34 PROCEEDINGS OF THE ACADEMY OF [1881. 



surface as six and eight inclies. Tliey vary in size, but for the 

 most part, are about five or six inches in length and three or 

 four in width. They are irregular in their outlines, but have a 

 general tendency toward the oval. One of tiie most irregular is 

 figured at PI. Y, fig. '21, HR, a large chamber which lay nearly 

 underneath the gate. The gallery, g g, into which the vestibule 

 opened, debouched into tliis room, and a portion of the gallery 

 roof unbroken is shown at ug. At B, appeared a bay-room, or 

 enlargement of a gallery, which penetrated the earth horizontally 

 at one end and at the other seemed to wind into the vestibule. 

 The height of the rooms at the w^alls or sides is from one-half to 

 three-fourths of an inch. The roof is vaulted, thus causing the 

 height to increase gradually until at the centre it is one and one- 

 half inches, which is the greatest distance that I measured. 



Floors and Roof. — The floors and walls are well nigli invari- 

 ably smooth, quite smooth some of them. The roof, on the con- 

 trary, is rough, presenting the natural condition after the sandy 

 pellets of earth and the little pebbles had been picked out by the 

 workers. This can hardly be otherwise than b}' purpose, precisely 

 as with the smoothness of the floors. The roughness of the roof 

 evidently greatly favors the use to which the honey-bearers put it 

 as a perch. So the smoothness of the floor and walls much better 

 adapts them for the use of gangways. The amount of travel to 

 and fro must be enormous, it is true, in a large formicary ; but I 

 cannot think that the resulting friction will account for the smooth- 

 ness, independently of the purposed masonry of the ants. In the 

 galleries the entire surface, above and below, is smooth, a condi- 

 tion which might be anticipated on the ground of adaptation. 



Galleries and Rooms. — The galleries are tubular openings, 

 varying somewhat in size, from one-half to three-fourths of an 

 inch, and even more, in diameter. A vertical section, however, 

 uniformly shows a quite perfect circle. The underground formi- 

 cary ma}^ be described in general terms as a system of galleries 

 and rooms, arranged in several horizontal series, one above another, 

 approximating the order of " stories " in a house, and intercom- 

 municating at many points by vertical galleries. The chai'acter 

 of the interior architecture can, perhaps, be best shown further by 

 giving somewhat in detail my studies of one nest. 



The nest selected for exhaustive exploration w^as situated upon 

 the summit of Adams Ridge, just above the nook within which my 



