BIOLOGIC RELATIONS BETWEEN PLANTS AND ANTS. 427 



genistas). In the United States the caterpillar of a species of Lyccena 

 has upon its last abdominal segments two or three pairs of small pro- 

 jecting buttons provided Avith a central opening from which exudes, 

 under the influence of the caresses of Formica fusca, a small drop of a 

 special liquid. 



As we have seen it is incontestable that the ants thus protect a num- 

 ber of insects injurious to vegetation against the attacks of their 

 natural enemies, but in certain cases it seems probable that ants, by 

 transporting these sucking insects from developing to older parts of 

 the plant may considerably aid the vegetable to sustain the attacks of 

 these parasites. For example, an aphis, by living on the young leaves 

 of a bud, will frequently cause them to develop abnormally, while if it 

 lived on the adult stem it would be much less prejudicial to the plant. 

 Oases will hereafter be cited in which ants transport aphides and 

 cochineals from one organ of tlie rdant to another. 



All these nectar-producing insects may be considered, generally, as 

 walking nectaries. They attract ants much more powerfully than do 

 the extrafloral nectaries. Traversing almost the entire surface of the 

 plant, they determine the goings and comings of the ants, which thus 

 indirectly protect the entire plant by their very presence instead of 

 remaining massed at special points where nectaries are found. But it 

 seems to us going too far to consider with Lundstrom that these walk- 

 ing animal nectaries are profitable to the plant. The quantity of nutri- 

 tive materials they take from it, and the deformations they cause in a 

 number of organs, are not compensated for by the protection, doubtful 

 indeed in many cases, offered by the ants they attract. 



The true plant-protecting ants are those which do not borrow (even 

 indirectly by means of aphides) their aliment from the vegetable king- 

 dom — those which are frankly carnivorous. Such ants are quite numer- 

 ous in our climate, and their usefulness to agriculture and silviculture 

 is incontestable. We find that Formica pratensis is very destructive to 

 insects such as caterpillars and grasshoppers. A nest of this species 

 will destroy as many as 28 insects per minute, or about 1,G00 per hour. 

 And such a colony works day and night throughout the entire season. 



In the midst of the arid savannas of America the beneficent action of 

 ants is shown by islets of verdure covering the hillocks raised by these 

 insects. The protection which they give to plants prevents the attacks 

 of leaf eaters. 



We have not touched upon many well-known points in the above- 

 mentioned biologic relations of ants. We prefer to concentrate our 

 attention upon their direct relations with a number of plants that may 

 be called myrmecophilous, since they afford shelter and often food to 

 ants. The history of these lodging plants is generally little known, and 

 they present a number of peculiarities which deserve to be studied in 

 detail. The instinct of ants leads them to attempt to establish them- 

 selves in cavities where they may be sheltered. These cavities will be 



