Mr, W. F. Kh-hy on Odonata from Hainan. 531 

 LibelliUidaB. Agrionidae. 



LlBELLULIN^. AgUIONIN.'E. 



Neurothemis, Brauer. 



1. TidUa, Dru. Matrona, De Selys. 

 Trithemis, Brauer. 9- *«*'»^«''«s, De Selya. 



2. trioialis, Rami). Baijadera, Ue Selys. 

 Crocothemis, Brauer. l"- ^P- 



3. servUia, Dru. Pseudophcea, Kirb. 

 Ztjyonidia, g. n. H- <^eco''«^«, De Selys. 



4. msj^rtw, sp. n. Rhmocypha, Ramb. 

 Orthetrum, Newm. 1^- Whileheadi, sp. n. 



5. testaceum, Burm. 



Biplacodes Kirb. Ccenagrionin^. 



b. nebuiosa, rabr. 



Acisomu, Ramb. Ceriagrion, De Selys. 



7. panojyoides, Ramb. 13. coromandelianum, Fabr. 



Pseiidolestes, g. n. 

 .ffishziidae. 14, mirahilis^ pp. it. 



GOMPHIN.B. 



yEslma, Fabr. 



8. Thomassoni, sp. n. 



1, Neurothemis Tullia. 



Libellula Tullia, Drury, 111. Exot. Eat. ii. pi. xlvi. fig. 9 (1773). 

 NturotAemi's Tullia, Kirb. Cat. Neur. Odou. p. 8. n. 6 (1890). 



Five specimens were obtained of this common East-Indian 

 species. 



2. Trithemis (?) trioialis. 



Libellula trivialis, Ramb. Ins. N^vr. p. 115 (1842). 



Tritheynis tririalis, Kirb. Cat. Neur. Odon. p. 18 (1890); De Selys, 



Ann. Mus. Genov. xxx. p. 467 (1891). 

 Diplacodes trivialis, Karsch, Ent. Nachr. xvii. p. 246 (1891). 



Three specimens obtained. 



Dr. Karsch has pointed out that this species is not a true 

 Ti'ithemis, from which it differs in having the sectors of the 

 triangle distinctly separated on the hind wings. He refers 

 it to Diplacodes, the type of which is D. tetra, Ramb. ; but 

 from Diplacodes it differs in several important characters. 

 The triangle of the fore wings is traversed ; the subtriangular 

 space consists normally of 3 cells, and only exceptionally of 

 2, whereas in D. tetra the usual number is 2, and only 

 exceptionally 1 or 3 ; and the triangle is followed by one row 

 of 3 cells and then several of 2. The moi'e extensive genera of 

 Libellulinee need a further revision ; typical Tr ithemis hus the 

 triangle followed by 3 cells, increasing, and Diplacodes by 2 ; 

 and species with one row of 3 cells, followed by several of 2, 

 do not, strictly speaking, belong to either genus. But I do 

 not wish to subdivide existing genera in the present paper. 



34* 



