PETERS. — METABOLISM AND DIVISION IN PROTOZOA. 477 



employed in the course of this investigation, and always with essentially 

 the result above described. 



While it is possible to make an inference from the expeiiments with 

 salts and with water, as to the method of destruction by distilled water, 

 much more satisfactory objective evidence is furnished by the next ex- 

 periment. This consisted essentially in tiie measurement of the changes 

 in electrical conductivity exhibited by very pure water in which Stentors 

 were contained. It is true that my original object in making this experi- 

 ment was to obtain a quantitative expression for the respiratory activity 

 of these small organisms. The measurement of COo excretion seemed 

 a priori feasible. The result in the sense in which the question was put 

 was negative, but the facts developed are of material importance to a 

 general explanation of the action of water and salts as observed in my 

 experiments. To the subject of respiration I shall return in the section 

 on permeability. 



For measuring the conductivity of the medium I employed the Nernst 

 apparatus previously alluded to. Important sources of error that 

 threaten to vitiate the results of so delicate an experiment ne^d to be 

 carefully eliminated. One of these is the cliange in conductivity due to 

 change of temperature, and another is the introduction of contamination 

 into the medium from the solution of portions of the apparatus itself. 

 As is well known, increase in temperature increases the absolute con- 

 ductivity, K, of liquids, aud vice versa. A favoring circumstance in the 

 present experiment is that the temperature coefficient of pure water is 

 comparatively small, but constant values are not known (Kohlrausch u. 

 Holborn, '98, p. 115), and if any temperature change occurs during the 

 experiment its corresponding change in conductivity must be determined 

 for the case in hand. This latter I have done when necessary. 



The portion of the apparatus requiring the most scrupulous control of 

 conditions is the test-cell in which the medium with the Stentors is placed. 

 I have found by experience that good chemical glassware, such as is 

 used for beakers, frequently furnishes, especially after usage or prelimi- 

 nary treatment (Kohlrausch u. Holborn, '98, p. 113 fF), a glass vessel 

 of such resistance to solution as to give no electrically mea-urable 

 quantity of solute within the length of time occupied by my experiment. 

 The thimble-sized beaker here used, capable of containing about 5 cc, 

 was of this nature. Because of the high resistance of the medium (pure 

 water) I was able to use un platinized electrodes without experiencing 

 any difficulty from polarization in obtaining a good minimum in the 

 telephone. Aside from this, freedom from polarization is one of the 



