66o THE AMERICAN NATURALIST. [Vol. XXXVIII. 



lower bend than the outer end of the tube / is below the surface 

 of the solution in the culture jar ; for if it is not the air will pass 

 out through it instead of through the culture solution. 



The tube / serves to convey the air from the chamber G into 

 the culture solution. It may be of any size, length or form. 

 The lumsn through the constriction in it is very small ; though 

 not absolutely necessary, it serves to cause a more steady flow 

 of air. A rubber tube with an adjustable pinch-cock might serve 

 this purpose still better. 



All that is necessary to operate either apparatus described, is 

 to pour the solution from the overflow jar into the supply jar 

 from time to time. The frequency required in this operation 

 depends of course upon the size of the jars and the diameter of 

 the capillary siphon. If, as in the experiment mentioned above, 

 4-liter jars and a 0.75 mm. siphon are used, a transfer once in 

 three days will be quite sufficient. 



In closing I wish to thank Dr. E. L. Mark for furnishing 

 material which made the experiments referred to above possible, 

 and for valuable criticism of the manuscript of this article. 



