A PENDULUM FOR THE REDUCTION TO A VACUUM. 
427 
Seventh set . — Results with Rater’s Invariable and Convertible Pendulums. 
with wooden tail pieces, 
with do. reduced, 
with brass tail pieces. 
^■without any tail pieces. 
The mean result of the invariable pendulum differs from that deduced by 
Captain Sabine, who makes n = T655. This difference arises from two causes: 
in the first place, he adopts Sir George Shuckburgh’s determination of the 
relative weights of air and water; whereas I have preferred, in all these 
reductions, the more recent determinations of MM. Arago and Biot : and 
secondly, (which is the principal cause of the difference,) he has assumed the 
specific gravity of the pendulum equal to 8*600 ; whereas I do not consider 
that it can be correctly assumed greater than 8*400, as I have already stated 
in page 414. Captain Sabine made use of two different pendulums, marked 
No. 12 and 13 ; and the results of each accord very well together. 
With respect to the convertible pendulum, it is clear that the first deter- 
mination of the values of n (viz. 2*144 and 2*204,) must be used with all those 
experiments made by Captain Rater for determining the length of the seconds 
pendulum, and inserted in the Philosophical Transactions for 1818: subject 
however to the proper correction for the vibrating specific gravity. The last 
three for the knife edge A, and the last two for the knife edge B, can be applied 
only to the pendulum as it now exists, deprived altogether of the tail pieces, 
and its sliding weight. 
The detail of the experiments with the invariable pendulum will be found in 
Invariable. 
Convertible. 
22) 
23) Knife edge A 
or 
heaviest end below. 
24) Knife edge B 
or 
heaviest end above. 
Exp. 
n 
Exp. 
n 
Exp. 
n 
I. 
ir. 
hi. 
IV. 
V. 
VI. 
1-588 
1-589 
1-570 
1-574 
1-615 
1-606 
r 
2-144 
Ph. Trans. 
1829 ^ 
- 
2-204 
Ph. Trans. J 
1829 } 
1-840 
2-205 
i 
1-853 
L 
2-012 
1831 
1-811 
1-910 
1-905 
1831 < 
r 
2-109 
2-161 
Mean = 
1-590 
Mean = 
1-875 
Mean = 
2-135 
