468 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1912. 



cystus Tiortideorum) . Other ants, again, are harvesters, who col- 

 lect supplies of seeds and store them away in the granaries of 

 their nests. For over 100 years the account in the Holy Bible 

 of the harvesting ants of Palestine was supposed to be a fable, 

 until, tlirough the observations on the genus Messor in the Mediter- 

 ranean region, it was gloriously verified. Recently Neger has even 

 found that Messor harbarus works the seeds over into a kind of ant 

 bread. The classical "agricultural ants" of North America belong 

 to the genus Pogonomyrmex, and, although the exaggerated roman- 

 ticism of the "agriculture" of Pog. harhatus has been destroyed 

 by Wheeler's critical investigations, yet the habits of these har- 

 vesters are still very interesting. Even greater interest has been 

 excited by the fungus growing ants, particularly since the researches 

 of Moller (1893). The habits of the American leaf-cutting ants, 

 from the group of Attini, has been thereby placed in a new light, 

 especially since, through the observations of v. Ihering, E. Goeldi, 

 and particularly Jacob Huber (1905), we are also informed regarding 

 the ingenious manner in which the queen of Atta sexdens starts and 

 cultivates the new fungus garden, when, after her marriage flight, 

 she founds her new colony. That in this apparently highly intelli- 

 gent "cultivation of vegetables" by the ants we are dealing with an 

 hereditary instinct, is verified moreover tlirough the analogy with the 

 fungus growing of the termites, among whom this custom, particularly 

 in the large genus Termes, is still more widely prevalent, although 

 the termites stand psychically below the ants. A dainty fungus 

 garden, of the size of a walnut, from the nest of a small guest termite 

 (Microtermes globicola), which lives in the hills of Termes Redemanni 

 in Ceylon, is shown in the accompanjdng enlarged illustration (fig. 

 22). Many other ants, finally, live by the hunt, particularly for 

 insects. With our large heap-building hill ants {Formica rufa and 

 F. pratensis) this means of gaining a liveliliood is truly enough only 

 secondary in comparison with the visits to plant lice and scale in- 

 sects. In any case it is important enough to place these and other 

 acervicolous species of Formica of the rufa and exsecta groups as 

 eminently "useful" under the protection of forestry laws. The san- 

 guine robber ant (Formica sanguinea) even occupies herself almost 

 exclusively with hunting, and leaves the cultivation of plant lice to 

 her slaves. 



But the carnivorous hunting ants are far more numerous among 

 the tropical and subtropical species, among the Ponerinse (Lobo- 

 pelta, etc.), and most particularly among the Dorylinse. 



5.— the; dorylin^ and their guests. 



The subfamily of Dorylinse comprises the hunting ants, which, 

 partly above ground and partly subterraneanly, go about in quest 



