IXTRODUCTION. XI 



in the other subfamilies linear in the form of a slit. In the cf 

 the genitalia are often exserted, fig. xii, C h. 



Many works have been compiled on the habits of ants, one of 

 the latest and most interesting being ' Marvels of Ant Life,' by 

 W. F. Kirby, F.L.fS., F.E.S. This little book contains in a con- 

 densed form most of what was known uj) to three or four years 

 ago of the conditions of life, the economy and habits of ants in 

 general. For a popular special account, however, of Indian ants, 

 reference should be made to a paper by Mr. E. C. Wroughton in 

 vol. vii (]892) of the 'Journal" of the Bombay Natural History 

 Society. Most interesting accounts also of the habits of some 

 Indian ants are given by Mr. G. E. James Eothney in the ' Trans- 

 actions ' of the Entomological Society of London for 1889. 



The correct classification of the genera and subfamilies of ants 

 is still under discussion. The two great specialists in Myrmecology, 

 Dr. Forel and Professor Emery, hold opposite views as to the 

 relationship and arrangement of certain genera. In the present 

 work I have followed Dr. Forel in limitiug the subfamily Dorylinm 

 to two genera, Dor)/lus and ^nictiis, and including in the Ponerince 

 certain low forms of ants whose afiinities, owing to our ignorance 

 at present of their lives, development and habits, are doubtful. 

 The following table gives briefly the distinguishing characters of 

 the main divisions of the Fonnicidce : — 



Kei/ to tlu SuhfamUies of the Formicidic 



' ^ ^.' 



A. Pedicel of the abdomen one-jointed. 



a. A more or less marked constriction 



between basal two segments of 



abdomen Poiterina', p. 23. 



b. No constriction between basal t^o 



segments of abdomen. 

 a'. Anal aperture in form of a trans- 

 verse slit. 



a~. Eyes never present, blind .... Dorylhim *, p. 1. 



Ir. Eyes al\^"ays present DolicJioderince, p. 288. 



b'. Anal aperture circular Camponotina', p. 308. 



B. Pedicel of abdomen two-jointed Myrmechui', p. 105. 



In conclusion, my best thanks are due to Messrs. Smj^thies, 

 Hauxwell, Thompson, Allan and Craddock. of the Forest Depart- 



* The ^ of one of the genera, Mnictus, has the pedicel two-jointed, as in 

 the MyrmecincB, but being without eyes and ocelli, it can only be confounded 

 with the ^ 5 of the Myrniecine genera Lionn/rmex and Corebara. In Indian 

 species, however, Liomyrmrx ^ has 11-jointed, Carehara ^ 9-jointed antennae, 

 whereas yEnicfus ^ has these organs lU-jointed. The general appearance of 

 Mnidus ^ is also so distinctive, that once known it cannot be mistaken for a 

 species of any other genus. 



