Mr Buchanan ’ on the Strength of Materials. 1 61 
ported between these, its extremities resting on the top of them ; 
and being attached in the middle to the top of the bellows, by 
means of a bolt and shackle, was pulled down as the bellows 
descended, and was thus bent, or ultimately broken, by the pres- 
sure. As the air becomes exhausted in the bellows, the exter- 
nal air rushes in copiously through the pores of the leather, and 
would destroy the effect ; to prevent which, it is necessary to 
inclose the instrument in a tin-case, about an inch wider all 
round, and this being filled with water, effectually excludes the 
air. (See Plate VIII.) 
In order to exhaust the air with greater regularity, a large 
tin vessel was employed, which had already been constructed 
for the institution, to perform experiments in rarified air. This 
vessel being fully exhausted of its air, forms a sort of store of 
vacuity, by communicating with which, smaller vessels may be 
exhausted in a moment, and by opening, therefore, by means of 
a stop-cock, a communication between the vessel and the bel- 
lows, any desired pressure could be readily attained. The hy- 
drostatic bellows is simple, and not much affected by friction, 
and although the sides pull downwards as well as the top, and with 
an effect somewhat varied according to the curve of their ten- 
sion, this irregularity was corrected by a previous trial with a le- 
ver and heavy weight, by means of which the scale of weights 
on the gauge was formed, and I found, that, in this manner, the 
strain on the bar could be estimated to within 20 or 30 pounds. 
This may be sufficient for many practical purposes ; and as it pre- 
sented the readiest means in my power of trying the efiects of 
the transverse strain, and shewing them to the students of the 
School of Arts, I took the opportunity of making with it the 
following experiments with care. A still more accurate appara- 
tus, however, might be formed on a similar plan, by using a sort of 
gasometer and tank, in which friction would almost entirely disap- 
pear, and the pressure could be measured with great exactness 
by the rise of the water, as the gasometer became exhausted. 
Suppose, for example, we had a gasometer 4 feet diameter, and 
6 feet deep, this might exert a pressure of nearly two tons, 
which,, by a single lever, could easily be augmented to 20 or 
30 tons. I mean shortly to try this experiment, and am per- 
VOL. XIT. NO. 23. JANUARY 1825. 
L 
