192 



THE EARL OF BERKELEY ON SOME 



method of running a thread of mercury along the bore, measuring its length, and 

 then weighing it. 



The capacities were found by weighing the pyknometers filled with water at C. 

 and at 90 C. respectively ; the volume occupied by the water was taken to be that 

 o-iven in LANDOLT and BORNSTEIN'S tables for water which is air-free. The difference 



to 



between the capacities thus determined gave the expansion from C. to 90 C., and 

 for intermediate temperatures it was assumed to be proportional to the temperature 

 interval ; this assumption was tested with one of the pyknometers, and it was found 

 that the resulting difference was within the experimental errors. With the 11 cub. 

 centim. pyknometers, however, it was deemed advisable to examine the error more 

 closely, and for this purpose the capacities were determined at five approximately 

 equal intervals of temperature between C. and 90 C. The numbers obtained were 

 plotted against the corresponding temperatures, and a bent-ruler curve passed 

 through the points ; the capacities for intermediate temperatures were taken from it. 

 The maximum difference between this curve and a line joining the penultimate 

 observations represented a difference of '0015 cub. centim. This is a quantity which 

 is barely greater than the experimental errors, as will be seen from the following 

 numbers obtained with one of the pyknometers : 



On re-determining the capacities after an interval of several months no change was 

 apparent. 



As the table used for the expansion of water gives numbers derived from air-free 

 water, and as the pyknometers had been filled with water which had not been freed 

 from dissolved air, it was thought possible that an error had been introduced in this 

 way ; a pyknometer was therefore filled, in a vacuum, with water which had been 

 boiled in that vacuum for three-quarters of an hour ; it was then withdrawn and 

 brought to a constant temperature in the thermostat and weighed in the usual 

 manner. The results of three observations carried out thus did not differ from those 

 obtained with ordinary water by more than the latter differed among themselves. 

 Taking into consideration that the solutions themselves are not air-free, it was 

 considered unnecessary to pursue the matter any further. 



* This observation was one made with air-free water. 



