1920] Wheeler — Subfamilies of Formicidae and Other Taxonomic \otex 47 



from the Ponerinse and raised to the rank of an independent sub- 

 family, between the Doryhna? and Ponerinse. A number of 

 reasons may be adduced for making these changes. 



In 1899 Emery, after a comparative study of the larvae of several 

 Formicid genera, concluded that "Those of Sima and Pseudomyrma, 

 besides their extremely hypocephalic development, exhibit a very 

 special character in the presence of rudiments of antennae. I 

 believe that this very noteworthy fact, together with the well- 

 known peculiar characters of the head of the imagines, will justify 

 the separation of these genera from the remainder of the Myrmi- 

 cinse, to form the new subfamily of the Pseudomyrminse." My 

 study of numerous species of this group, which now embraces four 

 genera, Tetraponera Smith { = Sima Roger), Pachysima Emery and 

 Viticicola \Mieeler of the Old and Pseudomyrma Lund of the New 

 World, shows that Emery was far from realizing the full import of 

 their larval characters. Not only have the larvae peculiar long, 

 straight, cylindrical, distinctly segmented bodies with blunt ante- 

 rior and posterior ends, a large, usually subquadrate head, ventrally 

 placed and with rudiments of antennse (which are also present in 

 the larvae of many other ants, notably in the Ponerinae), but the 

 thoracic and first abdominal segments are furnished with peculiar 

 exudatory papillae (exudatoria) , which form a cluster around the 

 mouth. I have described and figured these organs in Viticicola 

 and Pad y sima (1918b) and have shown that they have the form 

 of extraordinary appendages in the first larval stage (trophidium) 

 of the two known species of the latter genus, and that the swollen 

 ventral portion of the first abdominal segment, just behind the 

 mouth, forms a pocket in which the workers place a pellet of food. 

 The exudatoria, the pocket, which I call the trophoihylax, and the 

 unusual method of feeding are characteristic of all four genera and 

 no distinct traces of such conditions have been found in any other 

 ant-larvae. 



More recent study has added two very interesting facts, which, 

 in advance of a complete account to be published in collaboration 

 with my colleague, Prof. I. \V. Bailey, may be briefly considered 

 in this p'ace. The food pellet proves to be merely the small pellet 

 ("corpuscule enroule," or "corpuscule de nettoyage" of Janet) 

 which the worker ant moulds in its own infrabuccal pocket 

 and consists of the solid food-particles from which the juices are 



