AN ESSAY ON THE CLASSIFICATION OF INSECTS. 191 



mandibulate, orders combined. They have their origin from the 

 common stock, but were always haustellate or emandibulate in 

 all stages, forming the first and lower of my main divisions. 

 With the development of this branch, after its distinctive feature 

 became established, I have nothing to do at present. It seemed 

 adapted for variation in special lines only, and, as the method 

 of feeding was practically fixed from the beginning, there is a 

 remarkable similarity in mouth parts throughout. 



The maadibidata possessed much greater powers of variation, 

 and a mouth structure in which all the parts were developed and 

 capable of modification, containing possibilities of much greater 

 range in obtaining food. They lived therefore under all sorts of 

 conditions, in all sorts of media, and all kinds of modification 

 were produced; some of them short-lived, adapted only to 

 surroundings then existing; others with greater possibilities, 

 that exist to the present time. 



The first mandibulate insect had the thoracic segments 

 similarly developed, all of about the same size, and each of them 

 free ; but the advent of wings gave opportunity for radical 

 divisions. I have no desire to go into details here more than 

 necessary to explain my views of classification, hence will not 

 pretend to account for the origin or development of wings. They 

 did appear, however, and independently at several different 

 points. In all cases the wings were net-veined or neuropterous 

 in type, a peculiarity which is explicable if the venation be con- 

 sidered of a tracheal origin. With the appearance of wings 

 many divergences in habit were made possible, and new types 

 began to appear. Three main lines branched almost simul- 

 taneously from the common stock, each of them fairly well 

 marked from the beginning, retaining its peculiarities and even 

 intensifying them in all future subdivisions to the present time. 

 In the first of these the prothorax, bearing no wings, became 

 separated from the other rings and movable, or in a sense domi- 

 nant. In both the others it tended to a reduction in size or to 

 become agglutinated with or united to the others. In a general 

 way it may be said that the series in which the prothorax is free 

 is lower in the scale of development, as retaining a more primi- 

 tive type. The orders belonging to this subdivision or branch 

 are the Dermoptera, Coleoptera, Plecoptera, Platyptera, and 

 Orthoptera. 



[The orders placed in the first division are then discussed in 

 detail, and remarks made on phylogeny.] 



The second branch from the Thysanuran stem started with 

 all the thoracic segments nearly equally developed. While the 

 prothorax was of good size and in the lowest forms quite free, 

 yet the tendency was from the very start to unite it at its base to 

 the other thoracic segments. In this series it is alwnys fairly 

 weSI developed, sometimes even very long ; but it is always 



