144 SOME EEFINEMENTS OF MECHANICAL SCIENCE. 



little variation in the height of water in the second jar, and in the 

 third the height remains practically uniform, thus insuring a con- 

 stant head for the water which drops into the registering jar. At 

 the beginning of each day the water is taken from below and carried 

 up a flight of steps to the top. 



That such an arrangement has some elements favorable to the ac- 

 curate measurement of time there can be no doubt. It certainly has 

 the element of simplicity, and notwithstanding its long -service, the 

 only wear noticeable was confined to the steps leacling to the up- 

 per jar. 



Clocks of the present type, although used as far back as the twelfth 

 century, and possibly earlier, were but fair timekeepers until several 

 centuries later. Those Avhich the astronomers used in their observa- 

 tories at the end of the fifteenth century were so unreliable that mod- 

 ified forms of the clepsydras of the ancients were used, and as they 

 did not prove to be satisfactory, most of the observations were made 

 without the use of clocks. 



Galileo's beautiful discovery of the iso^-hronism of the pendulum 

 from the swinging chandelier in the cathedral at Pisa was of great 

 value in many respects, but in none more so than in its application 

 to the measurement of time. 



Soon after that great discovery the English clock maker, Graham, 

 invented the mercurial peudulum. by Avhich the variation in its 

 length caused by the difference in temperature was fully compensated, 

 and some years later Harrison, another English clock maker, invented 

 a compensating ])endulum, which consisted of a series of metal bars 

 having different coefficients of expansjon — so that two liundred years 

 ago, as it is to-day, the pendulum was the nearest perfect of all the 

 devices that have been employed for governing or controlling the 

 motions of a clock mechanism. 



Every part of the clock down to the minutest detail has been the 

 subject of study and improvement, and clocks are uiade and adjusted 

 with such precision and delicacy that in testing them the question is 

 within how small a fraction of a second will they run. Not content 

 with their marvelous performance when under normal conditions, 

 some of the finest astronomical clocks are surrounded by glass or 

 metal cases, in which a partial vacuum is maintained, and in order 

 that the cases may not be opened or disturbed the winding is done 

 automatically by means of electricity, the frequency of the winding 

 in some cases being as often as once every minute. These clocks are 

 set up in* especially constructed rooms or underground vaults, where 

 they are free from jar or vibration, where the temperature and 

 barometric conditions remain practically constant, and where every 

 possible precaution is taken to further minimize the errors of the 

 running rate. 



