TKRMITl': EC()N;)MV 



6 1 



■seme species have l)een erected upon winged forms alone and 

 others ui)on the soldier caste only. Of ccmparatively few kinds 

 are the win^-ed and soldier castes definitely associated, a matter 

 which is onlv jDOssible when the op])ortunity occurs to secure 

 both castes from the i^arent nest. Ag-ain the list is without doubt 

 amplified, because many of the criteria relied upon for the difiter- 

 cntiation of species are not so reliable as thev have been taken to 

 be. 



Upon the other hand the country as a wdiole has been so 

 little surveyed for termites that there must remain a number of 

 kinds still unknown to science. For example, I know of no 

 species having been recorded from Basutoland or Gri(.|ualand 

 East. Very few indeed have been recorded from the Caj^e, and 

 but two of some seven that I have obtained from the Orange Free 

 State haAC been jjreviously mentioned from that Province. 



It occurs tc me that a further knowledge of distribution mav 

 show that environment has a marked efl:'ect upon the variation of 

 several species, in both form and habit. 



The conditions under which the nests of certain rermites and 

 shrubs and trees are associated have led me to the conclusion 

 that '' park-formation," as we know it in this country, is entirely 

 due to these insects. It is extremely probable rhat, apart from 

 untoward circumstances, a colony of termites exists for a consi- 

 derable period ; some species thriving much longer than others. 

 What that period is, cannot be said for any kmd, still there is 

 some evidence to show that Haviland's natalciisis thri^■es for a 

 decade or more. But it must be conceded that ultimately the 

 colony dies out. In the case of mound-building kinds, as long as 

 the colony exists, the natural weathering is repaired, but when it 

 dies the mound erodes and incidentally forms a piece of culti- 

 vated ground. Ample evidence has been collected to show that 

 these old sites are again and again re-colonised, and with every 

 successive occupation they increase in area, if but little in height. 



Both the weathering of the occupied mound and the erosion 

 of the deserted mound afiford shrubs and trees, that cannot gain a 

 foothold in normal grass-veld, a suitable spot in which to thrive. 

 How their seeds get to the mound matters little ; how the variety 

 of plants becomes so great as it is, is readily understood, because 

 such park-formations are only met with in grassy countrv more 

 or less continguous to timber belts and in and about what we in 

 South Africa call bush-veld. 



That certain termites derive some direct advantage from the 

 presence of trees and shrubs u]^on their mounds is evidenced by 

 the fact that whilst they do not permit grass to grow upon the 

 mound, they do not interfere with deep rooting and stronger 

 growing plants. To a large extent the advantage is against the 

 full effect of storms, and possibly the root svstems of the trees 

 afford some barrier to the predatory incursions of the aard- 

 vark. 



Up to the i)resent I have not been able to locate a clump of 

 trees of typical park-formation in which the soil is not raised 



E 



