CH. IX] THE RESPIRATORY SYSTEM 167 



are so closely interwoven as to produce the appearance of a 

 complete whole. The cross-section of the spiral thread in one of 

 the larger tracheae of a Dragonfly larva is easily seen to be quadri- 

 lateral. 



The tracheal system of the Dragonfly imago is of the type 

 known as an "open" system, i.e. the spiracles are functional, and 

 air is drawn into the system directly through them. It seems to be 

 quite clear that the larva had originally an exactly similar system. 

 With the adoption of the aquatic habit, however, the spiracles 

 ceased to be useful, except on such special occasions (e.g. meta- 

 morphosis) as the larva was able to leave the water and breathe 

 air directly. Hence the spiracles have persisted in the larva, 

 although not generally functional, and often indeed quite closed 

 up. The requisite oxygen is obtained by the agency of a system 

 of specially developed tracheal gills, from which special branches 

 known as efferents pass to the main tracheal trunks. Thus the 

 tracheal system in the larva becomes more complicated than that 

 of the imago. 



Tracheal gills are developed in the larvae of Dragonflies in three 

 places : 



(i) In all Anisoptera, within the anterior portion of the 

 rectum, by specialization of the rectal epithelium and pads. 

 The whole formation makes up the beautiful structure known as 

 the branchial basket. 



(ii) In most Zygoptera, on the three anal appendages of the 

 larvae, i.e. the two cerci and the appendix dorsalis. These are 

 known as caudal gills. 



(iii) Laterally, on certain of the abdominal segments, in a 

 very few archaic Calopterygidae. These paired gills are remnants 

 of original paired abdominal appendages, of which, in other 

 Dragonflies, the only pair left are the cerci. They are homologous 

 with the gills of the May-fly larva. The larvae possessing them 

 also have caudal gills. 



Besides these recognized forms of gills, there are the three 

 rectal pads or folds of Zygopterid larvae, which probably function 

 as blood-gills. These were dealt with on p. 116, and need not be 

 further discussed. 



We shall deal with the respiratory system under five headings : 



