22 proceedings: philosophical society 



and by keeping the furnace temperature reasonably uniform by means 

 of platinum-faced partitions and a suitable arrangement of the furnace 

 winding. Duplicate results seldom differed as much as 0.001. Sys- 

 tematic errors in calorimetry are usually much larger than the acci- 

 dental, but numerous and varied intercomparisons seem to indicate that 

 in this case there were few errors of any sort greater than 0.0015. The 

 specific heats at high temperatures show in general a curvature con- 

 cave to the . z-axis, which is merely the upper part of an »S-shaped 

 curve characteristic of these as of all substances and explainable on the 

 basis of the quantum hypothesis. Accepting that explanation, these 

 curves show that the silica and silicon compounds investigated have 

 atomic vibration periods of high frequencies comparable with those 

 characteristic of the atoms of the diamond, whose specific heat curve 

 is similar. These vibration frequencies are due to the oxygen in the 

 compounds and are characteristic of oxygen compounds generally. 

 Platinum shows an altogether different curvature at ordinary tem- 

 peratures. The relation between the specific heats of the various 

 substances when crystalline and when in the form of glass showed 

 marked differences. There was also a tendency for the specific heat 

 of the glass to increase 10 per cent or more at some rather high tem- 

 perature; this as yet remains unexplained. Two definite kinds of in- 

 version or transformation in the solid state were demonstrated, char- 

 acterized by the presence in one case and the absence in the other of 

 large variations in the specific heat below and at the temperature of 

 inversion. 



Discussion. Mr. Swann noted that the agreement of the specific 

 heats fitted in well with the quantum theory. Mr. Sosman referred 

 to the different slopes of the heat curves for polymorphic substances 

 and cited in particular the three forms of quartz; the question whether 

 the cause is of atomic or molecular nature is an open one. Mr. L. J. 

 Briggs spoke of the atomic heat of platinum at the higher tempera- 

 tures. 



Mr. N. S. Osborne then presented an illustrated communication on 

 A calorimeter for the determination of latent and specific heats of fluids. 

 The principle of the unstirred or "aneroid" type of calorimeter has 

 been embodied in an instrument especially designed for determinations 

 of the specific heat and latent heat of vaporization of several sub- 

 stances in general use as refrigerating media. Heat developed elec- 

 trically in a coil located in the central axis of the cylindrical shell com- 

 prising the calorimeter is distributed by conduction to the calorimeter 

 and contents whose initial and final temperatures are measured, when 

 in thermal equilibrium, by a platinum resistance thermometer. Heat 

 from other sources is excluded by enveloping the calorimeter with 

 a metal jacket, separated by an air space, and keeping this jacket dur- 

 ing measurements at the same temperature as the calorimeter surface, 

 using multiple thermocouples to indicate this equality. The measured 

 heat added is used either to change the temperature of the contents 

 or to evaporate a portion of the contents withdrawn as superheated 



