254 Dynamics. 



and contraction according to the state of the thermometer. But 

 a more convenient method has been devised by which the expan- 

 sion of one metal is made to counteract that of another. The 

 expansion of iron and brass being to each other as three to five, 

 Fig.isi.if we make the rod FB of iron, and the rod AO of brass in the 

 proportion of 5 to 3 ; they being connected at the lower extrem- 

 ities, and the weight being attached at O, the rod AO will expand 

 upward just as much as the rod FA expands downward, and the 

 point O where the weight is applied, will consequently remain 

 amid all changes of temperature at the same distance from jP, 

 the point of suspension. A number of rods of each kind is usu- 

 ally employed as represented in figure 182 where the rod which 

 supports the weight, is attached at Fand free at ), )', the brass 

 rods expanding upward and the iron ones downward as before ; 

 so that if the proper proportion as to length be observed, a com- 

 pensation for the effect of temperature will be obtained. Other 

 means have been invented for accomplishing the same purpose. 

 Of these we shall mention only one which has been attended 

 with great success. The weight JIB is made to consist of a glass 

 . g tube about two inches in diameter, and from 4 to 7 inches long, 

 filled with mercury. As the rod of the pendulum supporting this 

 weight, expands downward, the mercury expands upward, as in 

 the contrivance first mentioned, and the quantity may be increas- 

 ed or diminished till a compensation is effected. A dock pro- 

 vided with a pendulum of this construction, made by T. Hardy of 

 London," for the Royal Observatory at Greenwich, was found after 



* The expansions of glass and mercury being as 1 to 10 very near- 

 ly, if the suspending rod be of glass, the column of mercury must be 

 y 1 ^ of the length of the pendulum or about 4 inches. If the rod be of 

 iron, as this substance has a greater expansion in the ratio of 3 to 2 

 nearly, the column of mercury should be about 6 inches. A steel rod 

 would require a column 6,4 inches in length, which, on the suppo- 

 sition of a diameter of two inches, would weigh lOlbs. From accurate 

 calculation, it is found that if such a pendulum should keep perfect- 

 ly true time, when the thermometer is at 30, and that it should gain 

 or lose 1" a day when the thermometer is at 90, the imperfection 

 would be remedied by the subtraction or addition, as the case requir- 

 ed, of 10 ounces of mercury. 



