334 WHEELER— AXT LARV^. 



is trophallactic, it is clearly essentially the same as the relation 

 between host and symphile. It becomes unnecessary, therefore, to 

 assume that in the ants and termites the primitive nursing- instinct, 

 which is a mutual feeding, has been specialized or modified during 

 the phylogeny in adaptation to particular symphiles. Slight onto- 

 genetic modifications, well within the limits of the plastic, or " intel- 

 ligent " behavior of the ants, as responses to the specific organiza- 

 tion of the symphiles, seem amply sufficient to account for the 

 phenomena. 



In the second place, trophallaxis is, of course, traceable to a 

 mutualistic hunger, or " exudate hunger " as Holmgren calls it, and 

 therefore to an appetite, in the sense in which this term is employed 

 by English psychologists. In view of the fact that psychologists 

 have universally regarded the appetites as very primitive and fun- 

 damental it is rather strange that they have received so little atten- 

 tion from the animal behaviorist. Very recently, however, Drever 

 (1917) and Craig (1918) have emphasized their importance in con- 

 nection with instinct in two valuable contributions. Drever regards 

 the appetites as very simple or primitive instincts or " as repre- 

 'senting an earlier stage of conscious life, which in the human being 

 .and the higher animals, is overlaid by the stage to which the develop- 

 ment of the specific ' instinct ' tendencies belong." He enumerates 

 the hunger, thirst and sex appetites, the appetite for sleep or rest, 

 for exercise or activity, " nausea," or " primitive disgust " and 

 James' " instinct of personal isolation." Craig's contribution is par- 

 ticularly interesting because he reaches his conclusions from a study 

 of birds (doves) and deals with the matter more thoroughly. Ac- 

 cording to him the appetites and aversions are constituents of the 

 instincts. " Each instinct involves an element of appetite, or aver- 

 sion, or both." Perhaps his view is not essentially different from 

 Drever's, since the most typical appetites, those of hunger and sex, 

 represent the basic reactions of organisms, and what are ordinarily 

 called " instincts," /. e., the chain-reflexes, or more elaborate mechan- 

 ized behavior of animals, are evidently later and superposed activi- 

 ties that, so to speak, adopt the general movement or pattern of ex- 

 pression characteristic of the appetites. Craig, in fact, resolves the 

 behavior of animals into cycles which run their course according to 



