( 156 ) 



primitive nests. Besides building between (he branches and the roots 

 and in the pitcher leaves of Dischklia Rnfjlesiana, these insects 

 make their nests from enlargements of the tunnels, which diverticula 

 are built with the same bitten-off particles of bark as the tunnels 

 themselves. They prefer to make these enlargements on the lower 

 side of a lateral branch, where such a branch emanates from a 

 thicker one, but also on the lower side of the point, where two 

 thinner twigs come off together. It would of course have been very 

 peculiar, if the seeds had become attached in laiger numbers just 

 at these places and had afterwards been covered by the ants. By 

 these discoveries the problem was solved and later journeys and 

 observ^ations confirmed us more and more in our view. 



When this fact has once been noticed, it takes very little trouble 

 to collect a few hundreds of seedlings or more witliin a short time. 

 Although germination takes [)lace very rapidly, the further growth 

 is, at least in the beginning, very slow. In the teak forest of 

 Tempoeran we found very good evidence that we had not been 

 mistaken. Among thousands of pitcher leaves of D. Rnfjiesiana we 

 found a single specimen which had a narrow slit in its wall. This 

 pitcher was inhabited by a lai'ge number of ants, which had divided 

 its lumen into various chambers and passages. The j)eculiar thing 

 was that from this slit thei'e issued the two cotyledons of a seedling 

 of D. nummularia, while the hypocotyl axis was long drawn out 

 and grew out from the in.uermost part of the ant passages. 



The ants often live in large uuuibers on the small trees of 

 Protium javanicum, which have been planted along the village 

 roads and owing to continual pruning have assumed the aspect of 

 pollard-willows. The widened, often half mouldered portions aj)pear 

 to be excellent nesting-places for the ants. Dbichidia's indeed grow 

 on these small trees in large numbers. We do not know why, but 

 in various places one finds numerous ants on one side of the road 

 and none or hardly any, on the other side. Without exception there 

 were in such cases numerous Disclddia'.s on one side of the road 

 and few on the other. It is indeed in these places that the dissemi- 

 nation by ants is readily observed. We were moreover able to take 

 a large number of photographs which will be reproduced in the 

 more detailed publication. 



This point in the investigation was reached very rai)idly but the 

 principal evidence was still wanting. We had never directly observed 

 the ants dragging away the seeds. Expei'iments made in our garden 

 were unsuccessful, for the Iridomyrmex was always driven away 

 by other species of ants. But after a long and abortive search we 



