THE TRA CHEAT ION OF WINGS 49 



radial connection (c-r) with the trunk of the vestigial metathoracic spiracle 

 {v. sp'i). The posterior wing has, therefore, forsaken the best source of air 

 supply and retained a connection with a trunk which no longer has an 

 opening to the exterior. 



In all the cases where the wing tracheation has been reduced it is the 

 cubito-anal trachea which has been lost. If the air supply is the factor 

 causing the one connection to be lost and the other connection to be retained 

 it is not easy to understand how it operates in the case of the metathoracic 

 spiracle, for here the cubito-anal trachea to the front wing is lost and the 

 costo-radial trachea to the hind wing is retained. 



To return to the case of the Odonata where the air-supply theory seems 

 to find its best support, the posterior connection to the front wing should 

 be considered. If the accessory cubito-anal trachea (Fig. 34, a. cu-a), 

 which is enlarged, connected with the dorsal longitudinal connective {d. It) 

 as Tillyard (/. c.) thought, it would be favorable to the air-supply theory, 

 but it connects with a small transverse connective between the two dorsal 

 longitudinal tracheae. This makes an indirect course for the air to follow 

 in going to the wing and the condition would seem to be more easily 

 explained as having arisen as an enlargement of a simple anastomosis of the 

 accessory cubito-anal tracheae of both sides, a condition found in Melanoplus 

 (Fig. 33) . The union of these tracheae might easily have anastomosed with 

 a transverse connective to form a course which later became enlarged 

 because it was the least obstructed route for the air to follow to the wings. 



It should be remembered that there are two anterior tracheae leading to 

 the wing pad {c-r and a. c-r), each of which is about half the size of the 

 posterior trachea (a. cu-a) and the two together must supply about as much 

 air to the wing as this posterior trachea does. 



The condition of the leg tracheae in the Odonata is very instructive in 

 the study of the values of the air supplies in the thorax. The legs, like the 

 wings, each have two tracheal connections which typically derive their air 

 supply from the same sources as do the tracheae to the wings. But in the 

 Odonata the posterior stem of the leg tracheae, which should offer the more 

 direct course for the air to follow in passing to the legs from the gills at the 

 posterior end of the bod}', is greatly reduced, when the thoracic angle is 

 great, while the anterior stem with its less direct cotuse is enlarged. This 

 would indicate that the directness of the air supply has little to do with the 

 modification of the leg tracheae in the Odonata. 



The belief that the migration of the medial trachea along the transverse 

 basal trachea toward the cubito-anal group of tracheae might be due to the 

 better air supply to that group was referred to in the introduction of this 

 paper. In view of the fact that the basal connection of the posterior group 

 of tracheae is the one to be lost in all cases where there is a reduction of the 

 tracheal connections, it seems more reasonable to ascribe this migration of 



