j^ EVOLUTION 341 



Belmont, New South Wales, and other insects relegated 

 by Tillyard to an extinct order Paramecoptera, " evidently 

 representing the common ancestors at that period of 

 the three orders Diptera, Trichoptera, and Lepidoptera." 

 While there is as yet no definite fossil evidence of the 

 existence of Trichoptera until the Necrotauliidae from 

 English and European deposits of Liassic age, comparative 

 study leaves little room for doubt that the Trichoptera 

 arose in Triassic times as a primitive side branch from 

 the stem that developed into the great order Lepidoptera ; 

 the statement often made that caddis-flies are directly 

 ancestral to moths and butterflies is misleadingly simple. 

 The " record of the rocks " is disappointingly incom- 

 plete with regard to the history of the Lepidoptera ; 

 apart from problematic remains of Jurassic age there are 

 no fossil Lepidoptera until the Tertiary (Cainozoic) 

 era is reached ; then in the well-known Miocene beds of 

 Florissant in Colorado and Oeningen in Baden, occur 

 a variety of Lepidoptera belonging to existing famiUes. 

 From comparison of the two sub-orders of Lepidoptera it 

 appears certain that the primitive Homoneura preceded the 

 now dominant Heteroneura in the time of their development. 

 The Paramecoptera appear also to have been ancestral 

 or approximately so to the Neuroptera. Of the two 

 neuropteran sub-orders the Megaloptera, with their com- 

 paratively primitive larvae and simpler wing-neuration, are 

 clearly older than the Planipennia ; we find that while the 

 former are represented by fossil wings {Triadostalis) from 

 the Lower Trias of Germany, the latter have no known 

 representatives until Archepsychops and its allies from the 

 Upper Trias of Queensland. There are still very great 

 gaps in our knowledge of these fossil insects ; when we 

 remember the delicacy of their build and the improbability 

 of their preservation in large numbers in water-formed 

 deposits, this " imperfection of the record " is not sur- 

 prising. It is all the more satisfactory that such information 

 as we have from fossil insects confirms so definitely the 

 conclusions drawn from a study of the structure and life- 



