246 A. A. SCHAEFFER 



is much more important than the succeeding stimulus. Thus 

 in experiments 6-13 (Schaeffer, 1916c) the sensing of tyrosin 

 grains affected the ameba in such a way that the succeeding 

 grain of globulin was treated at first like a grain of tyrosin. 

 But after the ingestion of the globulin, the succeeding grain of 

 tyrosin was treated like the grain of globulin, and promptly 

 eaten. The behavior of the ameba was ordered with definite 

 reference to its past. It is clear that each of these items of 

 behavior, taken by itself, could not be explained. The process 

 of selection cannot be understood by treating in a quantitative 

 manner a number of separate acts of choice. On the contrary, 

 selection must be looked upon as a connected, developing process 

 involving at least the entire past of the individual ameba. 



SELECTION OF FOOD IN STENTOR AND PARAMECIUM AS 

 COMPARED WITH THAT IN AMEBA 



In this section are summarized the chief facts bearing especially 

 on the selection of food in stentor, Paramecium and ameba. A 

 comparison of the main results of investigations on these 

 organisms serves # to give a better conception of this problem of 

 choice of food, I believe, than can be had in any other .way. 

 Incidentally, such a comparative study gives also a hint as to 

 the probable path of development in a species of whatever 

 method of selection may be present in that species. The reason 

 for selecting these three organisms is that they serve as types 

 of three methods of selection, and what is equally important, 

 the food selective processes are better known in these three 

 organisms than in any other lower forms. 



Paramecium. Paramecium feeds on bacteria and other very 

 small particles, but bacteria form much the larger part of the 

 normal diet. 



A very great number of particles, hundreds of thousands or 

 millions, must be eaten every day in order to maintain life. 



If the particles were tested individually, an enormous amount 

 of time would be consumed in the process, and there would not 

 be sufficient time to eat the optimal amount of food. The 

 ingestion of particles of sand, debris, mud, silt, etc., if each 

 particle were tested individually, could therefore not be con- 

 sidered an advantageous reaction, since it would entail a 

 dangerous decrease in the amount of food eaten. In times of 



