chi;mistry — morse. 197 



closing period precedes the measurement of pressure, while that which 

 occurs at the end of the experiment is subsequent to the measurement and is 

 to be neglected. No sure means of discriminating between the two had been 

 discovered and the resulting uncertainties were sufficient to mask the effect 

 of temperature upon pressure. It was highly desirable, therefore, to sup- 

 press all dilution, or at least to reduce it to an insignificant minimum. This 

 has been accomplished by making radical changes in the construction of the 

 porous cells and their attachments. The work along this line has been de- 

 scribed in the September number of the American Chemical Journal. 



The third large source of error in the measurement of osmotic pressure 

 was due to fluctuations in the temperature of the bath. Such changes during 

 an experiment produce the so-called thermometer effects which have been 

 much discussed in the various papers already published. It is sufficient for 

 the present purpose to state that a rising temperature in the bath produces 

 an overpressure in the cell, while a falling temperature has the opposite 

 efifect, and that these changes, unless they are very minute and exceedingly 

 slow, may lead to errors large enough to make the determination of a tem- 

 perature coefficient impossible. The older arrangements for temperature 

 regulation were found to be quite inadequate and entirely new ones were 

 devised, which enable us to secure and automatically maintain for any length 

 of time any temperature between that of the hydrant water and 50°. The 

 unavoidable fluctuations do not exceed 0.1° in the largest baths. A paper in 

 which the new arrangements are described is in process of preparation. 



Briefly stated, the system consists in rapidly passing the water in a bath 

 first over a cooled surface and then over a heated one. The former is se- 

 cured by the regulated flow of hydrant water through coils of pipe, or a closed 

 chamber, which is submerged in the bath. If necessary, as when a low 

 temperature is to be maintained in hot weather, the hydrant water is previ- 

 ously cooled by passing it through pipes lying in ice water. The heating of 

 the second surface is electrical, and is under the control of a thermostat im- 

 mersed in the water of the bath. The flow of hydrant water is so regulated 

 that it would normally maintain in the bath a temperature somewhat below 

 that required, while the heating arrangement is capable of carrying it con- 

 siderably higher. It will be seen that, under these conditions, any tempera- 

 ture for which the thermostat is set will be automatically maintained, the 

 exactness of the regulation depending solely on the sensitiveness of the 

 thermostat. When temperatures above that of the air are to be maintained, 

 the flow of the hydrant water is cut off. The exterior of the bath then be- 

 comes the cooled surface and the regulation of temperature proceeds without 

 change in principle. The system has been found to be generally applicable to 

 bath control in the laboratory. 



