IN THE LARV.E OF ANISOPTERID DRAGONFLIES. 175 



rounded or elliptical shape and deep orange-brown colour, such as are readily 

 noticeable in the gill-baskets of the Libellulidge. The pads do not, in fact, 

 show np until the basket has been sectioned and stained. It is clear then 

 that they are widely diffused structures forming a more or less general 

 support to the gill-folds. Judging by their occurrence in a long series of 

 cross-sections, I conclude that they support the main longitudinal folds 

 alternately on either side, sending off side-branches along the anterior sides 

 only of the cross-folds. In this connection it is important to notice that, in 

 ■the Libellulidfe, the localized oval pads occur only on the anterior sides of 

 the gill-lamellse. Also, the form of the basal pad and its nuclei in the 

 Gompldnoj and Petalnra (apart from its actual size, shape, and position) is 

 very closely similar to that of the Libellulidge. These two points are of great 

 importance when we come to consider the homologies of the gill-lamellge in 

 Libellulidse. 



Passing on now to the basal pads found in those ^schnince which have 

 gills of the Foliate Type (^sclma and Anax), we have to notice two very 

 important points of difference between them and those just studied above. 

 Plate 22. fig. 2Q shows a very typical cross-section taken from a foliate gill 

 near the posterior end of the gill-basket in ui^Esclma hrevistyla. The out- 

 standing feature of this section is the occurrence of basal pads on both anterior 

 and posterior faces of the folia. Of these, tluit on the anterior face is the 

 larger and longer, extending through a greater number of sections and 

 showing everywhere a greater length along the gill-base. The secoiid point 

 is the more rounded form of the meganuclei {mi^). These, though nearly 

 always distinctly oval in transverse sections of the pad, are not of the 

 elongate-oval form seen in the pads of Austrogomphiis. Correlated with 

 this rounder form of the meganuclei is the greater narrowness of the pad 

 compared with its length, when contrasted with that of Austrogomphiis. 

 Micronuclei are also present, of the usual size and form. 



In ^-Eschna no part of the pad lies along the rectal circumference, but 

 the whole mass stands up as a su[)port for the base of the gill. Between the 

 two pads lies a large mass of hypobranchial tissue (%), while a very distinct 

 cavity (s/:>) is often to be seen between the upper ends of the pads. At both 

 ends the pads narrow to pass into the epithelial syncytium, which is very 

 deejjly j^if/mented (pg) . 



The last type of basal pad to be studied is that found throughout the 

 Libellulidae (Plate 22. fig. 27). This type differs from all those studied above 

 in its intense localization into the form of separate oval or rounded swellings 

 situated at the base of each lamella — a localization which is evidently 

 correlated with the separation of the gills into distinct lamellse. 



Sadones, who studied this form in Libelhda, made a peculiar error in 

 stating the position of the pads on the lamellse. He states that they lie near' 

 the axis of symmetry of each double series of lamellse. By reference to 



