119 



after the process of excavation the cortex and tips of the thorns 

 harden and turn brown, and the ants take up their dwelHng in 

 the hollow structure. A single branch may display in sequence 

 the various stages in this excavation and habitation, the old 

 thorns at the base being ñlled with ants and their brood, the more 

 distal thorns completely hollowed out, and only just tenanted, 

 and the green thorns at the tip of the branch still intact or with 

 merely the beginnings of an aperture or a small excavation in 

 the pulp itself. 



It occasionally happens that the ants overlook thorns which 

 have reached the right stage for excavation. These nevertheless 

 mature and turn brown exactly- like the inhabited thorns, and 

 when cut open are seen to have become hollow through a drying 

 up of the pulp. It is evident, therefore, that unless the ants 

 utilise the pulp as food, they are really wasting their time and 

 energy in excavating the green thorns, since they would achieve 

 the same results much more easily by boring through the cortex 

 of old thorns. They would then merely have to remove the few 

 fibrous remnants of the pulp, and the thorns would be ready for 

 habitation. Such behaviour would, of course, require a greater 

 initial effort in perforating the harder cortex of the old thorns. 



The structure of the Beltian bodies has been carefully studied 

 by Meneghini and Savi (1884), Franxis Darwin (1877), '^^id 

 SCHIMPER (1888), who all agree in regarding these peculiar 

 structures as the homologues of the serration-glands on the leaf- 

 borders of many other plants. At first I had some difficult}' in 

 finding the Beltian bodies, because I looked for them on large 

 trees, from which they had been removed by populous colonies, 

 of ants ; but later I detected them readily. Only on one occasion, 

 however, was I fortunate enough to see the ants in the act of 

 collecting them. This was while I was walking in the out- 

 skirts of Patulul, along a road which was bordered with a hedge 

 of Erythrina trees. Among these stood two A. cornígera bushes,, 

 about 8 ft. a])art, with their trunks connected by barbed wires, 

 along which were jxissing processions of Ps. fulvcsccns workers,, 

 each bearing a minute yellow body in its mandibles. Closer 

 inspection showed that one of the trees was peoi)led by a large 

 colony of Pseudomyrmas, and that they had just discovered, du 

 the young leaves of the other uninhabited twv, an abundant 



