OF CERTAIN TERMITE NESTS. 193 



to prevent the raids of predatory (true) ants. Sections of nests 

 are very common, especially on roadsides, but in no case is 

 there any connection between the exposed chambers and the 

 still-inhabited part. If the comb is left in an exposed 

 chamber, a fresh wall is built in the position of the old one : 

 if only a small fragment is left, this is enclosed by a new 

 chamber built within the former one ; but if all the comb 

 is removed, the termites simply stop up the small passages 

 which lead to an adjoining chamber. Doflein's explanation 

 fails also in the numerous cases in which the nest is entirely 

 underground. 



There is no apparent relation between the number of chim- 

 neys and the size and situation of the nest, species of termite, 

 &c. Underground nests of Termes redemanni consisting of 

 only a dozen chambers may have three or four chimneys, 

 while mound nests of the same species, six feet high with an 

 extensive underground portion and in the same locality, have 

 only one. In general, mound nests have the greatest number 

 of chimneys in their earlier stages, when the number of their 

 chambers is comparatively small and the hill is inconspicuous. 

 This lends support to the view that the chimney is merely a 

 permanent scaffold to guide the direction of further additions 

 to the hill. Any ventilation which takes place is more pro- 

 bably due to an exhaustion effect produced by the wind 

 blowing across the opening of the chimney. , 



There are no doubt narrow passages extending underground 

 for long distanc.es from the subterranean part of the nest, 

 though these have not been actually traced. It is easily seen 

 that this must be the case, since the narrow covered galleries 

 of the termites appear on the surface of the ground or on the 

 trunks of trees at long distances from the hill. Except when 

 the winged termites leave the nest,* the insects do not use 

 the chimneys as a means of exit, but advance underground 

 until they reach dead wood or other suitable food, when 



* See Appendix. 



