HOROLOGY. 309 
*loek afcfwering to every probable height of the thermo- 
tneter, the correfponding correction may be obtained. It 
is alfo neceffary to oblerve, that the mean height of the 
thermometer during the interval ought to be ufed. In 
Six's thermometer this height may be eafily. obtained; but 
in thermometers of the common conltruCtion it will be 
more difficult to find this mean. 
It had been found, by repeated experiments, that a brafs 
Tod equal in length to a fecond-pendulum will expand 
or contrad one thoqfandth part of an inch by a change 
of temperature of one degree in Fahrenheit’s thermome¬ 
ter ; and, fince the times of vibration are in a fubduplicate 
ratio of the lengths of the pendulum, hence an expanfion 
or contraction of -fooo P art of an inch will anfwer nearly 
to one fecond daily : therefore, a change of one degree in 
the thermometer will occafion a difference in the rate of 
the clock equal to one fecond daily. Whence, if the 
■clock be fo adjufted as to keep time when the thermome¬ 
ter is at $5°, it will lofe 10 feconds daily when the ther¬ 
mometer is at 65°, and gain as much when it is at 45 0 . 
Hence, the daily variation of the rate of the clock from 
fuminer to winter will be very confiderable. It is true 
indeed that molt pendulums have a nut or regulator at the 
lower end, by which the bob may be railed or lowered a 
determinate quantity ; and therefore, while the height of 
the thermometer is the fame, the rate of the clock will be 
uniform. But fince the ftate of the weather is ever varia¬ 
ble, and as it is impoffible to be raifing or lowering the 
bob of the pendulum at every change of the thermometer, 
feveral contrivances have been propofed, by conftruCting 
a pendulum of different materials, and fo difpofingthem 
that their effeCts may be in oppofite directions, and 
whereby counterbalance each other; and by this means 
the pendulum will continue of the lame length. Thefe 
are called compound pendulums. 
Mercurial Pendulum. The firft of thefe inventions is that 
by the celebrated Mr. George Graham. In this, the rod 
of the pendulum is-a hollow tube, in which a fufficient 
quantity of mercury is put. Mr. Graham firft ufed a 
glafs tube, and the clock to which it Was applied was 
placed in the molt expofed part of’the houfe. It was kept 
conftantly going, Without having the hands or pendulum 
altered, from the 9th of June 1712 to the 14th of Octo¬ 
ber 1725, and its rate was determined by tranfits of fixed 
ftars. Another clock made with extraordinary care, hav¬ 
ing a pendulum about 60 pounds weight, and not vibrat¬ 
ing above one degree and a half from the perpendicular, 
was placed befide the former, in’order the more readily to 
compare them with each other, and that they might both 
be equally expofed. The relult of all the obl'ervations 
was this, that the irregularity of the clock with the quick- 
iilver pendulum exceeded not, when greateft, a fixth part 
of that of the other clock with the common pendulum, 
but for the greateft part of the year not above an eighth 
or ninth part; and even this quantity would have been 
lelfened, had the column of mercury been a little ffiortet: 
for it differed a little the contrary way from the other 
clock, going fafter with heat, and flower with cold. To 
confirm this experiment more, about the beginning of 
July 1723, Mr. Graham took off the heavy pendulum from 
the other clock, and made another with mercury, but with 
this difference, that inftead of a glafs tube he ufed a brals 
one, and varni&ed the inlide to l'ecure it from being in¬ 
jured by the mercury. This pendulum he ufed afterwards, 
and found it about the lame degree of exactnefs as the 
other. 
The ingenious Mr. Gumming objects to the mercurial 
pendulum, that, as the cylinder of mercury moves quicker 
through the air than the mean motion of the pendulum- 
rod, it will be l'ooner heated and cooled by the air than 
the pendulum will. And he is of opinion, that neither 
the mercurial pendulum, nor any thermometer that does 
not. move with the pendulum, can have a proper effeft. 
The Gridiron Pendulum was the invention of the cele¬ 
brated Mr. John Harrifon, about the year 1725. It con- 
Vol. X. No. 662. 
fifts of five round folid rods of fteel r and four of brafs, placed 
in an alternate order, the middle tod being of fteel,by which 
the pendulum ball is fufpended ; thefe rods of brafs and 
fteel, thus placed in an alternate order, and fo connected 
with each other at their ends, that, while the expanfion of 
the fteel rods has a tendency to lengthen the pendulum, the 
expanfion of the brafs rods, afting upwards, tends to 
ihorten it. And thus when the lengths of the brafs and' 
fteel rods are duly proportioned, their expanfions and con¬ 
tractions will exaCtly balance and correCt each other, and 
l"o prefervC the pendulum invariably of the fame length. 
The Simplicity of this ingenious contrivance is much in 
its favour; and the difficulty of adjuftment feems the only 
objection to it. Mr. Harrifon, in his firft machine for mea- 
furing time at fea, applied this combination of wires of 
brafs and fteel, to prevent any alterations by heat or cold ; 
and in the machines or clocks he afterwards made for this 
purpofe, a like method of guarding againft the irregulari¬ 
ties arifing from this caufe was ufed. 
The principle of the gridiron pendulum is thus accu¬ 
rately laid down by Gumming. The expanfion of brafs 
is allowed to exceed that of fteel generally in the propor¬ 
tion of 5 to 3 ; on which fuppofition, if any piece of fteel, 
by the application of a certain degree of heat, expands 
of an iucli, a piece of brafs will, with the fame heat, ex¬ 
pand '5 or half an inch; fo that the expanfion of a bar of 
fteel is equal to the expanfion of a bar of brafs ■§■ of its 
length.. Therefore, when the length of the bars is inverfe- 
ly as the expanfion of the relative metals, the whole ex¬ 
panfion in each bar will be equal, and confequently may 
be applied to balance each other. Whatever be the length 
of the pendulum, three half-lengths of fteel, and as much 
of brals, mutt be added, in order to have as much expan- 
fion upwards as downwards; and, when the whole length 
of the bars is thus afeertained, they may be cut into Inch 
feveral lengths as may belt fuit the purpofe, and avoid the 
inconvenience of having any part of them below the ball 
of the pendulum ; always obferving, that the expanfion of 
the fteel tend downwards, and that of the brafs upwards ; 
in which cafe they will balance each other, without re¬ 
gard to the number of pieces into which they are cut; 
for, the effect of the whole confifts of the effefts of ^11 its 
parts. M. Berthoud’s improved gridiron pendulum is 
correftly reprefented at fig. 33. ABCD is a frame of 
fteel, forming the two outmoft bars of pendulum; it is 
lengthened to VX, and enters the bob T, which is move-' 
able up and down upon C D V X. A B carries the fufpen- 
lion-piece m, fixed to G. EFGH is another frame of 
fteel, with two elbows, E and F, fuftained by the two brafs 
rods PP, PP. The lower extremities of thefe brafs rods 
reft upon the frame ABCD, near CD. The rod KL, 
which is of fteel, with a double elbow or crofs-piece at K, 
is cylindrical, and formed into a ferew at its lower extre¬ 
mity. This rod is fuftained by the two brafs rods RR, 
R R, whole lower extremities reft on the bottom part of 
the'frame E F G H, near G and H. The pieces, S S, S S, S S, 
o round the frames and rods to keep them in their places, 
ut not fo as to hinder their free play upwards and down¬ 
wards ; they are ferewed to the frame ABCD, to keep 
them in their places. The bob T, which may be railed 
and lowered at will upon C D V X, is flip ported by the 
piece M, ferewed on at N,aud the piece M is fuftained at 
rs by the nut g g, which is applied to the fcrew-part L 
of L K, and by turning this ferew the bob is railed or 
lowered. The pieces pq, pq, are faftened to M by ferews 
at 0 and 0, and by two other ferews above at S and S. 
Thefe two pieces, pq, pq , do not prefs fo hard againft the 
frame ABCD as to prevent the bob from rifing or fink¬ 
ing, but they hinder its lhaking from fide to fide. It is 
to be obfervedalfo, that the ferews of this lower piece SS 
do not enter the frame ABCD, but into the pieces pq,pq t 
fo that they offer no impediment to the motion of the 
bob T. The nut^ is graduated, and the little index n 
fhows the number of degrees, or portion of a degree, it is 
turned, for adjuftment. The lower extremity of the bob 
4 K x ’. alfo 
