ON TIMEKEEPERS. 189 



semicylindrical counterpoise, so that the time might be determined either 

 from the measure or weight of the quantity of water discharged, or from the 

 position of the counterpoise. Various other modes might also be devised for 

 making cheap and simple timekeepers on similar principles, dependent on the 

 motion of various liquids or elastic fluids; but great accuracy could scarcely 

 be expected from them. A candle sometimes serves as a coarse measure of 

 time; and by burning a thread whicli passes through it, it may easily be 

 made to answer the purpose of an alarm. 



Clocks and watches are machines in which wheelwork is employed for the 

 measurement of time, being driven by a weight or by a spring, and regulat- 

 ed by a pendulum or a balance. Watches differ from clocks, in being port- 

 able, and this condition excludes the pendulum and the weight from their 

 construction. 



It is conjectured that the Saracens*: had clocks which' were - moved by 

 weights, as early as the eleventh century. Trithemius mentions an orrery, 

 moved by a weight, and keeping time, which was sent, in 1232, by the 

 Sultan of Egypt, as a present to the Emperor Frederic II. Wallingford, in 

 1326, had made a clock which was regulated by a fly. The use of such a fly 

 in equalising motion depends on the resistance of the air^ which increases ra- 

 pidly when the velocity is increased, and therefore prevents any great ine- 

 quality in the motion, as long as the moving power varies but little; and if 

 the action of the weight were transmitted with perfect regularity by the 

 wheels, and the specific gravity of the air remained unaltered by pressure or 

 by temperature, a fly clock might be a perfect machine, the weight being 

 always exactly counterbalanced by the resistance of the air, attending a certain 

 velocity of the fly ; and it might even be possible to regulate the inequalities 

 of the action of the weight, by causing the fly to open and shut, or to turn on 

 an axis, by means of a spring, according to the magnitude of the resistance. 

 The unequal density of the air would however still remain uncompensated"; 

 and in this respect a liquid would be a better medium than an elastic fluid. 

 For experiments which are but of short duration, and which require great 

 precision, a chronometer regulated by a simple fly is still a useful instru- 

 ment. Mr. Whitehurst's apparatus for measuring the time occupied in the 

 descent of heavy bodies, is governed by a fly ; the index is stopped by the 



