462 



of the segments has been reduced as compared with that of the thorax. 

 In the Anisoptera there are other modifications besides the reduction 

 in diameter. Here, the abdomen is sometimes triangular in cross-sec- 

 tion, and different portions of the abdomen of the same species have 

 different cHameters. Considering shape alone, the following line of 

 development may be recognized, beginning with the more generalized : 

 Agrionidae, Coenagrionidae, Aeshnidae, Libellulidae, and Gom- 

 phidae. This order of specialization is followed throughout in the 

 abdomen. 



30. — The approximation of the terga on the ventro-meson, is a 

 mark of specialization most frequently found in the Anisoptera, as is 

 also the appearance of the secondary ridges on the terga. 



31. — The anal appendages of the abdomen are interesting, and 

 the line of specialization indicated by them seems to coincide in general 

 with that already outlined for the suborders in 29 and 30. The series 

 has already been given for the two groups in paragraph 13. Within 

 the Anisoptera, two different lines are found, both probably represent- 

 ing specialization. In one of these the inferiors are fused, as in the 

 Libellulidae ; in the other the superiors are enlarged and expanded, as 

 in the Aeshnidae. In the Zygoptera the forcipate appendages of the 

 Agrionidae probably represent the most primitive forms, and the short 

 and frequently greatly modified appendages of the Coenagrionidae, the 

 more highly specialized. 



32. — Accessory male genitalia of the second segment are important. 

 The statement that this organ has been derived from the sexual organs 

 of the progoneates is substantiated by the reported connection of the 

 proximal end of the penis with the visceral cavity. This occurs in 

 Zygoptera and seems not to have been observed in the Anisoptera, 

 the connection supposedly having been lost through specialization. 

 Further specialization has been suggested in the tracheation of the 

 appendages, which occurs in some Anisoptera according to Backhoft' 

 ('10) but not in Zygoptera. Other differences indicating specialization 

 in Anisoptera are to be noted in the segmentation of the penis and 

 in the position and connection of the seminal vesicle with the latter. 

 The structure of the hamules and the genital lobes, and of the portions 

 of the genitalia arising from the third abdominal segment, seems to be 

 simpler in the Zygoptera and not so much reduced or changed from 

 the original plan of the sterna of these segments. The tip of the intro- 

 mittent organ is much simpler in structure in the families of Zygoptera. 



33. — As mentioned in paragraph 15, the presence of the ovipositor 

 in the early stages of the nymphs of Zygoptera and its absence in the 

 nymphs of Anisoptera suggest that the anisopterous appendages have 



