446 PROCEEDINGS OP THE AMERICAN ACADEMY. 



determined that from two to four transfers were sufficient to prevent any 

 perceptible contamination. The number necessary depends much upon 

 the experimenter's personal method of working. The construction of the 

 pipettes is also an important factor. It appears from the following con- 

 sideration that the amount of contamination resulting upon the dilution 

 following the use of the pipette must be exceedingly small. Capillary 

 pipettes can be made of such small diameters as to admit but one Stentor 

 at a time, and under the magnifier five to ten animals can be drawn in 

 with a very small quantity of water. This small amount of liquid is 

 furthermore a very small proportion of the 4 to 5 cc. with which it is 

 diluted in the next watch-glass. It follows that successive transfers 

 rapidly multiply the dilution of the original contamination, which at 

 length becomes infinitesimal. This source of matter foreign to the de- 

 sired medium can be reduced to as low a value as that resulting from 

 the practically unavoidable contact with air and glass. The subject of 

 the solubility of glass will be treated of in the experiments where that 

 factor becomes of importance in drawing conclusions. 



The selection of the Stentors for experiment is also a matter of much 

 importance. No other factor that might be overlooked would so quickly 

 bring discrepant results. If the metabolic efi^ects of different reagents 

 are to be compared, it is important that the experiments be conducted, 

 if possible, upon animals in the same condition. When that is not prac- 

 ticable, it is important to have a means of comparison by which diff"erent 

 results can be reduced to a common standard. The following experi- 

 ments offer illustration of both procedures. The experimenter who 

 maintains a continuous supply of material (Stentor) will have on hand 

 cultures in various stages of development. For comparative experiments 

 the animals should be taken invariably from the same culture. A good, 

 well-established culture has considerable longevity and makes the above 

 rule practicable. The results are still more favorable when the experi- 

 menter has learned to raise healthy cultures by a uniform method. In 

 any case Stentors presenting abnormal or special conditions, such as the 

 undersized, the pale, those that are conjugating or dividing, should be 

 rejected. Material should always be abundant, for if it is, practice soon 

 secures a very uniform selection of normal animals. Among the exam- 

 ples which follow there are some which show the possible reduction of 

 results from different sources to a common metabolic standard. 



The question as to the number of organisms that are necessary to a 

 given experiment is one that cannot be answered by a general statement. 

 In experiments upon metabolism where division is the criterion, cumula- 



