The riddle of Lyriothemis bivittata (Rambur, 1842): Lyriothemis kameliyae spec. nov. (Odonata: Libellulidae) Author Kompier, Tom text Zootaxa 2017 4250 4 315 326 journal article 33291 10.11646/zootaxa.4250.4.2 ce4e62c5-1d2a-478e-a0f3-1d2637e76f8e 1175-5326 495277 CE45B4DB-641B-4D37-AD35-93CCB384BE1D 2. Lyriothemis bivittata ( Rambur, 1842 ) ( Figs. 1 A, 1D, 2A, 2D, 4C, 5B, 6A, 7) Specimens used for description. Male : Huu Lien Nature Reserve, Lang Son Province , northern Vietnam ( 21.662 N , 106.373 E ), 23.VII.2016 , T. Kompier leg; Female : Xuan Son National Park, Phu Tho Province , northern Vietnam ( 21.110 N , 104.994 E ), 31.VII.2016 , same collector. Description of male ( Figs. 1 A, 5B). Head . Labium yellowish white, with brown inner edge. Labrum, anteclypeus, postclypeus and antefrons white, postfrons metallic dark green, corners somewhat rounded. Vertex metallic green, with two small tubercles, antennae black, occiput orange-brown. Compound eyes blackish grey, chestnut-brown dorsally. Prothorax . Black, but center of posterior lobe pale brown and anterior lobe pale. Synthorax. ( Fig. 1 A) Blackish brown with markings as follows: dorsal side uniform warm brown, but close to antehumeral suture blackish brown like sides. Mesepimeron with an oblong yellow spot on ventral half, metepimeron almost completely yellow. Legs black. Femora with short spines growing progressively longer distally, tibiae with longer black spines. Wings . ( Fig. 1 A) Slightly tinted amber, somewhat stronger at apices. Two brown streaks above and below median space in both fore and hindwing as follows: First two cells between subcosta and radius blackish brown, third and fourth cell tinted brown and cubito-anal space blackish brown until first crossvein and brownish to second. Veins black. Membranule brown. Anal loop 19–21 celled. 2–3 bridge crossveins in all wings, last antenodal crossvein complete. Forewing with 2 cubito-anal crossveins, triangle 2- celled, subtriangle 3-celled. Hingwing with 2 cubito-anal crossveins, triangle 2-celled. Nodal index in forewing 9– 17:15–9 and in hindwing 10–12:13–10. Pterostigma black, covering 2.5-3 cells, slightly over 5 times longer than wide. Abdomen. ( Fig. 1 A) Slightly broader than thorax in dorsal view, S3 broadest. S1 ochre-yellow. S2 dorsally bright red to somewhat above lateral carina, lower part orange, lateral carina thinly black. Hamulus ( Fig. 5 B) black, large, shaped like a rectangular dish, clearly depressed in center, with a backward-directed spine somewhat lateral of its inner posterior corner and a large curved incision slightly anterior of middle of inner edge. Orange genital lobe very slender, pointed and nearly 4 times as long as wide. S3 dorsally bright red to thinly black lateral carina, ventrally orange. S4–8 bright red dorsally, with almost imperceptibly black lateral, dorsal and posterior carina, ventrally orange. S9 red dorsally and orange ventrally, but with some darker blotches and black carinae. S10 black but dorsally red. Anal appendages . Cerci dark brown, twice length of S10, extending just beyond epiproct, pointed, subapically with 4 small ventral teeth. Epiproct of normal type , brown. Measurements (mm): Abdomen incl. appendages 32, hindwing 36, pterostigma 5. Description of female ( Figs. 1 D, 2A, 4C, 6A). Head . Labium black, but anterolateral corners whitish. Labrum black. Anteclypeus brownish white, postclypeus blackish brown. Antefrons, postfrons and vertex metallic dark green, antefrons with whitish edges along compound eyes. Vertex with two small tubercles, antennae brown, occiput orange brown. Compound eyes grayish black with chestnut top. Prothorax . Black, but anterior lobe with anterior edge pale yellow and middle and posterior lobe with brown spot on center. Synthorax. ( Figs. 1 D, 4C) Blackish brown with markings as follows: dorsal side uniform warm brown, but close to antehumeral suture blackish brown like sides. Mesepimeron with an oblong yellow band, more brownyellow on its dorsal half, metepisternum more warm brown towards its dorsal edge, metepimeron almost completely yellow. Legs black. Femora with short spines growing progressively longer distally, tibiae with longer black spines. Wings . ( Figs. 1 D, 2A) Tinted amber, strongest at apices. Two brown streaks above and below median space in both fore- and hindwing as follows: first three cells between subcosta and radius blackish brown, fourth to sixth cell tinted brown, cubito-anal space blackish brown until first crossvein. Veins black. Membranule brown. Anal loop 33–34 celled. 5 bridge crossveins in forewings, 2–3 in hindwings. Last antenodal crossvein complete. Forewing with 1 cubito-anal crossvein, triangle 3-celled, subtriangle 5–6 celled. Hindwing with 2 cubito-anal crossveins, triangle 2-celled. Nodal index in forewing 11–17:19–12 and in hindwing 11–13:14–13. Pterostigma dark brown, covering 2–3 cells, 6 times longer than wide. Abdomen. ( Figs. 1 D, 2A, 4C) S1 ochre yellow. S2–7 red dorsally and yellow-orange ventrally, carinae thinly blackish. S8 dorsally red, but carinae somewhat thicker black, and ventrally black. S9 dorsally with a red triangle on either side of a black line over dorsal carina, remainder black. S10 and cerci black, slightly longer than S10. S8 somewhat dilated. Vulvar lamina ( Fig. 6 A) short, slightly curled, only narrowly incised. Measurements (mm): Abdomen incl. appendages 34, hindwing 41, pterostigma 6. FIGURE 5. Hamuli in anteroventral view: (A) L. tricolor , Taiwan, courtesy Wen-Chi Yeh; (B) L. bivittata , Huu Lien NR, 23.VII.2016; (C) L. kameliyae , paratype, Huu Lien NR, 24.V.2014. Diagnosis. Van der Poorten (2008) provided an overview of all known Lyriothemis species. Lyriothemis kameliyae spec. nov. can be separated from almost all other Lyriothemis species on the basis of the combination of red abdomen (in the male), dark streaks at the wing base, and thoracic pattern with two yellow lateral markings. These characters are shared with Lyriothemis bivittata , but it can be separated from that species on the basis of many clear differences, as described below. There are two other similar species. The first occurs also in Vietnam and is most likely another undescribed species ( Figs. 2 E, 4D). It can readily be separated by its distinct dorsal thoracic stripes, which are absent in L. bivittata or vague and more rectangular in L. kameliyae . See also the discussion section. The other is Lyriothemis tricolor Ris, 1919 ( Figs. 1 C, 1F, 5A). That species is thought to range from India in the west to the Ryukyu Islands of Japan in the east, although currently not known from Vietnam , and can be separated on the basis of, amongst others, distinct yellow dorsal stripes on the thorax, a black line over its abdomen, its more orange abdominal color and differently shaped hamulus, of which the inner edge is completely straight without incision ( Fig. 6 A). Note that contrary to the key in Fraser (1936, p. 265) , L. tricolor sometimes displays dark streaks at the wing base and has 3 rows of cells at the beginning of the discoidal field, rendering these characters useless for identification purposes ( Figs. 1 C, 1F). Males of L. kameliyae differ from L. bivittata by having black labium and occiput, and more square-shaped postfrons marking, black prothorax and blackish-brown dorsum of the thorax, sometimes with hard to discern indication of dorsal stripes. The abdomen is as wide as the thorax, S1–2 are yellow with black markings, S8 is distinctly marked with black, S9–10 are all black, and the ventral side of the abdomen is yellow. Structurally they are distinctly smaller, with proportionally shorter pterostigma and differently shaped hamulus ( Fig. 6 C) with vshaped incision and posterior tooth at inner corner. The differences in wing venation seem too variable to be of much use. The male of L. bivittata by contrast has white labium and brown occiput, rounded corners to the postfrons marking, pale markings on anterior and posterior lobes of prothorax, and warm brown thoracic dorsum. The abdomen is wider than the thorax, S1–2 are ochre yellow to orange, unmarked, and S3–10 are red dorsally and orange ventrally. Structurally L. bivittata is larger, with proportionally longer pterostigma and hamulus with curved incision and a small posterior tooth in different position (fig. 6B). Females of L. kameliyae have black labium and occiput. Anterior lobe of prothorax with pale marking, remainder black. Dorsum of prothorax blackish brown with vague somewhat paler dorsal stripes. Abdomen yellow and pale orange to brownish orange, with more or less distinct longitudinal stripes, S8–10 all black, ventral side of S1–7 yellow. In L. bivittata the labium has pale edges, occiput is brown, and the middle and posterior lobes of prothorax have brown markings. Dorsum of thorax is warm brown without dorsal stripes. S1 is ochre yellow and unmarked, S2–7 are dorsally red, ventrally orange. S8–9 are dorsally red with black markings, but black ventrally. S10 is black. Structurally L. bivittata is much larger bodied, with relatively longer pterostigma and the vulvar lamina only narrowly incised. Finally, in both males and females of L. bivittata the dark streaking at the base of the wings tends to be more extensive. The characters are summarized in Table 1 . Discussion. Thanks to the help of Professor Jean Legrand I was able to study photos of Rambur’s type ( Fig. 7 ). It displays warm brown dorsum to the thorax, pterostigmata approximately 6 times longer than wide, and unmarked abdomen with orange visible on S8. These characters enabled me to identify my Vietnamese specimens. In his description of specimens in Selys’, Martin’s, and British Museum’s collections Ris (1909) described two males from Tonkin (“a. 2♂ ”) that are characterized, amongst others, as follows: “Thorax vorne kupferbraun…Abdomen anfangs sehr breit…leuchtend karminrot…unterseite 1–8 gelbrot, 9–10 dunkel….” (Front of thorax copper-brown…Abdomen first very broad…bright red…ventral side S1–8 yellow-red, 9–10 dark). He notes abdomen 33–37mm and pterostigma 4.5–5mm , respectively. According to him a third male (from Darjeeling) also belongs to this group on the basis of coloration. It measures 30 resp. 4.5mm . Although the description of the hamulus mentions a triangular incision, the accompanying drawing shows a hamulus identical to that of the male of L. bivittata here described from Vietnam . It is likely these males were indeed L. bivittata . In his description of the second group (“b. 3♂ ”) Ris does not further discuss the hamulus. It is not clear he noticed differences, nor is it evident to which of the two groups the hamulus in his drawing belongs. The three males under “b. 3♂ ” he describes as “Thorax vorne ganz schwarz, eine bronzebrauner Schulterstreif nur undeutlich erkennbar…Abdomen schlanker, Segment 9–10 oben, 8–10 unten schwarz…vordere Hälfte van Segment 2 schwärzlich.” (Dorsum of thorax completely black, a bronze-brown shoulderstripe only just noticeable…Abdomen more slender, S9–10 black above, S8–10 black below…First half of S2 blackish.) Measurements are 28, 30, 32 for abdomen and 3, 3.5 and 3.5 for pterostigmata, respectively. The illustration of the wings of the specimen from Than-Moi (p. 112) shows a pterostigma that is 4x longer than wide. It seems that although he described the underside of the abdomen as bright red, these smaller males therefore most likely refer to L. kameliyae spec. nov. FIGURE 6. Vulvar lamina: (A) L. bivittata , Xuan Son NP, 31.VII.2016; (B) L. kameliyae allotype. FIGURE 7. Rambur’s type of L. bivittata , courtesy of Jean Legrand. TABLE 1. Comparison of main differences between males (left two columns) and females (right two columns) of L. kameliyae spec. nov. and L. bivittata ( Rambur, 1842 ) .

L. kameliyae ♂	L. bivittata ♂	L. kameliyae ♀	L. bivittata ♀
labium	black	yellowish white	black	black edged pale
occiput	black	brown	black	brown
postfrons mark	more square	more rounded
prothorax	black	pale markings	black with pale mark on anterior lobe	also brown marks on middle and posterior lobe
dorsum of thorax	blackish brown	warm brown	blackish brown with indistinct dorsal stripes	warm brown
S1–2	yellow, marked with black	ochre to orange, unmarked	yellow, marked with black	ochre yellow and red
abdomen	red, same width thorax, venter yellow	red,>width thorax, venter orange	yellow, pale orange to brown, often striped, venter yellow	red, unmarked, venter orange
S8–10	S8 with black, S9–10 black	red	black	S8–9 with red, remainder black
hamulus	v-shaped incision and tooth at corner	curved incision and tooth removed from corner
vulvar lamina	widely incised	narrowly separated
pterostigma mm	3–3.5 about 4x longer than wide	5 about 5x longer than wide	4 about 4x longer than wide	6 about 6x longer than wide
extent of streaking at wing base	3 cells	4–5 cells	3–4 cells	6 cells
abdomen mm	26.5–29	32	27–27.5	34

The five female specimens described by Ris are all somewhat different. Female “a.” (from Sibsagar, Assam ) has yellowish brown dorsum to the thorax and bright red abdomen (although S7–10 are missing). The pterostigma is with> 4mm possibly short for L. bivittata , but the coloration seems to suggest this species. Female “b.” (from Than-Moi, Tonkin ) has similar dorsum, but with vague shoulder stripes (sic.) and S1 black dorsally, S2– 7 brownish and S8–10 black. Its abdomen is 29mm , pterostigma 4mm . This likely is L. kameliyae . Female “c.” (from Tonkin ) has a darker thorax, but with a vague copper-brown stripe. S1 is black dorsally, S2 black with yellow spots, S3–7 are yellow with black dorsal and posterior carinae and black longitudinal stripe, S8–10 black. Below S1 is black, S2–7 are yellow and 8–10 are black. Its abdomen is 30, pterostigma is 4mm . This is unmistakenly L. kameliyae . Female “d.” (from Tonkin ) has S1–8 bright red, S8 with black sides, S9–10 black. Its abdomen is 33mm and pterostigma < 5mm . This could be L. bivittata . Female “e.” (from Tonkin ) has S2–7 brown, S8–10 black and is again smaller, abdomen 29 and pterostigmata 4.5mm , respectively. This is another L. kameliyae candidate. Fraser (1936) provided a more extensive description of L. bivittata . His description of the male is based on large ( 33–37mm ) specimens and mentions bright yellow labium, prothorax with posterior lobe ferruginous at middle, bright red abdomen with black borders on S9–10, all suggesting L. bivittata . The illustration of the wings on p. 264 shows a pterostigma 5x longer than wide and extensive dark streaks at the wing base, also supporting L. bivittata . His description of the female contains characters of at least two specimens from widely different origin ( Assam and Laos ), but does not contain definite characteristics of L. kameliyae , or it must be the statement that in some cases S8–10 are black. It is possible that some of the characteristics he mentions do not belong to L. bivittata but concern another species. Asahina (1969 , 1996 , 1988 ) provides us with something of a puzzle. Asahina (1988) features a Thai male from Chanthaburi that is relatively large (abdomen 31mm ) with entirely red abdomen. However, important parts of the description are not provided, and the illustration of the hamulus is very sketchy (and does not show the posterior tooth). It is possible this specimen is true L. bivittata . It illustrates the hamuli of a male from Nepal and a male from Thu Duc in southern Vietnam but does not further describe these. The male from southern Vietnam had earlier been described in his 1969 paper. The illustrations of the hamuli in his 1988 paper for both specimens suggest L. kameliyae . Likewise, the description of the male specimens from southern Vietnam in his 1969 paper (p. 13–14) refers to “front side of pterothorax uniformly dark brownish” and abdomen length 26—29.5mm . This also indicates L. kameliyae . The depiction of the hamuli in this (1969) paper shows the posterior tooth in a different position from that in Asahina (1988) . Either one of the illustrations must be incorrect, or a different individual (from Trang Bom in southern Vietnam ?) is illustrated in his 1969 paper. If the illustration is correct, the identity of the concerned specimen is best left undecided. It is also not likely that it is L. sp. (see below), of which at least the female has bright yellow markings on the dorsum of the thorax. The 4 Thai females (also from Chanthaburi ) Asahina (1988) describes are relatively small, 25–29mm , and have shining black labium and chocolate brown thorax. Their abdomen is described as orange yellow when young, reddish brown when aged, with black edges on S8–9 and black S10. Although somewhat of a mix, it is well possible these females belong to L. kameliyae . The female depicted in Asahina (1996, p. 76–77) from Cuc Phuong in Ninh Binh Province looks very much like L. kameliyae . It has yellowish abdomen with black stripes and very dark dorsum to the thorax. The streaks at the wing bases seem short for L. bivittata too and more in line with L. kameliyae . But the large size (abd. 39mm ) is puzzling. It seems plausible this is a mistake. Lastly, the immature female mentioned in Asahina (1969) is, based on wing pattern, size, coloration of thorax and labium a good candidate for L. kameliyae , although he does not mention distinct striping. Lahiri (1987) reported a possible L. bivittata female from Meghalaya in India. Judging from the size (abdomen 33mm and hindwing 42mm ) in combination with the extensive dark streaking at the wing base, it is certainly not L. kameliyae . Do et al. (2011 , p. 15) also recorded L. bivittata from Cuc Phuong National Park, but it is obvious from the photo provided that this is in fact L. kameliyae . Many of these old descriptions cannot be conclusively attributed to either species, although some can. It is evident that both species had been lumped into one for over a century. It is thus necessary to check specimens in collections to ascertain their true identity. The male L. bivittata specimens made available on the website of the British Museum of Natural History (www.data.nhm.ac.uk; BMNH(E)1201870 & 1201871) are clearly true L. bivittata . Other females from Chanthaburi and one female from Kanchanaburi not discussed here can for instance be found in the collection of Bro. Amnuay Pinratana in Bangkok (M. Hämäläinen in litt. ). Within Vietnam there is one more Lyriothemis species that looks like both L. kameliyae and L. bivittata and is somewhat intermediate in size. The male still eludes capture. The female ( Figs. 2 E, 4D) shares the lateral thorax pattern and dark streaks on the wing bases with both species, and a bright red unmarked abdomen (S1–7) with L. bivittata . But it has all black S8–10 and two bright yellow dorsal stripes on an otherwise blackish brown dorsum of its thorax. Although these dorsal stripes are similar to those of L. tricolor ( Figs. 1 C, 1F), the abdomen pattern does not fit that species. Until its male is found, this species remains an enigma. In addition to the specimen illustrated here I have recorded it also from Ha Tinh Province . Note on behavior and occurrence. Males and females of both species inhabit tropical rainforest habitats in mountainous areas, where they occur in very low densities. They often perch on branches in areas of tree fall, or otherwise sunny open spots inside the forest. Some individuals were observed striking pools of water on hot afternoons, likely to cool off. L. kameliyae has been verified from reserves in northern Vietnam , especially with karst mountains. In Xuan Son National Park it was found at approximately 1000m asl, but in Huu Lien Nature Reserve the elevation is only 200–300 meters. At both these locations it occurs together with L. bivittata . In addition to the other records mentioned previously, L. kameliyae has also been photographed at Cuc Phuong National Park and Tam Dao National Park by Sébastien Delonglée (pers. comm.). Both species appear in early April. L. bivittata has been recorded at least into September and L. kameliyae into early August. The range of L. kameliyae possibly extends westward to Nepal ( Asahina, 1988 ), but it certainly occurs in the south-east of Yunnan in China , judging from photos of both a male and female published on http://nc.kl.edu.tw/bbs/ showthread.php?t=45178. Van der Poorten (2008) called L. bivittata widespread in the southern part of continental Asia , known from Bangladesh , Nepal , NE India , Lao PDR, Myanmar, Thailand, Laos and Vietnam . In addition Haomiao Zhang (2011) reported L. bivittata from Southern Guizhou , China . Although it likely occurs over this complete range, given the specimens from India on the one hand, and its occurrence in the north of Vietnam on the other, making it plausible it also occurs in China , it is advisable to re-examine all specimens in collections. It can be concluded that L. bivittata and L. kameliyae at least co-occur over large parts of their known ranges.