6 Prof. R. Bunsen on a Method of defer mining the 



body, and somewhat thicker than the glass tube which it is to 

 line. After the portion of tubing which projects beyond the 

 knot on the thread has been cut away with scissors, the glass 

 tubes are strung on to the thread. If the thread is now pulled 

 tight while the extremity of the caoutchouc tube is held fast, 

 the latter is lengthened out and becomes so thin that the glass 

 tubes can be conveniently pushed over it at measured distances 

 from each other. If the extended tube is now allowed to regain 

 its original position, it is pressed with such force against the in- 

 terior of the glass tubes that perfect contact is formed where 

 the glass and caoutchouc touch. Such caoutchouc tubes, sur- 

 rounded on all sides by glass, scarcely change their weight at 

 all in the air ; and they form, when the one end is closed with 

 its glass stopper, and the other with the fine point of the glass 

 vessel, a perfectly air-tight connector capable of being readily 

 opened at any time. In order that these connectors of caout- 

 chouc may be weighed together, their weight must be equalized. 

 Each one is distinguished by the number of the vessel to which 

 it belongs, and the adjustment for weight is made in the follow- 

 ing manner : — 



Suppose we have five such connectors, I k , II*., Ill/,, IVj., V&, 

 prepared of nearly the same size, the stoppers to each, made out 

 of heavy glass rods (fig. 2,7), must not at first be made to end in 

 a ball, as represented in the figure, but in a long drawn-out glass 

 thread. By breaking oft and again melting the glass thread, it 

 is easy to obtain the connectors with their stoppers of nearly 

 equal weight. Connector l k is now placed on the right pan as 

 tare, and connector V* is laid on the left ; the position of equi- 

 librium of the balance is observed by the oscillations, and the 

 differences in weight which II fc , III*., IV* show with Y k are ac- 

 curately estimated in the manner described above for the vessels 

 themselves. If T 5 be the unknown weight of Vj., A 2 , A 3 , . . . , 

 the differences of weight between Y k and the separate con- 

 nectors, then the weight g of the same is 



II, . . ^ = T 5 + A 2 , 



III, . . ^ 3 =T 5 + A 3 , 



IV, . . *=T,+A 4 . 



The weighings obtained in this manner gave 



-0-00059 -0-00060 -0-00041 



-0-00062 -0-00046 -0-00043 



-0-00070 -0-00055 -0-00053 



—0-00040 -0-00043 -Q-Q0033 



Mean - 0-00056 =^A 2 -0-00051 = A 3 ^0-00043 = A 4 



