28 Dr. Mills on the Ascent of Hollow 



thermometers used could be read directly to tenths, and to 

 hundredths by estimation. The times were taken with a me- 

 tronome beating thirds of seconds. 



4. The following experiments were made at 16° with a mix- 

 ture of alcohol and water of sp. gr. "9120 at that temperature ; 

 they are compared in the table with numbers calculated from 

 the equation 



(*-T>P = * \ 



0-6-10)d 2 =23-59.J 



Table I. 



d. t obs. t calc. 



5-66 6-56 6-84 



4-76 6-83 7-18 



3-37 8-17 8-18 



2-79 8-62 9-13 



2*20 12-08 10-97 



Probable error of a single experiment, 0*44. 



The experiments were actually made on a height of 11 

 centim. ; and the values obtained for t were half those given in 

 the table. Each result is a mean of seven determinations. 



From the above comparison, it appears that the product of 

 the number of seconds t (less a constant r) into the square of 

 the diameter d of the cylinder is a constant quantity c. The 

 time of ascent through a cylinder of infinite diameter would 

 have been t=6'10 seconds. The law, however, shows indica- 

 tions of failure when the diameter is narrowed to 2*2 centim. 



5. Table II. contains the results of some experiments with 

 recently-boiled water, at 15°, and another bulb. The time is 

 calculated to seconds. Each result is a mean of these deter- 

 minations. 



Table II. 



d. t obs. t calc. 



17-16 26-2 26-17 



9-24 26-1 26-71 



5-66 26-3 27-97 



4-76 27-7 28-80 



3-37 29-9 31-63 



2-79 37-6 34-24 



2-20 41-0 39-20 



Equation, (*-25'95)d 2 = 64-56. 



Probable error of a single experiment 1*29. 



Sum of the errors (t obs. — t calc), +0*08, 



