3-16 
H O R O 1 
temperature, conftitutes its liability to vary its rate of 
going in different months of the year. Berthoud, who 
attempted to proportion the thicknel's of his pivots to fuit 
the oil with a given balance, fo as to produce an approx¬ 
imation towards a good compenfation, has called this the 
natural compenfation, which is a very proper appellation, as 
it exifts in a certain degree in ail chronometers and watches 
that have oil applied to them. This natural compenfation, 
however, is found to be a very imperfect one, and unfor¬ 
tunately interferes with the effects produced by the better 
compenfations, which, in oppoiition to this, may be called 
-the artificial compenfations. To banifh the interference of 
the variable eflefts produced by oil, is one of the difficul¬ 
ties, and indeed now the principal one, that oppofes the 
efforts of the chronometer-maker, to make die aftion of 
his mechanifm permanently uniform. 
To underltand the effeft of a compcnfation-balance, we 
mull be acquainted with the nature-and ufe of the curb, 
or that mechanifm, laid on the outfide of the upper plate, 
which is called the hide and curb, and is afted upon, 
when needful, by a wheel, fumilhed with a graduated cir¬ 
cle, commonly called the regulator. We have leen that 
the -quicknels of the balance's vibrations depends upon 
the ftrength of the balance-fpring, a Itrong lpring pro¬ 
ducing fwift vibrations, and vice verfa. Now the ftrength 
of the fpring depends upon its thicknefs and upon its 
length. The .thicknefs of the fpring being known, its 
ftrength will vary according to its length, increaling as 
the length is diminiihed, and the contrary. To produce, 
-therefore, in the balance, a certain number of vibrations, 
we furnifh it with a fpring of a ftrength proportionate to 
.the effeft required ; and, in order to attain to this pro¬ 
portionate ftrength the more readily, we apply to the fpring 
a curb, wliich, by lengthening or fhortening the outer 
coil of the lpiral, produces more of lei's ftrength, as may 
be .-required. 
The aftion of the common curb on the balance-fpring, 
-and on the general going of a time-piece, is well known 
to every one who has occasionally moved with his 'key 
that graduated circle called the regulator, to make his 
■watch go fafter or flower. The compenfation-curb is made 
in a different way ; and is applied, fometimes to the ba- 
lance-fpring, fometimes to the balance itfelf, and fome- 
time part!}' .to both. 
A comperfation-curb applied to th o. fpring of the balance, 
is Ihown in the Horology Plate VIII. at fig. 3.7. B B is 
the balance, b the fpring, whole outer coil is fixed to the 
cock A at c, and the inner fpiral is fixed to the axis of 
the balance by the collar in the ulual manner. Upon the 
cock A is a .cylindrical piece e, with a hole to receive the 
upper pivot of the balance. The curb is a piece of fteel, 
■gm, fo adjufted on the piece e, as that, while m is made to 
move along the divifions from F towards L, the part h 
delcribes tire arc of a circle front h towards /; that is to 
fay, k follows a part of the outer coil of the fpiral. To 
keep the curb y»z in its place, a bar might be ferewed 
over it, as d d, fig. 38. 
Fig- 39> is :m under view of the cock, the fpring be¬ 
ing Ihown as when the balance vibrates: b is the arm of 
the curb, on which are two pins, m and n. The outer 
coil of the fpiral fnould have but very little play be¬ 
tween thefe pins, which prevent the end bo of the fpiral 
from moving during the vibrations. It is now evident, 
that, by pulhing the curb towards p, the afting part of 
the fpring is Ihortened, and thereby becomes ftronger, 
and communicates quicker vibrations to the balance; the 
contrary effeft, of courfe, is produced by moving the 
atm of the curb towards g. Thus, then, we remedy the 
natural ftrength or weaknefs of the fpring, or its length¬ 
ening or fhortening by the influence of heat or cold. This 
figure fhows the pins m and n at a fmall diitance from 
each other: if the diftance were very great, they would 
have no effeft, fince the outer coil of the fpiral would not 
■reach either of them during its vibrations. The diftance 
fhould be fuch as to allow the fpiral a little play. If the 
LOGY. 
two pins were moveable, fo that they could approach to 
or recede from each other during the vibrations, this 
would evidently influence the velocity of them, as the 
part bo would then move more or lels with what we have 
called the afting part of the fpring. The receding of the 
pins from each other would make the vibrations flower, 
and their approach would render them quicker. This 
idea has given fife to a compenfation equally Ample and 
ingenious; which we fhall prefently defcribe. 
The equality of a pendulum’s vibrations is difturbed 
by the influence of heat and of cold, or the variations of 
temperature : but the balance is affected by thefe caufes 
in a higher degree. The variations in the pendulum are 
occafioned merely by the dilatation and contraftion of 
the rod ; but thofe of the balance arife not merely from 
the dilatation and contraftion of the balance itfelf, but 
from fimilar effefts of heat and cold upon the balance- 
fpring. Heat enlarges the balance, and makes its vibra¬ 
tions flower; heat at the fame time lengthens the fpring, 
makes it weaker, and thus the vibrations are flower Hill. 
Cold has the contrary effeft; and experiment has ihown 
the influence of temperature to be fo great, that a watch 
made to keep time in the greateft heat of fummer will 
gain 5 or 6 minutes a-day in the depth of winter; and 
the contrary. But it fhould be noted, that in thefe ex¬ 
periments the watches were expofed during winter night 
and day to the open air ; a watch worn in the fob would 
not be affected in this manner. Stiff, however, it is very 
necelfary to contrive fome compenfation for thefe effefts 
upon time-pieces which are required to meafure time with 
great exaftnels. 
This compenfation may be produced various ways, and 
different means are employed according to the greater or 
lefs degree of exaftnels required. A Ample compenfa¬ 
tion applied to the balance-fpring is fufficient to make a 
watch for common ufe perform extremely well; but the 
compenfation by the balance is the molt complete, and is- 
rnoftly applied to marine or longitude watches. 
The. better to underltand the effeft of a compenfation 
applied to the balance-fpring, let us Aril obferve the ac¬ 
tion of heat and cold upon two metals combined. AB, 
Ag. 40, reprelents a thin lamina of which the Ade a a is' 
l’uppoled to be brafs, and the fide bb fteel. This com¬ 
pound lamina, being expofed to heat, will dilate; but by' 
the Table before given we obferve that brafs dilates by 
heat, as well as contrafts by cold, almoft twice as much 
as fteel; conlequently the fide a a, by its greater expan- 
Aon, would force the part B towards Df and, if expofed 
to cold, the brafs would contraft more than the fteel, l'o 
as to make the part B tend or bend towards C. Now fup- 
pole this piece of compound metal to be faflrioned into 
the lhape ABCD, fig. 41 ; the outfide or dark part being 
the fteel, as in the former figure, and the inner part brafs. 
Suppofe the end B immoveably fixed; then the effeft of 
heat will be to make the part A take a direction towards 
c, by which D mult move towards e. Now the heat, aft¬ 
ing alfo upon CD, will make D approach ftill more to¬ 
wards e ; lo that the total motion of D becomes very con- 
fiderable, and that more or lefs according to the degree 
of heat to which ABCD is expofed. The direftion of 
the motion of D, when ABCD is expofed to edd, is 
contrary; D will take a direftion towards f, or, what is 
the fame thing, it will approach B; while heat will make 
D recede from B. The effeft of heat is to make ABCD 
open, of cold to make it dole. Now we have noticed, in. 
deferibing fig. 30, that the lpring having lefs play between 
two pins would thereby vibrate fafter, and the contrary. 
If then we apply ABCD to the curb fo that B may be 
fixed by a lcrew, while the part D may be free to move 
as afted upon by the different degrees of temperature, as 
Ihown at fig. 42, it is evident that the play of the balance- 
fpring mult vary according to the influence of the heat 
or cold upon the compenfator. By heat, D will approach 
towards the fixed pin a, and the fpring will have lefs 
play, and make quicker vibrations ■, by cold it will recede. 
