ON SOME AUSTRALIAN TERMITES. 371 



were fully developed on 15tli February, but on 8th March, up to which date only 

 light rain fell, these forms were still present in all the nests examined. On this 

 date the tips of the wings showed marked damage due to prolonged occupancy of 

 the parent nest. Similar conditions prevailed in the nests of Hamitermes perplexus, 

 sp. n. Heavy showers fell on the night of the 9th March and throughout the day and 

 night of 10th. Swarming of H. perplexus took place during the afternoon and evening 

 of the latter date. The swarming of D. silvestrii was not observed, but probably 

 took place about the same time, since none of the nests contained imagines on 

 14th March. 



It is not intended to discuss in detail here the numerous other forms of life which 

 have been found in termitaria, but brief mention may be made of one species which 

 plays an important part in the economy of two species of termites dealt v/ith 

 in this paper. At a very rough estimate it may be said that 80 per cent, of the 

 termitaria of D. silvestrii and H. perplexus are invaded and permanently occupied 

 by the verv common and widely distributed ant, Iridomynnex sanguineus, Forel, 

 which is particularly abundant on the low-lying country in the vicinitj' of TownsviUe. 

 The termitaria are entered by means of holes burrowed into the walls (PL xii, fig. 2), 

 in and out of which pass endless streams of ants in their journeys from one nest to 

 another. If a termitarium is cut open vertically, it will be found that the ants 

 have greatly enlarged the original galleries so as to form large flattened chambers 

 in tier upon tier, until finally the greater part of the structure is in their undisputed 

 possession. The floor of each cell is thickly covered with the eggs, larvae and 

 pupae of the invaders, and immense numbers of ants throng all parts not actually 

 in possession of the termites. As the ants extend their sphere, the termites are driven 

 back from chamber to chamber and destroyed, until but a few stragglers are left. 

 The complete, or nearly complete, occupation of a termitarium is evidently a matter 

 of time, during which the advance is being constant^ delayed by the termites walhng 

 up their galleries and passages as they retreat. The remains of the dead termites 

 in the chambers occupied by ants show clearly that the nests are not attacked merely 

 to provide a dry and safe shelter, but that the original occupants are used as food. 

 Immediately the walls are broken with the pick the ants swarm out in countless 

 thousands, destroying and carrying off the dislodged termites, crawhng up one's 

 legs and attacking one's hands, head or any skin surface to which they can gain 

 access. Others of their kind gather from all directions to take part in the onslaught, 

 until the nest and the surrounding ground is a seething mass of insect hfe. Under 

 these conditions a close examination of the nest or its occupants is impossible, and 

 it is only by finding an ant-free nest that one can hope to investigate its interior. 

 Within a few minutes of the nest being broken into all the neighbouring ant- 

 infested termitaria of these two species will be found to contain the bodies of freshly 

 killed termites, while files of ants pass to and fro so long as a termite remains exposed 

 to attack. Plate xii, fig. 1, shows ant tracks made on the surface of the ground 

 approaching a mound of H. perplexus. It is a remarkable fact that the mounds 

 of a certain species of Eutermes, which are very common amongst those of the 

 Drepanotermes and Hamitermes, are never molested by Iridomyrmex. 



The imagines of the beetle, Cryptodus grossipes, Fairm., have been found in the 

 cells of D. silvestrii, and Mandalotus germinatus. Lea, has been taken on two occasions 

 in the nests of H. perplexus. Their relationship to their hosts is not known. Bubaris 

 indemnis, Pascoe, has been found under the walls of termitaria of several kinds, 

 but they appear not to come directly into contact with the termites. 



Ciliates (? Trichonympha), which occur in vast numbers in all the workers 

 and soldiers and in many imagines of Mastotermes danviniensis, Frogg., in 

 TownsviUe, have not been found in Drepanotermes silvestrii, or in any other 

 local species of termite. 



