26 MORRIS LOEB 



toplasm of such cells, placed in a solution whose solvent only 

 can penetrate the membrane, will yield water to the solution 

 if the latter be concentrated, while it will take it up again if 

 the solution be diluted. The protoplasm will therefore re- 

 cede from the walls of its cell, or again approach it, and this 

 expansion or contraction can be observed with the micro- 

 scope. By systematic dilution, a point may be determined 

 for every substance where it is isotonic with another, i.e., will 

 neither expand nor contract a protoplasm which had come 

 to rest in the other. Interesting as this method is, and ca- 

 pable of yielding good results in the hands of a skillful micro- 

 scopist, it is hardly useful in the chemical laboratory, where, 

 aside from lack of familiarity with microscopic work, the 

 investigator would be hampered by the exclusion of all 

 substances which will kill plant life. 



The freezing-point method has recently been reviewed in 

 this magazine; it therefore only remains for the writer to 

 express his view of its scope. While originally only those few 

 liquids were used as solvents whose freezing-points approached 

 that of water, recent investigators have successfully em- 

 ployed substances like paraffine and the more fusible metals 

 as the solvent; there are, therefore, few substances which are 

 not amenable to the method. But the following errors should 

 be avoided: the use of thermometers not sufficiently sensi- 

 tive to admit of observations at high dilutions; contenting 

 one's self with observations within too limited a range of con- 

 centration to exclude a chance of overlooking abnormal be- 

 havior at some point; allowing too great a change of concen- 

 tration to occur through the separation of the solvent in the 

 solid state; using solvents which have a chemical effect upon 

 the substance under investigation. Substances dissociate in 

 one liquid which remain in complexer molecules in another. 

 Chemists, in applying Raoult's methods, must remember 



