Fkbeuary 7, 1902.] 



SCIENCE. 



217 



and, for present purposes, practically in- 

 variable unit of time. For the subdivision 

 of the day into the arbitrary units of time 

 called hours, minutes and seconds, re- 

 course is had to artificial mechanical de- 

 vices known as clocks. 



It may perhaps be stated in general, 

 without serious danger of dispute, that the 

 pendulum clock is the most accurate and 

 reliable of all types of timekeeping mechan- 

 ism. Chronometers have the advantage of 

 portability and often run remarkably well 

 for considerable periods of time, but they 

 cannot compete with the pendulum clock 

 in carrying an even rate during a series of 

 months or years. 



Yet a still higher degree of accuracy than 

 that now prevalent in the performance of 

 astronomical clocks is attainable, and is 

 necessary in the present state of astronomy. 

 There seems to be no reason why improve- 

 ments in timekeeping should not take place 

 along with the general progress in other 

 directions, where scientific results depend 

 on the perfection of mechanical appliances. 

 The sidereal clock is one of the main fea- 

 tures of an astronomical observatory, and if 

 it is to continue to be used to measure the 

 angular distance in right ascension between 

 the fixed stars, greater uniformity in its 

 rate than is now usual must be secured. 

 It is also important in time service work 

 to have clocks which will carry time with 

 greater accuracy during long intervals of 

 cloudy weather when observations of the 

 stars cannot be made. The develop- 

 ment of the pendulum clock dates from 

 the time of Huyghens, the celebrated 

 Dutch astronomer, who, in 1656, pub- 

 lished his theory of the pendulum. From 

 that time until the present the per- 

 fecting of the pendulum clock has received 

 the attention of the best mechanical artists 

 in Europe and America. Important im- 

 provements in clock-making were made 

 early in the eighteenth century, when the 



mercurial compensation and dead-beat es- 

 capement were invented by Graham, of 

 England. The gridiron pendulum, pre- 

 viously suggested by Graham, was soon 

 after constructed by an Englishman named 

 Harrison. 



Excellent practical work was done a cen- 

 tury later by a German named Kessels, of 

 Altona, who improved the dead-beat es- 

 capement by modifying the form of the 

 ' anchor.' The mechanical work of Kessels 

 is remarkably fine. He made a clock for 

 the observatory at Pulcowa in Russia, 

 and another for the celebrated astron- 

 omer, Bessel at Konigsberg. Bessel in- 

 vestigated the running of the clock with 

 his usual thoroughness and was much 

 pleased with it. He writes of Kessels as 

 ' der kenntnissreiche und vorsiehtigeKunst- 

 ler.' Kessels also made a clock for the 

 Naval Observatory in Washington, which, 

 after running for half a century, is in per- 

 fect condition and is still giving good ser- 

 vice. 



Later Tiede, of Berlin, and Hohwii, of 

 Amsterdam, attained great success in mak- 

 ing astronomical clocks, and there are now 

 two or three English and American makers 

 who are doing work of great merit. 



The Dennison gravity escapement, which 

 has recently come into use, is supposed to 

 be an improvement on the dead-beat es- 

 capement, because any small irregularity 

 in the action of the train of wheels should 

 theoretically have little or no effect on the 

 pendulum. It should, for this reason, be 

 better adapted for use in clocks provided 

 with an electric contact, worked, as is usu- 

 ally the case, by a toothed wheel on the 

 seconds arbor for transmitting signals for 

 record on the chronograph. This is an im- 

 portant practical advantage, and to more 

 certainly secure it, American clocks are 

 usually made strong and heavy and are run 

 with heavy weights. The relative merits 

 as timekeepers of the best American and 



