BY R. J. TlLLYARD. 405 



members has branched off early or late from the main reduction- 

 movement, has stopped far short or has travelled far along the 

 road to this end, so will it be found to possess a greater or less 

 portion of the original characters of the Odonata. The great 

 reduction of the preenodal area in the Zygoptera is the best illus- 

 tration of this statement. Starting with a prasnodal area roughly 

 equal to half the wing-surface, we find this portion of the wing 

 gradually reduced both in length and breadth, so that the number 

 of antenodals rapidly fell below that of the postnodals, and finally 

 became reduced to two; that being, as far as we can see, the 

 irreducible minimum beyond which evolution has not been able 

 to go, without sacrificing the species altogether. Now de Selys 

 divides the whole of the Zygoptera up into two families, accord- 

 ing to whether they possess more than two, or only two, ante- 

 nodals. Neither the utility nor tlie naturalness of this was 

 questioned, until forms like Thaiimatoneura were discovered, 

 possessing all the cliief characteristics of the Agrionidre, but 

 with moj-e tha7i two antenodals. The need for attention to, 

 and improvement in, the classification then became evident. 

 The proposition may be placed before us, in the light of our 

 present knowledge, as follows : — Assuming the Zygoptera to 

 be ail asthenogenetic group, and hence, that species possessing 

 only ttvo antenodals have all descended from forms possessing 

 more than two antenodals (not, of course, the same forms as now 

 exist, with more than two antenodals), let there be a; + y separate 

 lines of descent in the .Suborder possessing representatives at the 

 present day. Of these, suppose that x have reached the stage of 

 possessing onh^ two antenodals, while y have not yet reached that 

 stage. Then, if we adopt de Selys' classification, we should have 

 X groups of Agrionidce, and y groups of Caloptei-ygidce. But 

 from any of the x lines of descent that led to Agrwnidff, offshoots 

 may have proceeded which found their natural position of equili- 

 brium in some form still existing with more than two antenodals. 

 Let there be z of these. By de Selys' classification, these z are 

 included in the y Calopterygida'; whereas it is quite evident that 

 they are more closely related to some of the x forms in the 

 Agrionidce. 



