394 NOTES OF A BOTANIST CHAP. 



of this species that the plants grow sometimes where 

 they are totally overwhelmed by the periodical floods, 

 rendering them a precarious dwelling-place for the 

 ants. This leads to the suspicion that some of the 

 sacciferous species, growing far away in the forest, 

 may have sprung originally from T. planifolia, 

 which grows on the river- banks ; and even that 

 some of the epiphyscous, anaphyscous, and hypo- 

 physcous species may be mere varieties of one 

 another, or may have had a common progenitor 

 at no very remote epoch. This and many other 

 interesting problems can only be solved when 

 naturalists shall become permanent members of the 

 fauna of Equatorial America, and not as now have Jo 

 be classed among "occasional visitants"; for their 

 solution would require observations to be carried on 

 through many consecutive years on the same spot. 



Besides Tococa, there are other allied genera of 

 Melastomes, viz. Myrmidone, Mart., Majeta, Aubl., 

 and Calophysa, DC., which have sac-bearing leaves 

 infested by ants. They are all found in the forests 

 of humble sparse growth called "caatingas," and 

 especially where the soil of white sand, or the 

 granite floor almost bare of herbs, lies low and is 

 liable to get transformed into a shallow lake in the 

 time of heavy rains, thus driving ants and other 

 insects to take refuge in the trees and bushes. Of 

 Myrmidone I gathered four species, including the 

 original M. macrospenna of Martins. They are 

 low-growing, sparingly-branched shrubs of 3 to 

 S teet ; the leaves of each pair are very unequal 

 in size, the smaller one sometimes even obsolete, 

 the larger saccate, as in the Tococa Anapkysctf, 

 but the sac always rugose as well as unisulcate ; 



