XVI] THE GEOLOGICAL RECORD 307 



the larvae of Protephemeroidea and Protodonata first began that 

 series of adaptive changes which finally led them to adopt a 

 purely aquatic mode of life. 



The Mesozoic Record. 

 A. The Liassic Dragonflies. 



With the close of the Palaeozoic Age, the Protodonata pass 

 from our ken, and the Giant Age of Insects gives place to a period 

 of great activity, in which reduction in size begins to work hand 

 in hand with specialization in structure and function. The record 

 of the Trias proper is very poor in insects, and no Odonata are 

 known from it. But, in the Lias, the curtain is once again raised. 

 We see before us a Transition Period, in which the giants, already 

 considerably reduced in size, are either dying out, or evolving 

 to higher forms. 



Three Liassic beds yield fossil Dragonflies. These are the beds 

 of Cheltenham and Dumbleton in Britain, and Dobbertin in 

 Mecklenburg. The first belongs to the Lower, the other two to 

 the Upper Lias. The fauna of all three is closely similar, and is 

 remarkable for the occurrence of an undoubted Gomphine, 

 Gompho'ides brodiei Buckmann. This establishes the first con- 

 nection with present-day Anisoptera. For the rest, putting 

 aside a number of incomplete wings which were very probably 

 also Anisopterous, but in which the base of the wing is missing, 

 there remains that remarkable Dragonfly fauna called by Handlirsch 

 the Anisozygoptera. These consist of three separate subfamilies, 

 the Architheminae, Tarsophlebiinae and Heterophlebiinae. Judged 

 by the condition of the radius, which is unbranched, and by the 

 absence of any true triangle-formation, these all seem to be 

 undoubtedly Zygoptera of an early unreduced form. As such 

 I propose to treat them, placing them as subfamilies of the 

 Calopterygidae and Lestidae. 



The Architheminae 1 (fig. 157) had a complete arculus, but 

 suffered from weakness in the region of the quadrilateral, which 

 seems to have tended to become merely an area of undifferentiated 



1 = Diaslatommidae Handlirsch. Diastatomma Giebel is preoccupied by 

 Diastatamma Brauer, a genus of recent Gomphinae. 



202 



