70 



HORN EXPEDITION — NARRATIVE. 



Ill iiddition to tlic ,L;allrfics I'uiiiiiiiji; along the ii,njuiKl Home of tlic tussock.s of 

 l\ircii|(iiio have their long sjiiny leaves more or less wholly eiieloseil in little 

 eylinders of sand, formed in the same vvay, to sueli an 

 extent that the whole tussock looks like a network of sand 

 tubes. In other eases thei'e were only small eyliiidrieal 

 cases of sand here and there on the spiny leaves. Each 

 of these was perliajis half an ineli or an inch long and a 

 (juarter of .in inch in diameter, and so built that tiie grass 

 blade, formed one part of the wall, a, sjiace Ijeing enclosed 

 between it and the sand. The eylindei' was always closed 

 at tiie top and had a small opening at the bottom, so that 

 if rain i-aiiu; it would not get into the clnunbers. 



Watching tiie ants, which are very small and l)iack- 

 Ijodied with yellowish feet, I saw them constantly running 

 ill and out of these chambers, juid on opening the latter 

 found that they wei(^ always Ijuilt over two or more Coccidie 

 attached to the h^af of the grass. Here, as in the case of 

 the ants tlescribed by Belt in Nicaragua, the Cocciihe abstract nutriment from 

 the leaf, ami the ants take advantage of the exudation from the body of the 

 Coccus. This arrangement is without d(jubt of advantage to both parties 

 concerned. The Coccidie gain protection from enemies, to whom they are made 

 invisible, and also from the great heat of the sun, and at the same time the ants 

 get without mucli trouble to themselves a supply of food. 



I think after examining a considerable number of tussocks of porcupine grass 

 both here and elsewhere that the; network of sainl tubes, which as above said some- 

 times cover the whole tu.ssock, always commences in the form of a number of 

 chambers specially built o\er the Coccid;u which it is (juite likely — though I had 

 not the meajis of testing it — are actually In-ought on to the leaf by the ants. Then 

 covered passages are ma.de up the leaves, leading from one chamber to another and 

 so gradually the whole tussock is enclosed. 



Ti'acing the passag(;s down to the roots the ant nest is seen to be built around 

 the latter or rather part of it as the tussocks are often of large size. Tiie nest 

 consists of a more or less conical mass of material built up of sand particles 

 agglutinated by the resin derived from tiu^ leaf sheaths with remains of the roots 

 luiming througii it. The largest nest dug up measured a foot and a half in depth 

 and about a foot acnjss at the toji which was just below the surface of the ground 

 from which it gradually tajiercd downwards. It was riddled with passages of 



