939 MODERN CLASSIFICATION OF INSECTS. 
cently proved exceedingly troublesome, infesting the houses in some 
parts of London and the suburbs, Brighton, &c., to such an extent, 
that the inhabitants have been compelled to quit their abodes. (See 
Bostock, in Trans. Ent. Soc. vol.ii. p. 66., for various details and 
experiments for its destruction.) 
The account given by Madame Merian of the annual visits of im- 
mense swarms of the visiting ant (Atta cephalotes) from house to 
house, in South America*, and of their habit of forming large 
troops, each individual carrying a piece of a leaf in its jaws, was 
long considered fabulcus, but has since been fully confirmed by 
Homberg, Smeathman, Hancock, Stedman, and Lund, the two last- 
named authors having been eye-witnesses to the entire defoliation of 
a tree by this species, which is thence called the parasol ant in To- 
bago. Lund has particularly described their mode of operation, and 
has also observed these marches extended through several days. 
A species of this family, sufficiently common in France (F. rufes- 
cens Latr., forming the genus Polyergus), constitutes a remarkable 
exception to the remainder of the family, in respect to its habits. 
This species, which Huber names the Amazon ant, is distinguished by 
the structure of its mouth, provided with slender simple jaws ; where- 
by it is rendered incapable of constructing its nest, and attending to 
the duties of the community (which are in a great degree performed 
by the latter organs). They are therefore under the necessity of form- 
ing themselves into large armies, and of attacking the nests of Formica 
fusca and cunicularia, their object being to carry off the pupa; the 
insects hatched from the latter acting, in all respects, as their slaves ; 
and as they are brought to the imago state within the nests of the 
Amazons, they do not feel the desire to quit their masters, but labour 
for the support of their abode as though it were their own ; increasing 
* M. Lund states that he never observed a species of the restricted genus Formica 
migrating, or marching in close columns, in Brazil; and that the migratory species, 
and those which form these compact columns, belong to the section which have the 
abdominal peduncle formed of two nodes, and the antennez uncencealed. 
+ The fact of these Amazons carrying off only neuter pup seems to me to offer a 
more striking instance of instinct; for were they to introduce a single male or 
female pupa into their own nest, the consequences may be easily conceived. As 
it is, the proceedings of these neuter slaves, acting for their masters and their 
progeny with as much tenderness as they would exhibit to their own species, 
seems to prove that their labours are but the effect of circumstances, independent of 
any sense of philoprogenitiveness, as already suggested in p. 181. The situation 
of these slaves, toiling in a strange territory for strange masters, might at first 
