58 ANTS AS TKEE-GUAEDS. 



eaten by herbivorous mammalia. Delpino mentions 

 that on one occasion he was gathering a flower of Clero- 

 dendrum fragrans, when he was himself ' suddenly 

 attacked by a whole army of small ants.' ^ 



Moseley has also called attention^ to the relations 

 which have grown up between ants and two ' curious 

 epiphytes, Myrmecodia armata and Hydnophytum 

 formicaruTn. Both plants are associated in their 

 growth with certain species of ants. As soon as the 

 young plants develop a stem, the ants gnaw at the base 

 of this, and the irritation produced causes the stem to 

 swell; the ants continuing to irritate and excavate 

 the swelling, it assumes a globular form, and may 

 become even larger than a man's head. 



' The globular mass contains within a labyrinth of 

 chambers and passages, which are occupied by the ants 

 as their nest. The walls of these chambers and the 

 whole mass of the inflated stem retain their vitality 

 and thrive, continuing to increase in size with growth. 

 From the surface of the rounded mass are given off 

 small twigs, bearing the leaves and flowers. 



' It appears that this curious gall-like tumour on 

 the stem has become a normal condition of the plants, 

 which cannot thrive without the ants. In 3Iyrmecodia 

 armata the globular mass is covered with spine-like 

 excrescences. The trees I referred to at Amboina had 

 these curious spine-covered masses perched in every 



' Scientijic Lectures, p. 33. 



* J\"^utes by a Katuralist on the ' Challcmjcr,' p. 389. 



