Febbuaey 7, 1902.] 



SCIENCE. 



219 



that the average deviation of the mean daily 

 rates fo^ the whole period is only 0^.015 

 and with a maximum deviation of 0^.03. 

 The clock was dismounted for cleaning in 

 1894 after running continuously for eight 

 years. The pressure of the air in the case 

 has been kept below the normal atmospheric 

 pressure, and mention is made of the pres- 

 sure having been made at one time as low 

 as 180 mm., about 7 inches. Little difSculty 

 seems to have been found in keeping the 

 cylinder air-tight. Indeed a slight progres- 

 sive diminution of the pressure in the 

 cylinder has been observed, and is at- 

 tributed by Professor Foerster to oxidation 

 of the metal parts of the clock and to 

 absorption by the glass walls of the cylin- 

 der of particles of moisture from the air 

 within. This clock has been for thirty-six 

 years the normal clock of the Berlin Ob- 

 servatory. 



Soon after, Tiede succeeded in mounting 

 a clock run by weights in an air-tight glass 

 cylinder, and it was exhibited at the Paris 

 Exposition of 1867. In his report of the 

 Pulcowa Observatory for 1867 Otto Struve, 

 the director, announced, with enthusiasm, 

 Tiede 's success, and stated that a clock run 

 by weights and enclosed in an air-tight 

 case had been ordered for that observatory. 

 It appears subsequently that much diffi- 

 culty was experienced from various causes 

 in getting the clock into working order. 

 But it was finally set up, about the year 

 1880, in the basement of the Pulcowa Ob- 

 servatory, where the temperature changes 

 only four or five degrees a year, and was 

 found to run with a satisfactory rate. This 

 was formany years, and presumably is still, 

 used as the principal clock of that observa- 

 tory, which is an institution widely known 

 for the high quality of its work. The 

 pendulums of these clocks at Berlin and 

 Pulcowa were compensated, of course, for 

 change of temperature. 



The Riefler clocks mentioned above are 



constructed so as to be easily mounted in 

 air-tight cylinders, which together with the 

 clock itself rest on a shelf bolted to the 

 clock pier. There is one of these clocks 

 mounted in the usual way at the George- 

 town University Observatory at Washing- 

 ton. It is run by a weight which is wound 

 up every few minutes by electricity. But 

 it is not found practicable, under the con- 

 ditions there, to keep the temperature 

 strictly constant. 



The standard clock of the Greenwich 

 Observatory by Dent, of London, is 

 mounted in the basement of the observa- 

 tory, where the temperature changes are 

 small and very gradual, and is fitted with 

 an electrical device for barometric com- 

 pensation. 



The standard clock of the Paris Observa- 

 tory, by Winnerl, enjoys the unique distinc- 

 tion of being mounted in a vault at a depth 

 of 27 meters underground. The tempera- 

 ture changes at that depth are of course 

 very small, being, according to Tisserand, 

 not more than one or two hundredths of a 

 degree during the year, but the effect of 

 barometric changes on the rate of the clock 

 has been found to be serious. 



There seems to be no case where an at- 

 tempt has been made to keep both tempera- 

 ture and barometric pressure strictly con- 

 stant. There is, I think, no doubt that it 

 is entirely feasible to maintain a suitably 

 constructed vault at a practically constant 

 temperature throughout the year by arti- 

 ficial means. Then, with an air-tight case, 

 the barometric pressure could be kept 

 practically uniform and the clock would be 

 completely protected from dust and moist- 

 ure. Even if it were not practicable to 

 get the ease perfectly air-tight, a prac- 

 tically uniform pressure could be main- 

 tained by exhausting the air from time to 

 time, provided that the leakage is very 

 small. 



Accurate comparisons of clocks running 



