If we examine the tracheal gills of the smaller dragon-fly (Agrion), or the May-flies, or Sialidæ, or Perlidæ, or Phryganeidæ, we see that they are developed in a very arbitrary way, either at the end of the abdomen, or on the sternum, or from the pleurum; moreover, in structure they invariably have but a single trachea, from which minute twigs branch out;[30] in the wings there are five or six main tracheæ, which give rise to the veins. Thus, in themselves, irrespective of their position, they are not the homologues of the gills. The latter are only developed in the aquatic representatives of the Neuroptera and Pseudoneuroptera, and are evidently adaptive, secondary, temporary organs, and are in no sense ancestral, primitive structures from which the wings were developed. There is no good reason to suppose that the aquatic Odonata or Ephemerids or Neuroptera were not descendants of terrestrial forms. To these results we had arrived by a review of the above-mentioned facts, before meeting with Fritz Müller’s opinions, derived from a study of the development of the wings of Calotermes (Fig. 158). Müller[31] states that “(1) The wings of insects have not originated from ‘tracheal gills.’ The wing-shaped continuations of the youngest larvæ are in fact the only parts in which air tubes are completely wanting, while tracheæ are richly developed in all other parts of the body.[32] (2) The wings of insects have arisen from lateral continuations of the dorsal plates of the body-segments with which they are connected.”