3'o ANTS. 



>cem to render the construction of carton nests altogether superfluous. 

 In the second place, certain species of Cecropia, having essentially the 

 same structure as ( '. adcnopus, are nevertheless free from ants. This 

 is true, for example, of ( . pel fata, as I have observed in Porto Rico 

 ( H)O/d }. 1 1 ere no species of Asteca occurs and the tree is almost never 

 tenanted by ants of any description, although it has well-developed 

 prostomes and distinct Miillerian bodies. It thrives on the mountains 

 of the island, even when its foliage is much eroded and perforated by 

 insects. ( )f the Brazilian species, C. lyratiloba, according to von 

 Iliering, resembles adcnopns in having Miillerian bodies and in being 

 regularly inhabited by Aztcca. The same is true of C. sciadophylla, 

 which is peopled with A. erncryi according to Forel. Both Schimper 

 and von Ihering, however, found C. hololcnca without trichilia and 

 without ants. The former author also describes and figures this tree 

 as lacking the prostomes. The allied genus Porouina is imperfectly 

 known. Rettig says that it has Miillerian bodies like Cecropia, and 

 Forel ( 1904/1 mentions Aztcca dnroifc as occurring in its twigs. 1 



The " myrmecophilotts " Rubiacese, embracing the genera Mynnc- 

 codia ( Fig. 176), Hydnophytum (Fig. 175) and Myrmephytum, with 

 about sixty species, confined to the Austromalayan region, have been 

 studied by Rumphius (1750), Gaudichaud (1826), Camel (1872), H. 

 O. Forbes (1880, 1886), Beccari (1884, 1885), Treub (1883, 1888). 

 Burck (1892), Haberlandt (1893), Karsten (1895), Dahl (1901) and 

 Rettig (1904). These plants are epiphytes on trees or rocks in hot, 

 sunny places and grow from large, bulbous stems full of cavities that 

 communicate with the outside by means of small holes. These bulb- 

 like structures are nearly always occupied by ants. Treub found that 

 the cavities arise in the very young plant and are not started, though 

 they may be subsequently enlarged, by the insects. Iridomyrmex 

 in\nnccodi(C, a subspecies of /. cordatus. is the ant most frequently 

 found nesting in these plants, but species of the same or other 

 genera have also been recorded (Camponotus maculatus, Crcmasto- 

 (jastcr diffonnis and Pheidolc jarana). In some species of Myrmc- 

 codla the bulb bristles with spines, as if for protection, but notwith- 

 standing the presence of these structures and the ants, no one has been 

 able to detect the existence of any enemy that might injure or devour 

 the plants. Treub observed that specimens grown in localities inac- 



1 In a very recent study (" Cecropia peltata nnd ihr Verhaltnis zu Azteca 

 Alfari, zu Atta sexdens und anderen Insekten," etc., Biol. Ccntralbl., 29, 1909, pp. 

 I _ I 6, 33-55, 65-77, pis. 1-5), Fiebrig is even less inclined than von Ihering to 

 accept the theory of myrmecophily among plants. He shows that Cecropia pel- 

 tata of Paraguay (apparently not the C. peltata of the Antilles!) is not protected 

 from its great number of insect and other enemies by the Azteca alfari, though 

 this ant constantly occupies its cavities and feeds on its Miillerian bodies. 



