IN THE LARV^ OF ANISOPTEKID DRAGONFLIES. 149 



ensure a complete circulation of aerated water around each single one o£ 

 them. [Compare the function of the tubercles found on the lamellae 

 of Libellulidse, for which see p. 162.] 



Arrangement of the Tracliece. — We have already shown how the separate 

 gill-tracliece divide each into five or six axial or papillar tracheae, each of 

 which then forms the axis of a separate papilla. The capillaries are almost 

 entirely confined to the papillte, only a few stray ones being detectable here 

 and there in the pigment-layer. Very possibly these may have become dis- 

 placed or pulled out from the bases of papillse during the process of dissection. 

 The capillaries arise along the whole course of the papillar trachea. They 

 are of excessive fineness, and can only be clearly detected under a high 

 power (text-fig. 4, tl). They run almost parallel to the main axis of the 

 papilla. But, whereas the papillar trachea itself does not reach the extreme 

 tip of the papilla, all the capillaries, on the other hand, run right up to the 

 very tip, where they curve sharply over in small loops and run back to 

 rejoin the papillar trachea near its distal end. 



In the everted gill-basket, the papillas are seen to project in a slanting 

 position directed somewhat posteriorly. The angle between the papillar axes 

 and the longitudinal axis of the gill-fold varies between 130° and 140°, being 

 greatest at the posterior end of the gill-basket, where it terminates in a series 

 of long papillse projecting beyond the end of the basal fold. 



Origin of the Papillate Type. — The type seems to have been called forth in 

 the course of evolution by the assumption of peculiar habits on the part of 

 the larva. Whereas the larvae of the genus Austrogomphus all live on the 

 river-bottoms, hidden in loose debris, or buried only partially so that their 

 heads and' anal ends project freely into the water, the larvae of Hemigomphus, 

 on the other hand, hury themselves completely in pure sand well below the 

 surface of the river-bed. In this they burrow or remain motionless all day, 

 only emerging at night to look for food. The anal opening is protected, as 

 in Petalura, not only by valves, but by crossed hairs, which allow of the 

 inspiration of water from the damp sand, while keeping the sand itself out. 

 When the larva is buried in the sand, the process of respiration must neces- 

 sarily be very slow, so that it is of the utmost importance that the larva 

 should command the most efficient and complete extraction of oxygen from 

 the small quantity of available water. This objective is certainly marvellously 

 well ensured by the papillate type of gill, which can be easily seen to have 

 two very distinct advantages over the undulate type : — 



1. As a result of the splitting-up of the gill-tracheae, a very much greater 

 number of capillaries is developed. 



2. As the capillaries lie in groups isolated in separate papillse projecting 

 far into the rectal cavity, their ability to absorb oxygen is very much 

 increased, especially as the water bathes each papilla on all sides, whereas in 

 the undulate type it can only course along the two sides of each fold. 



LINN. JOURN. — ZOOLOGY, VOL. XXXIII. 11 



