II. OSMOTIC PRESSURE MEASUREMENTS 57 



A = A' 



(^ - 3 



where s = specific heat of the sokition, u = the observed undercool- 

 ing in °C., X = latent heat of fusion of the pure solvent, and a = 

 weight of solvent in a gram of solution. Since, for actual solutions, 

 the above assumptions are only approximate under the most favor- 

 able circumstances, it is necessary that undercooling be avoided for 

 most accurate results. Since undercooling is almost impossible to 

 avoid, and frequently, w^hen body or tissue fluids are involved, may 

 be of large magnitude, it is obvious that this method should not be 

 considered more than approximately exact. Unless the volume of 

 solution employed in the determination is large, further corrections 

 would have to be applied (in case of undercooling) for the specific heat 

 of the mercury in the thermometer bulb, for the specific heat of the 

 stirrer wire used, and probably for the energy added by the mechanical 

 process of stirring while the solution warms up from the temperature 

 of undercooling to the final temperature at which A' is read. 



A second method for estimating freezing point depressions of dilute 

 solutions, susceptible to considerably greater accuracy, is one intro- 

 duced by Adams {12-14)- In this method, samples of the pure 

 solvent and of a solution are placed in highly lagged vessels (Dewar 

 flasks) and brought to temperature equilibrium with the solid phase of 

 the pure solvent present in each. The temperature difference is 

 measured very accurately by use of a thermocouple (or thermopile) 

 and at this stage a sample of the solution is removed for analysis. 

 The accuracy of this method is dependent upon the exactness with 

 which the temperature difference can be estimated and the accuracy 

 of the analysis of the equilibrium sample of solution. With careful 

 work the temperature difference may be obtained within an error of 

 perhaps 0.0025 °C. For use in the determination of the freezing point 

 depression of a tissue fluid such as blood serum, an analysis for some 

 component (e.gr., chloride) of the original sample of serum would have 

 to be made. The serum would then have to be concentrated by re- 

 moval of water and a series of freezing point determinations made 

 such that the equilibrium concentrations (again followed by chloride 

 determinations on the equilibrium samples) studied would span a 

 range on either side of that of the original sample. A plot of the ob- 

 served values of A versus chloride concentration, with interpolation 

 to the chloride concentration of the original serum, should yield as 



