58 CLASSIFICATION Of INSECTS. 



This argument is in our opinion very strong, espec- 

 ially when it is considered that the existing Thysanura 

 and all Thysanuriform larvae, even in the earliest larval 

 stages of aquatic forms, have tracheae before they 

 acquire external gills. Tracheae are exclusively inter- 

 nal air-breathing organs, and this indicates that the 

 primitive Thysanuroid ancestor, if it were similar to 

 Thysanura and the Thysanuriform larvae, must have 

 been unprovided with gills or any means of breathing 

 in the water, but was at first provided with air-breathing 

 organs, and consequently must have been a truly ter- 

 restrial and wingless air-breathing animal. 



This view has also another important corollary, 

 namely, that existing aquatic larvae are not in any 

 sense primitive, but that their adaptive and peculiar 

 characters — gills, etc. — are secondary specializations 

 and that they themselves were derived from ancestors 

 having purely terrestrial habits and organs. In other 

 words, the insects of the existing faunas belong to an 

 exclusively terrestrial type, even those now living in 

 the waters, either during their larval or in their adult 

 stages, having been evolved from air-breathing terres- 

 trial forms. 



The larvae in May-flies respire through the skin dur- 

 ing their earlier stages, and do not at first have any ex- 

 ternal gills. In most stone-flies the larvae are destitute 

 of external gills throughout life or have only external 

 respiratory filaments : the internal tracheae are, how- 

 ever, developed at very early stages even in embryo. 

 The giUs have not a fixed form or position in the in- 

 sects, as in true aquatic types in other subdivisions of 

 animals. The gills of Mollusca, the respiratory water- 



