440 PROCEEDINGS OF THE AMERICAN ACADEMY. 



frontal carlnae make with each other, the shape of the clypeus, and the 

 pubescence, this being even more marked than in the soldiers. 



4. On the other hand, the mandibles are toothed and in shape are 

 much like those of the small workers, and the mandibular muscles do 

 not cause mouncls on the dorsal surface of the head. 



5. There are three well marked ocelli, and the compound eyes are 

 much larger than those of either soldiers or workers. 



6. From these facts it seems probable that the large workers orig- 

 inally resembled the queen more than did the small workers, and later 

 developed to an exaggerated extent certain of her characteristics, which 

 the small worker, on the other hand, lost. 



IV. Methods in General. 



In order that I might not be biased when considering the subject of 

 division of labor among ants as a whole by conclusions deduced from a 

 single method of work, it was thought well to make observations under 

 several different conditions, partly by placing the ants in more, partly 

 in less natural surroundings. My observations may thus be arranged 

 in three gi'oups as follows : A. work with Fielde (:00; :04^) nests, with 

 aluminum nests (Buckingham, :09) and other apparatus ; B. Work with 

 Barth (:09) nests; C. Out-door work. 



A. Work with Fielde Nests, Aluminum Nests and Other Apparatus. 



The Fielde and aluminum nests used are shown in plan in Figure 6 

 (p. 447) ; they were ten inches long and six inches wide. 



1. Marking. — In order to study the activities of each member of a 

 colony, when experimenting with ants in the Fielde nests and in my 

 own, the following method was employed : Each individual was marked 

 and its head measurements were recorded together with the data of the 

 various experiments, so that it might be possible to tell which sizes 

 were concerned in the different activities and to keep account of the 

 special activities of each individual. Length of head multiplied by width 

 of head was used as a criterion of size. The classes of the species of 

 Pheidole which I used are so distinct that I have thought no marking 

 necessary, and have simply recorded the numbers of ants of each class 

 taking part in each activity. With Camponotus, however, the case 

 was quite different, since it was impossible to establish natural groups 

 or classes, owing to the continuous gradation of the forms into one 

 another. 



A common way of marking ants, described by Miss Fielde (:03^, 

 p. GIO, foot-note), is to af&x to them colored paint by means of var- 



