplex Hag. Holotype male and allotype female: Veracruz, Laguna de Santo Domingo, 4 km NW of Huatusco, alt. 1300 m, 9-VII-2000; deposited at CNIN, UNAM, Mexico.

At the rendezvous, males prefer stones as perches. Discrimination experiments with pairs of substrates showed that they land preferentially on perches that correspond in height to the flight level of female female appearing at the water. When they first landed, males preferred perches in the middle of the stream, but afterwards they also used those near the stream margin. The results are interpreted in terms of early recognition of females and rapid formation of tandem linkage.