188 MB. K. J. TILL YARD ON THE EECTAL BREATHING-APPARATUS 



clearly to be distinguished by their abundant tracheation and by the gradual 

 appearance of a darkish pigmentation localizing their position. 



Larvae of both jEschna and Anax were taken and sectioned at the seventh 

 instar (text-fig. 20, C). Also, the gill-baskets of others of the same age v^^ere 

 opened, and selected portions of the gills photographed, w^hen a very in- 

 teresting result was noticeable. In both genera the foliate type of gill was 

 completely established by the development of undulations on the original 

 leaflets, which by now project into the rectum in the form of a series of 

 definite foliae closely resembling those seen in the full-grown larva of 

 JEsc/ma, but not so deeply pigmented. But, in Aiiaa?, little bundles of 

 developing papillae were to be seen projecting outwards from the free edges 

 of the foliae. 



In order to trace back the origin of these papillae^ I took the only remaining 

 larva of Anaa? which I still possessed in the sixth instar. and dissected it 

 carefully. On opening the gill-basket I found, at the extreme edges of the 

 foliae, tiny groups of four or five papillae grouped together in a very immature 

 stage of development, with here and there a small swelling indicating the 

 beginning of papilla-formation along other parts of the leaf-edge. The 

 papillae are, at the first, small rather pyramidal tubercles, containing no 

 tracheae. As they increase in length (their area in cross-section near the 

 base remaining fairly constant), the tracheal loops lying immediately beneath 

 them are as it were sucked up out of their course into the papillae. Finally, 

 they become long enough to z-esemble the finger of a glove, with the capillary 

 loops rising up well into their interior. At the sixth instar I did not notice 

 a single papilla fully formed, and very few indeed had tracheal loops in them. 

 At the seventh instar a large number of the papillae were fairly complete, 

 others were in all stages of development. 



The further development of the typical papillo-foliate gill-type in Anaas is 

 very rapid. During the seventh instar the foliae tend to curve over poste- 

 riorly along their free edge, while numerous papillae arise at all points of the 

 curve. At the eighth instar this curved folia already bears a very close 

 resemblance to the hasal hump of the full-grown larva, and is very deeply 

 pigmented. A series of these foliae taken along the free edge of a hemi- 

 branch in semi-profile, at the eighth instar, is shown in Plate 19. fig. 8. As 

 will be seen, they differ little from the condition seen in the adult (Plate 21. 

 fig. 16). 



Phylogenetic Conclusions. (Text-fig. 21.) 



We are now in a position to review our study of the gill-basket with a 

 view to stating the phylogenetic conclusions to which it points. These may 

 be summarized as follows : — 



1. The most primitive type of gill-basket to be found in present-day 

 Anisopterid larvae is undoubtedly the Undulate Type in the Simplex Sijstem. 



