494 merwin: thermal dehydration of stilbite 



humidity of the surrounding air. For some materials the sur- 

 face resistance at 1 per cent humidity is lO^i times larger than 

 at 95 per cent humidity, though for the majority of materials 

 the surface resistance does not change by a factor of more than 

 10^ under these conditions. Since the change is not uniform it 

 has been found necessary to construct curves showing the change 

 of surface resistivity with humidity. 



PHYSICAL CHEMISTRY.— r/ie thermal dehydration of stilhite, 

 thaumasite, and the hydrates of magnesium sulphate and of 

 copper sulphate. H. E. Merwin. Geophysical Laboratory. 



The method of removing from substances the water which 

 comes off readily only at high temperatures, as outlined in the 

 study of metahewettite,^ is here applied to a part of the water 

 which is lost readily at ordinary temperatures. About 0.5 

 gram of material, powdered to pass 40-to-60 mesh, is held in a 

 thin layer between an outer 5-inch test tube and an inner, 

 inverted, 3-inch test tube, the bottom of the latter having been 

 drawn out and having a small opening. The combination is 

 immersed nearly to the top in an oil or nitrate bath, heated to 

 the temperature and for the time desired, and weighed. The 

 immediate object is to determine the rate of loss of water as the 

 temperature is gradually increased. Breaks in the rate should 

 represent definite hydrates if such exist, provided such hydrates 

 have sufficiently differing vapor pressures at the temperatures 

 considered. Some characteristic breaks are shown in figure 1. 

 Temperature is plotted in one direction and in the other either 

 the number of molecules of water to one molecule of anhydrous 

 salt or the percentage loss in weight. The circles or the shapes 

 of the closed areas represent the rate of dehydration. The 

 diameter of a circle is proportional to the amount of water lost 

 during the last heating divided by the time of heating. In 

 cases where the circles overlapped considerably they were re- 

 placed by two continuous lines through the ends of each diameter 

 normal to the curve. The periods of heating were 5 minutes. 



' Hillebrand, Merwin and Wright, Am. Phil. Soc. Proc, 53: 45. 1914. 



