G. W. Hough on an Automatic printing Barometer. 57 
The present barometer has recently been compensated for 
temperature, by supporting the tube on a brass rod. Previous 
to this, the instrument was set to the height given by the stand- 
ard reduced to 82° Fahrenheit; and differential corrections 
applied to the printed results. 
We propose, however, to construct a siphon which shall have 
the elements of compensation within itself. My attention was 
first called to this method by Dr. James Lewis, who has dis- 
cussed the subject, for a steel siphon, in the September number 
of this Journal. : 
The whole theory of a siphon compensation depends on this 
fundamental proposition, viz: If the atmosphere will support 
30 inches of mercury at 0° Centigrade, at 100° ©. it will support 
30-+30%&€e=30°540 inches; e being equal to 0'018, the expansion 
in volime of mercury for 100° C. If now ina siphon barome- 
ter, the increased length of the whole column, when the temper- 
ature is raised from 0° to 100° C., is equal to 305-0540 
inches ; the surface of the mercury in the short leg of “the siphon 
will remain at the same zero of height for all temperatures, at 30 
Inches of pressure. Put e’=0-016 the expansion of the mercurial 
column in a glass tube for 100° C. : : 
t 2m= length of mercury in the equal legs of the siphon in 
which the diameter is unity 
l= length of intermediate column. 
d= diameter of intermediate column. _ 
h= height for which the compensation is to be computed. 
Then we have the following general formula: 
(2m-+ld? )e'=he. 
It is readily demonstrated that all siphons of the same diameter, 
In the equal legs, will require the same volume of mercury for 
compensation. Dee 
If the siphon be of uniform diameter throughout, it will re- 
quire 33°7 inches of mercury to compensate at 30 inches of 
A tube of this form will hardly give a sufficient length of 
mercury in the short leg. In order to attain the necessary 
length, we connect the two equal legs with a tube of smalle: 
diameter. ean 
Ihe following values have been calculated, to aid in the con- 
Struction of a hs. 
a, m. D. Wi. 
0-70 8°5 in. 0-2 in. 0°62 Ibs. Troy. 
6075 Hg & o4 “* 2°50 “ 
0°80 6-0 “ 0-6 * hg Meee 
0°85 46 « og “ 1000 
Be ae Gio | Buigstes 1s6¢ = * 
Aw. Jour. Set.—Szcoxp Serres, VoL. XLI, No. 121.—Jan., 1866. 
: ; =e 
