The Termites, or White Ants 109 



the diflerentiation of individuals. To understand this claim it is necessary 

 to attend more closely to the feeding habits. The food of termites con- 

 sists almost exclusively, as has already been said, of wood. But this wood 

 may be taken directly from the walls of the burrow or secured indirectly 

 from another individual. In this latter case it consists of disjecta of undi- 

 gested material, which, while mostly wood, must be mixed with other or- 

 ganic material: because the termites keep their nests clean by eating their 

 cast skins and the dead bodies of other individuals. This undigested mate- 

 rial is called proctodeal food. In addition, a certain amount of evidently 

 very different matter is regurgitated through the mouth from the anterior 

 part of the alimentary canal. This is called stomodaeal food. As the young 

 receive all their food from the workers, it is apparent that there is oppor- 

 tunity for a choice, on the part of the nurses, in the kind of food given the 

 young. And it is presumed by Grassi that such a choice is made, and that 

 it results in the differentiation of the castes. As a matter of fact, such a 

 differentiation of individuals is accomplished in the honey-bee community 

 by feeding those larva; which the workers wish to make fertile queens "royal 

 jelly" — a rich food regurgitated through the mouth from the anterior part 

 of the alimentary canal. This is done for the queens during the whole 

 larval life, while larva; which are fed royal jelly for only one or two days> 

 and then mixed pollen and honey for the rest of larval life, develop into 

 workers. With the honey-bee, however, the workers are to be looked on 

 as probably only arrested females. But in the case of Termopsis angusli- 

 collis Heath has experimented by feeding members of 

 various colonies, both with and without primary rojal 

 pairs, "on various kinds and amounts of food — procto- 

 deal food dissected from workers, or in other cases from 

 royal forms, stomodeal food from the same sources, 

 sawdust to which different nutritious ingredients had 

 been added — but in spite of all I cannot," he says, 

 "feel perfectly sure that I have influenced in any un- 

 usual way the growth of a single individual." 



This is by all odds the most important and interesting 

 problem in termite economy, and the solver of it will do 

 much for zoological science. 



A singular and primitive family of small insects, the _ p h ■ 



Embiidas, of doubtful affinities, is represented by not tcxana. (.Vfter 

 more than twenty hving species, of which but four M dander; en- 

 occur in this country. The individuals do not live in 

 communities as the termites do, but in structural characters they probably 

 more nearly resemble these insects than any others. Fig. 139 illus- 

 trates a typical Embiid. This species, Embia texana, is about one-quarter 



