Selection of Food in Stentor Ccerideus (Ehr.) II3 



can poke around the pouch a good deal before a Stentor contracts, 

 but in a satiated specimen the faintest touch generally causes con- 

 traction. But mechanical stimuli on the sides of the body do 

 not seem to cause contractions more readily in replete than in 

 hungry Stentors. 



5 If the stimuli are not too strong, contraction is often re- 

 solved into stages, and only the first of these may be passed 

 through, instead of all of them as is the case when the Stentor is 

 hungry. Thus contraction may be resolved into the following 

 separate stages, (a) Cessation of action of the membranellae. {b) 

 Closure of the pouch, (c) Gradual rounding up of the anterior 

 and posterior portions of the Stentor into an oblong mass. If 

 the stimulus still continues, complete contraction follows; but 

 this act changes the form and size of the already partially con- 

 tracted Stentor very little. No such slow contraction takes place 

 in a hungry Stentor where, if the mechanical stimulus is strong 

 enough to cause cessation of action of the membranellap, the entire 

 process of contraction occurs instantaneously. 



EXPERIMENTS WITH MIXTURES OF PARTICLES 



In this part will be considered experiments in which the food 

 particles were not fed and observed individually, but in which the 

 Stentors were surrounded by mixtures of particles of various kinds. 

 The purpose was to determine whether the animals can make a 

 selection from among the different particles of such a mixture. 

 After remaining for various periods of time in such mixtures, the 

 Stentors were placed on a slide and compressed with a cover glass. 

 It was then possible to estimate with some accuracy the relative 

 amounts of the various kinds of particles that had been ingested. 

 For such experiments we can use only particles that remain long 

 in suspension, so as to maintain their relative distribution. Heavy 

 particles cannot be employed. 



It is obvious that this method gives less precise results than 

 employed in the foregoing experiments. But by its use cer- 

 tain additional problems can be attacked. We have seen that 

 Stentor discriminates betw^een different sorts of particles fed in 



