SOUTH AFRICAN TERMITES. 66 



«pp., it appears as a tine, slioi-t trachea, or fails altogether to 

 develop. The sub-costal trachea is often branched in Hodo- 

 termes ; it is more often simple in Calotermes and Cryp- 

 totermes. In Termes natalensis and Microternies it 

 is shortened and reduced, fig. 17G (1^1. XT), and in iUiino- 

 terines it is missing. In most of the Metatormitida) 

 examined it is either absent or very much reduced. 



The radial is sti^ongly bi-anched upon both sides in Hodo- 

 termes. The inner branches or sectors are wanting in 

 Calotermes, Cryptotermes, and Rhinotermes. In all 

 the other species except T. natal em sis, both innei- and outer 

 branches are missing or very much reduced. Thus, in the 

 case of Rhino termes there are very reduced outer branches 

 with no inner. In Termes natalensis there is a regular 

 series of outer branches, but ribs do not arise from them in 

 the final wing, and in this and certain don to termes there 

 is a tendency — possibly due to reversion — for the radial to 

 produce relatively strong inner branches ; in some wing-sacs 

 of one and the same insect these are present, in others absent. 



When present the anal trachea is much reduced in all the 

 species studied, and in T. natalensis it enters the thoracic 

 cavity and not the wing-sac. 



A further feature which may be mentioned in this con- 

 nection is the tendency for the tracheation of the wing-sac, 

 as a whole, to become reduced by the failure of the principals 

 to produce that multitude of tibril-like branches so charac- 

 teristic of the wing-sacs of the more generalised species, and 

 practically absent from those of the more specialised. 



VII. THE VENATION AND UNFOLDING OF THE 

 FINAL WING. 



Plate VI, figs. 61-69. 



Hitherto, in discussing the nymphal wing, a fairly fiat 

 organ has been under reference, and this, when cleared in 

 glycerine jelly, reveals the air-laden tracheas stretched out 



