158 MR. R. J. TILLYARD ON THE RECTAL BREATHING-APPARATUS 



leaf form seen in ^sclina, forms a transversely-elongated oval hmnp, borne 

 upon a slender stsilk whose axis is a secondary efferent trachea. It is this 

 character of the nan^oio stalk which is typical of the foliate type, and 

 distinguishes it from all others. In Anax, these leaf-stalks are not quite 

 so long as in JEschna, and consequently the leaves do not project so far into 

 the rectum. There are twelve leaves or humps in each row in Anax 

 papuensis. Rarely, thirteen or fourteen occur. The positions of the leaves 

 or humps in Anax correspond exactly to those of j^schna, except that they 

 are elongated in a direction at right angles to the longitudinal axis (those of 

 y^schna are elongated at a somewhat less angle). The relationships of the 

 rows of leaves in consecutive hemibranchs are the same as those already 

 explained in u:jEsclina. Fat-globules and dark pigmentation are present as 

 in ^schna, the pigmentation being even deeper than in ^schna, and of a 

 purplish-black colour. 



Structure of the Papillce. — The most striking feature in the gills of Aiiax 

 is the very remarkable formation of small papillae all over the free surface of 

 each leaf or " hump," which is called on that account by His the " basal 

 hump " (die basalen Hocker) — a very appropriate term, if we remember that 

 the hump is " basal'' to the papillae, but "distal" to the efferent trachea 

 forming the axis of the stalk. Oustalet studied these papillae very carefully, 

 and gives an accurate description and figure of them. They arise in large 

 numbers as small finger-like eversions of the wall of the whole free surface 

 of the hump. They are not homologous with the papillae already described 

 in Hemigomphus, but have been, like these hitter, called forth quite 

 independently, in the process of evolution, by the same necessity for 

 obtainino; a better extraction of oxvo-en from the water in the rectum. 



Each papilla is of cylindrical shape, with a rounded tip. In proportion, 

 they are shorter and thicker than the papillae of Hemigomphus. Within the 

 papilla there is no axial tracJiea, but only a small number of fine capillary 

 loops, which have been, as it were, drawn or sucked up into the papillar 

 eversion, out of their original course along the surface of the hump, so as to 

 complete their loopings within the papilla. Oustalet estimated that each 

 hump carries a hundred of these papillae. The number of capillary loops in 

 each separate papilla of Anax papuensis varies from five to eight, the average 

 being six (text-fig. 10). 



In the cuticle of these papillae there are no transverse chitinous rods such 

 as are to be seen in those of Hemigomphus. At their rounded ends^ however, 

 there are a number of little curved chitinous hooks, which probably serve 

 the same purpose as the rods. 



Owing to the comparative thickness of the basal humps and the density of 

 their pigmentation, they are even more unsatisfactory for photography than 

 are the foliae of ^schna. Plate 19. fig. 8 shows a profile view of one hemi- 

 branch of a half-grown larva, while fig. 7 of the same Plate shows the front 



