Measurement of changes in the dimensions of Solid Bodies. 458 
The probable error is 0°461 of a scale division, which corres- 
ponds to xz¢d,000 Of an English inch. 
No. 3. 
Actual readings. Diff. M’. MM”. M. 
ve re 29° 30°25 29°62 
set a 29:25 30°1 29°67 
ee: ry : 32°2 32°1 32°15 
pi alt 29°55 29°70 29°62 
ie: es 29°15 26°9 28-02 
oe stat 28°8 30°9 29°85 
iy 27°5 Final average, 29°77 
The probable error is 0°31 of a a Neteg ii corresponding to 
11,30,000 Of an English inch. It is to be remarked that while 
these last readings were being sakes thie heavy trains passed. 
No. 4. 
Av. of 25 read. _Diff. M’. M”, M. 
75°04 29°54 : ‘ ‘ 
45°50 29°37 29°45 29°27 29°36 
74°87 2971 oes : : 
72°40 32°80 Final average, 29°10 
39°60 
The probable error is 0°16 of a scale division, corresponding to 
21,985,000 Of an English inch. 
Combination of results, 1, 2, 3 and 4. 
It is evident that the weights of these results are quite dif- 
ferent; accordingly after taking all things into consideration, I 
assigned to Nos. 2 and 8 single weights respectively, to No. 1 
a double weight, and to No. 4 a three- fold weight. Under 
these conditions the average of He four sets of observations is 
28-99, with a sits ean of re - a scale division, corre- 
sponding to se,a0 7,008 n English inch 
In Nos. 1, 2, and in ibe first fear readings of No. 4, there is 
no decided evidence that the pendulum had any independent 
motion of its own; in No. 3, its independent srt seems to 
have been nearly uniform. 
