a vibrating body suspended from a fixed point. For the history of this invention, see the article CLOCK.

The theory of the pendulum depends on that of the inclined plane. Hence, in order to understand the nature of the pendulum, it will be necessary to premise some of the properties of this plane; referring, however, to Inclined Planes, and to the article MECHANICS, for the demonstration.

I. Let AC (fig. 1.) be an inclined plane, AB its perpendicular height, and D any heavy body: then the force which impels the body D to descend along the inclined plane AC, is to the absolute force of gravity as the height of the plane AB is to its length AC; and the motion of the body will be uniformly accelerated.

II. The velocity acquired in any given time by a body descending on an inclined plane AC, is to the velocity acquired in the same time by a body falling freely and perpendicularly, as the height of the plane AB to its length AC. The final velocities will be the same; the spaces described will be in the same ratio; and the times of description are as the spaces described.

III. If a body descend along several contiguous planes, planes, \(AB, BC, CD\), (fig. 2,) the final velocity, namely, that at the point \(D\), will be equal to the final velocity in descending through the perpendicular \(AE\), the perpendicular heights being equal. Hence, if these planes be supposed indefinitely short and numerous, they may be conceived to form a curve, and therefore the final velocity acquired by a body in descending through any curve \(AF\), will be equal to the final velocity acquired in descending through the planes \(AB, BC, CD\), or to that in descending through \(AE\), the perpendicular heights being equal.

IV. If from the upper or lower extremity of the vertical diameter of a circle a cord be drawn, the time of descent along this cord will be equal to the time of descent through the vertical diameter; and therefore the times of descent through all cords in the same circle, drawn from the extremity of the vertical diameter, will be equal.

V. The times of descent of two bodies through two planes equally elevated will be in the subduplicate ratio of the lengths of the planes. If, instead of one plane, each be composed of several contiguous planes similarly placed, the times of descent along these planes will be in the same ratio. Hence, also, the times of describing similar arches of circles similarly placed will be in the subduplicate ratio of the lengths of the arches.

VI. The same things hold good with regard to bodies projected upward, whether they ascend upon inclined planes or along the arches of circles.

The point or axis of suspension of a pendulum is that point about which it performs its vibrations, or from which it is suspended.

The centre of oscillation is a point in which, if all the matter in a pendulum were collected, any force applied at this centre would generate the same angular velocity in a given time as the same force when applied at the centre of gravity.

The length of a pendulum is equal to the distance between the axis of suspension and centre of oscillation.

Let \(PN\) (fig. 3,) represent a pendulum suspended from the point \(P\); if the lower part \(N\) of the pendulum be raised to \(A\), and let fall, it will by its own gravity descend through the circular arch \(AN\), and will have acquired the same velocity at the point \(N\) that a body would acquire in falling perpendicularly from \(C\) to \(F\), and will endeavour to go off with that velocity in the tangent \(ND\); but being prevented by the rod or cord, will move through the arch \(NB\) to \(B\), where, losing all its velocity, it will by its gravity descend through the arch \(BN\), and, having acquired the same velocity as before, will ascend to \(A\). In this manner it will continue its motion forward and backward along the arch \(ANB\), which is called an oscillatory or vibratory motion; and each swing is called a vibration.

Prop. I. If a pendulum vibrates in very small circular arches, the times of vibration may be considered as equal, whatever be the proportion of the arches.

Let \(PN\) (fig. 4,) be a pendulum; the time of describing the arch \(AB\) will be equal to the time of describing \(CD\); these arches being supposed very small.

Join \(AN, CN\); then since the times of descent along all cords in the same circles, drawn from one extremity of the vertical diameter, are equal; therefore the cords \(AN, CN\), and consequently their doubles, will be described in the same time; but the arches \(AN, CN\) being supposed very small, will therefore be nearly equal to their chords: hence the times of vibrations in these arches will be nearly equal.

Prop. II. Pendulums which are of the same length vibrate in the same time, whatever be the proportion of their weights.

This follows from the property of gravity, which is always proportional to the quantity of matter, or to its inertia. When the vibrations of pendulums are compared, it is always understood that the pendulums describe either similar finite arcs, or arcs of evanescent magnitude, unless the contrary is mentioned.

Prop. III. If a pendulum vibrates in the small arc of a circle, the time of one vibration is to the time of a body's falling perpendicularly through half the length of the pendulum, as the circumference of a circle is to its diameter.

Let \(PE\) (fig. 5,) be the pendulum which describes Fig. 5, the arch \(ANC\) in the time of one vibration; let \(PN\) be perpendicular to the horizon, and draw the cords \(AC, AN\); take the arc \(Ee\) infinitely small, and draw \(EFG, efg\) perpendicular to \(PN\), or parallel to \(AC\); describe the semicircle \(BGN\), and draw \(er, gs\) perpendicular to \(EG\); now let \(t\) be the time of descending through the diameter \(2PN\), or through the chord \(AN\): Then the velocities gained by falling through \(2PN\), and by the pendulum's descending through the arch \(AE\), will be as \(\sqrt{2PN}\) and \(\sqrt{BF}\); and the space described in the time \(t\), after the fall through \(2PN\), is \(4PN\). But the times are as the spaces divided by the velocities.

Therefore \(\frac{4PN}{\sqrt{2PN}} = \frac{Ee}{\sqrt{BF}} : t :: \frac{Ee}{\sqrt{BF}} : \text{time of describing } Ee = \frac{t \times Ee}{2\sqrt{2PN} \times BF}\).

But in the similar triangles \(PEF, Ere\), and \(KGF, Ggs\),

As \(PE = PN : EF :: Ee : er = \frac{EF}{PN} \times Ee\);

And \(KG = KD : FG :: Gg : Gs = \frac{FG}{KD} \times Gg\).

But \(er = Gs\); therefore \(\frac{EF}{PN} \times Ee = \frac{FG}{KD} \times Gg\).

Hence \(Ee = \frac{PN \times FG}{KD \times EF} \times Gg\).

And by substituting this value of \(Ee\) in the former equation, we have the time of describing \(Ee = \frac{1 \times PN \times \sqrt{BF} \times FN \times Gg}{2KD \times \sqrt{PN + PF} \times \sqrt{FN}}\).

But by the nature of the circle \(FG = \sqrt{BF} \times FN\), and \(EF = \sqrt{PN + PF + FN}\).

Hence, by substitution, we obtain the time of describing

\[Ee = \frac{1 \times PN \times \sqrt{BF} \times FN \times Gg}{2KD \times \sqrt{PN + PF} \times \sqrt{FN}} = \frac{1 \times \sqrt{PN} \times Gg}{2KD \times \sqrt{PN + PF}} = \frac{1 \times \sqrt{PN}}{2BN \times \sqrt{PN - NF}} \times Gg.\]

Put \(NF\), in its mean quantity for all the arches \(Gg\), is nearly equal to \(NK\); for if the semicircle described on the diameter \(BN\), which corresponds to the whole arch \(AN\), be divided into Pendulum into an indefinite number of equal arches, Gg, &c., the sum of all the lines NF, will be equal to as many times NK, as there are arches in the semicircle equal to Gg; therefore the time of describing \( E = \frac{1}{2} \times \sqrt{\frac{2PN}{2BN \times \sqrt{2PN - NK}}} \times Gg \). Whence the time of describing the arch AED

\[ = \frac{1}{2} \times \sqrt{\frac{2PN}{2BN \times \sqrt{2PN - NK}}} \times BGN; \]

and the time of describing the whole arch ADC, or the time of one vibration, is

\[ = \frac{1}{2} \times \sqrt{\frac{2PN}{2BN \times \sqrt{2PN - NK}}} \times 2BGN. \]

But when the arch ANC is very small, NK vanishes, and then the time of vibration in a very small arc is

\[ = \frac{1}{2} \times \sqrt{\frac{2PN}{2BN}} \times 2BGN = \frac{1}{2} \times \frac{2BGN}{BN}. \]

Now if \( t \) be the time of descent through 2PN; then since the spaces described are as the squares of the times, \( \frac{t}{4} \) will be the time of descent through \( \frac{t}{4} \)PN: therefore the diameter BN is to the circumference 2BGN, as the time of falling through half the length of the pendulum is to the time of one vibration.

**Prop. IV.** The length of a pendulum vibrating seconds is to twice the space through which a body falls in one second, as the square of the diameter of a circle is to the square of its circumference.

Let \( d = \text{diameter of a circle} = 1, c = \text{circumference} = 3.14159, \&c., t \) to the time of one vibration, and \( p \) the length of the corresponding pendulum; then by the last proposition \( c : d :: 1'' : \frac{d}{c} = \text{time of falling through half the length of the pendulum}. \)

Let \( s = \text{space described by a body falling perpendicularly in the first second}; \) then since the spaces described are in the subduplicate ratio of the times of description, therefore

\[ 1'' : \frac{d}{c} :: \sqrt{s} : \sqrt{\frac{t}{p}}. \]

Hence \( c^2 : d^2 :: 2s : p. \)

It has been found by experiment, that in latitude \( 51^\circ 20' \) a body falls about 16.11 feet in the first second: hence the length of a pendulum vibrating seconds in that latitude is

\[ = \frac{32.22}{3.14159^2} = 3 \text{ feet } 3.174 \text{ inches}. \]

**Prop. V.** The times of the vibrations of two pendulums in similar arcs of circles are in a subduplicate ratio of the lengths of the pendulums.

Let PN, PO (fig. 6.) be two pendulums vibrating in the similar arcs AB, CD; the time of a vibration of the pendulum PN is to the time of a vibration of the pendulum PO in a subduplicate ratio of PN to PO.

Since the arcs AN, CO are similar and similarly placed, the time of descent through AN will be to the time of descent through CO in the subduplicate ratio of AN to CO: but the times of descent through the arcs AN and CO are equal to half the times of vibration of the pendulums PN, PO respectively. Hence the time of vibration of the pendulum PN in the arc AB is to the time of vibration of the pendulum PO in the similar arc CD in the subduplicate ratio of AN to CO: and since the radii PN, PO are proportional to the similar arcs AN, CO, therefore the time of vibration of the pendulum PN will be to the time of vibration of the pendulum PO in a subduplicate ratio of PN to PO.

If the length of a pendulum vibrating seconds be 39.174 inches, then the length of a pendulum vibrating half seconds will be 9.793 inches. For \( 1'' : \frac{1}{4}'' :: \sqrt{39.174} : \sqrt{x} \); and \( x : \frac{1}{4} :: 39.174 : x \). Hence

\[ x = \frac{39.174}{4} = 9.793. \]

**Prop. VI.** The lengths of pendulums vibrating in the same time, in different places, will be as the forces of gravity.

For the velocity generated in any given time is directly as the force of gravity, and inversely as the quantity of matter. Now the matter being supposed the same in both pendulums, the velocity is as the force of gravity; and the space passed through in a given time will be as the velocity; that is, as the gravity.

Cor. Since the lengths of pendulums vibrating in the same time in small arcs are as the gravitating forces, and as gravity increases with the latitude on account of the spheroidal figure of the earth and its rotation about its axis; hence the length of a pendulum vibrating in a given time will be variable with the latitude, and the same pendulum will vibrate slower the nearer it is carried to the equator.

**Prop. VII.** The time of vibrations of pendulums of the same length, acted upon by different forces of gravity, are reciprocally as the square roots of the forces.

For when the matter is given, the velocity is as the force and time; and the space described by any given force is as the force and square of the time. Hence the lengths of pendulums are as the forces and the squares of the times of falling through them. But these times are in a given ratio to the times of vibration; whence the lengths of pendulums are as the forces and the squares of the times of vibration. Therefore, when the lengths are given, the forces will be reciprocally as the square of the times, and the times of vibration reciprocally as the square roots of the forces.

Cor. Let \( p = \text{length of pendulum}, g = \text{force of gravity}, \) and \( t = \text{time of vibration}. \) Then since \( p = g \times t^2, \) Hence \( g = p \times \frac{1}{t^2} \); and \( t = \sqrt{\frac{p}{g}}. \)

That is, the forces in different places are directly as the lengths of the pendulums, and inversely as the square roots of the times of vibration; and the times of vibration are directly as the square roots of the lengths of the pendulums, and inversely as the square roots of the gravitating forces.

**Prop. VIII.** A pendulum which vibrates in the arch of a cycloid describes the greatest and least vibrations in the same time.

This property is demonstrated only on a supposition that the whole mass of the pendulum is concentrated in a point; but this cannot take place in any really vibrating body; and when the pendulum is of finite magnitude, there is no point given in position which determines the length of the pendulum; on the contrary the centre of oscillation will not occupy the same place in the given body, when describing different parts of the tract it moves through, but will continually be moved in respect of the pendulum itself during its vibration. It may, however, be observed, that Huyghens, aware that Pendulum: a pendulum ball suspended at the end of a thread vibrating between cycloids, would not describe a cycloid with its centre of oscillation, gave a very beautiful and simple method of suspension, which secured its vibrations in that curve. Harrison, whose authority is next, insists on the advantage of wide vibrations, and in his own clocks, he always used cycloidal cheeks. This circumstance has prevented any general determination of the time of vibration in a cycloidal arc, except in the imaginary case referred to.

There are many other obstacles which concur in rendering the application of this curve to the vibration of pendulums designed for the measures of time the source of errors far greater than those which by its peculiar property it is intended to obviate; and it is now wholly disused in practice.

Although the times of vibration of a pendulum in different arches be nearly equal, yet from what has been said, it will appear, that if the ratio of the least of these arches to the greatest be considerable, the vibrations will be formed in different times; and the difference, though small, will become sensible in the course of one or more days. In clocks used for astronomical purposes, it will therefore be necessary to observe the arc of vibration; which if different from that described by the pendulum when the clock keeps time, there a correction must be applied to the time shown by the clock. This correction, expressed in seconds of time, will be equal to the half of three times the difference of the square of the given arc, and of that of the arc described by the pendulum when the clock keeps time, these arcs being expressed in degrees; and so much will the clock gain or lose according as the first of these arches is less or greater than the second.

Thus, if a clock keep time when the pendulum vibrates in the arch of $3^\circ$, it will lose $10\frac{1}{2}$ seconds daily in an arch of $4^\circ$ degrees.

For $\frac{4^2 - 3^2}{4} \times \frac{1}{4} = \frac{7}{4} \times \frac{1}{4} = 10\frac{1}{2}$ seconds.

The length of a pendulum rod increases with heat; and the quantity of expansion answering to any given degree of heat is experimentally found by means of a pyrometer; but the degree of heat at any given time is shown by a thermometer: hence that instrument should be placed within the clock case at a height nearly equal to that of the middle of the pendulum; and its height, for this purpose, should be examined at least once a day. Now by a table constructed to exhibit the daily quantity of acceleration or retardation of the clock answering to every probable height of the thermometer, the corresponding correction may be obtained. It is also necessary to observe, that the mean height of the thermometer during the interval ought to be used. In Six's thermometer this height may be easily obtained; but in thermometers of the common construction it will be more difficult to find this mean.

It had been found, by repeated experiments, that a brass rod equal in length to a second pendulum will expand or contract $\frac{1}{500}$ part of an inch by a change of temperature of one degree in Fahrenheit's thermometer; and since the times of vibration are in a subduplicate ratio of the lengths of the pendulum, hence an expansion or contraction of $\frac{1}{500}$ part of an inch will answer nearly to one second daily: therefore a change of one degree in the thermometer will occasion a difference in the rate of the clock equal to one second daily.

Whence, if the clock be so adjusted as to keep time when the thermometer is at $55^\circ$, it will lose $10$ seconds daily when the thermometer is at $65^\circ$, and gain as much when it is at $45^\circ$.

Hence the daily variation of the rate of the clock from summer to winter will be very considerable. It is true indeed that most pendulums have a nut or regulator at the lower end, by which the bob may be raised or lowered a determinate quantity; and therefore while the height of the thermometer is the same, the rate of the clock will be uniform. But since the state of the weather is ever variable, and as it is impossible to be raising or lowering the bob of the pendulum at every change of the thermometer, therefore the correction formerly mentioned is to be applied. This correction, however, is in some measure liable to a small degree of uncertainty; and in order to avoid it altogether, several contrivances have been proposed by constructing a pendulum of different materials, and so disposing them that their effects may be in opposite directions, and thereby counterbalance each other; and by this means the pendulum will continue of the same length.

Mercurial Pendulum: The first of these inventions is that by the celebrated Mr George Graham. In this, the rod of the pendulum is a hollow tube, into which a sufficient quantity of mercury is introduced. Mr Graham first used a glass tube, and the clock to which it was applied was placed in the most exposed part of the house. It was kept constantly going, without having the hands or pendulum altered, from the 9th June 1722 to the 14th of October 1725, and its rate was determined by transits of fixed stars. Another clock made with extraordinary care, having a pendulum about 60 pounds weight, and not vibrating above one degree and a half from the perpendicular, was placed beside the former, in order the more readily to compare them with each other, and that they might both be equally exposed. The result of all the observations was this, that the irregularity of the clock with the quicksilver pendulum exceeded not, when greatest, a fifth part of that of the other clock with the common pendulum, but for the greatest part of the year not above an eighth or ninth part; and even this quantity would have been lessened, had the column of mercury been a little shorter: for it differed a little the contrary way from the other clock, going faster with heat and slower 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 instead of a glass tube he used a brass one, and varnished the inside to secure it from being injured by the mercury. This pendulum he used afterwards, and found it about the same degree of exactness as the other.

The Gridiron Pendulum is an ingenious contrivance for the same purpose. Instead of one rod, this pendulum is composed of any convenient number of rods, as five, seven, or nine; being so connected, that the effect of one set of them counteracts that of the other set; and therefore, being properly adjusted to each other, the centres of suspension and oscillation will always be equidistant. Fig. 7 represents a gridiron pendulum composed of nine rods, steel and brass alternately. The two outer rods AB, CD, which are of steel, are fastened to the cross pieces AC, BD by means... PENDULUM

The next two rods, EF, GH, are of brass, and are fastened to the lower bar BD, and to the second upper bar EG. The two following rods are of steel, and are fastened to the cross bars EG and IK. The two rods adjacent to the central rod being of brass, are fastened to the cross pieces IK and LM; and the central rod, to which the ball of the pendulum is attached, is suspended from the cross piece LM, and passes freely through a perforation in each of the cross bars IK, BD. From this disposition of the rods, it is evident that, by the expansion of the extreme rods, the cross piece BD, and the two rods attached to it, will descend; but since these rods are expanded by the same heat, the cross piece EG will consequently be raised, and therefore also the two next rods; but because these rods are also expanded, the cross bar IK will descend; and by the expansion of the two next rods, the piece LM will be raised a quantity sufficient to counteract the expansion of the central rod. Whence it is obvious, that the effect of the steel rods is to increase the length of the pendulum in hot weather, and to diminish it in cold weather, and that the brass rods have a contrary effect upon the pendulum. The effect of the brass rods must, however, be equivalent not only to that of the steel rods, but also to the part above the frame and spring, which connects it with the cock, and to that part between the lower part of the frame and the centre of the ball.

M. Thiont. Another excellent contrivance for the same purpose is described in a French author on clock-making. It was used in the north of England by an ingenious artist about 40 years ago. This invention is as follows:

A bar of the same metal with the rod of the pendulum, and of the same dimensions, is placed against the back part of the clock case; from the top of this a part projects, to which the upper part of the pendulum is connected by two fine pliable chains or silk strings, which just below pass between two plates of brass, whose lower edges will always terminate the length of the pendulum at the upper end. These plates are supported on a pedestal fixed to the back of the case. The bar rests upon an immovable base at the lower part of the case; and is inserted into a groove, by which means it is always retained in the same position. From this construction, it is evident that the extension or contraction of this bar, and of the rod of the pendulum, will be equal, and in contrary directions. For suppose the rod of the pendulum to be expanded any given quantity by heat; then, as the lower end of the bar rests upon a fixed point, the bar will be expanded upwards, and raise the upper end of the pendulum just as much as its length was increased; and hence its length below the plates will be the same as before.

Of this pendulum, somewhat improved by Mr Croftswaite watch and clock-maker, Dublin, we have the following description in the Transactions of the Royal Irish Academy, 1788.—“A and B (fig. 8.) are two rods of steel forged out of the same bar, at the same time, of the same temper, and in every respect similar. On the top of B is formed a gibbet C; this rod is firmly supported by a steel bracket D, fixed on a large piece of marble E, firmly let into the wall F, and having liberty to move freely upwards between cross staples of brass, 1, 2, 3, 4, which touch only in a point in front and rear (the staples having been carefully formed for that purpose); to the other rod is firmly fixed by its centre the lens G, of 24 pounds weight, although it should in strictness be a little below it. This pendulum is suspended by a short steel spring on the gibbet at C; all which is entirely independent of the clock. To the back of the clock plate I are firmly screwed two cheeks nearly cycloidal at K, exactly in a line with the centre of the verge L. The maintaining power is applied by a cylindrical steel fluid, in the usual way of regulators, at M. Now, it is very evident, that any expansion or contraction that takes place in either of these exactly similar rods, is instantly counteracted by the other; whereas in all compensation pendulums composed of different materials, however just calculation may seem to be, that can never be the case, as not only different metals, but also different bars of the same metal, that are not manufactured at the same time, and exactly in the same manner, are found by a good pyrometer to differ materially in their degrees of expansion and contraction, a very small change affecting one and not the other.”

The expansion or contraction of straight grained fir Fir Pendulum wood lengthwise, by change of temperature, is so small, that it is found to make very good pendulum rods. The wood called sapadillo is said to be still better. There is good reason to believe, that the previous baking, varnishing, gilding, or soaking of these woods in any melted matter, only tends to impair the property that renders them valuable. They should be simply rubbed on the outside with wax and a cloth. In pendulums of this construction the error is greatly diminished, but not taken away.

Angular PENDULUM, is formed of two pieces or legs Angular like a sector, and is suspended by the angular point. Pendulum. This pendulum was invented with a view to diminish the length of the common pendulum, but at the same time to preserve or even increase the time of vibration. In this pendulum, the time of vibration depends on the length of the legs, and on the angle contained between them conjointly, the duration of the time of vibration increasing with the angle. Hence a pendulum of this construction may be made to oscillate in any given time. At the lower extremity of each leg of the pendulum is a ball or bob as usual. It may be easily shown, that in this kind of a pendulum, the squares of the times of vibration are as the secants of half the angle contained by the legs: hence if a pendulum of this construction vibrates half seconds when its legs are close, it will vibrate whole seconds when the legs are opened, so as to contain an angle equal to $15^\circ 24'$.

The Conical or Circular PENDULUM, is so called Conical or Circular from the figure described by the string or ball of the Pendulum. This pendulum was invented by Mr Huygens, and is also claimed by Dr Hoeke.

In order to understand the principles of this pendulum, it will be necessary to premise the following lemma, viz. the times of all the circular revolutions of a heavy globular body, revolving within an inverted hollow paraboloid, will be equal, whatever be the radii of the circle described by that body.

In order therefore, to construct the pendulum so that its ball may always describe its revolutions in a paraboloid surface, it will be necessary that the rod of the pendulum be flexible, and that it be suspended in such Pendulum. Such a manner as to form the evolute of the given parabola. Hence, let KH (fig. 9.) be an axis perpendicular to the horizon, having a pinion at K moved by the leaf wheel in the train of the clock; and a hardened steel point at H moving in an agate pivot, to render the motion as free as possible. Now, let it be required that the pendulum shall perform each revolution in a second, then the paraboloid surface it moves in must be such whose latus rectum is double the length of the common half second pendulum. Let O be the focus of the parabola MEC, and MC the latus rectum; and make AE = MO = ½MC = the length of a common half second pendulum. At the point A of the verge, let a thin plate AB be fixed at one end, and at the other end B let it be fastened to a bar or arm BD perpendicular to DH, and to which it is fixed at the point D. The figure of the plate AB is that of the evolute of the given parabola MEC.

The equation of this evolute, being also that of the semicubical parabola, is \( \frac{27}{16} p x^2 = y^3 \).—Let \( \frac{27}{16} p = P \); then \( P x^2 = y^3 \), and in the focus \( P = 2y \). In this case \( 2x^2 = y^2 = \frac{1}{4} P^2 \); hence \( x^2 = \frac{1}{8} P^2 \), and \( x = P \sqrt{\frac{1}{8}} = \frac{27}{16} p \sqrt{\frac{1}{3}} \) is the distance of the focus from the vertex A.—By assuming the value of \( x \), the ordinates of the curve may be found; and hence it may be easily drawn.

The string of the pendulum must be of such a length that when one end is fixed at B, it may lie over the plate AB, and then hang perpendicular from it, so that the centre of the bob may be at E when at rest. Now, the verge KH being put into motion, the ball of the pendulum will begin to gyrate, and thereby conceive a centrifugal force which will carry it out from the axis to some point F, where it will circulate seconds or half seconds, according as the line AE is 9.8 inches, or 2½ inches, and AB answerable to it.

One advantage possessed by a clock having a pendulum of this construction is, that the second hand moves in a regular and uniform manner, without being subject to those jerks or starts as in common clocks; and the pendulum is entirely silent.

Theory has pointed out several other pendulums, known by the names of Elliptical, Horizontal, Rotatory, &c., pendulums. These, however, have not as yet attained that degree of perfection as to supplant the common pendulum.

Observing that both the gridiron and mercurial pendulums are subject to many inconveniences and errors, Mr. Kater has attempted to construct one possessing such properties in respect of cheapness and accuracy as he thinks might justify giving it the preference to any other. As wood possesses a less degree of expansibility by means of heat than any other substance; on this account, if it could be rendered quite impervious to moisture, it would be the best of all substances for the rod of a pendulum; and as it also appears that zinc, above all other metals, possesses the greatest degree of expansibility by means of heat, he considered it the best substance which could be employed for a compensation.

His next object was to institute a set of delicate experiments, in order to ascertain the precise degree of the expansibility of wood by the application of heat, and he discovered by the use of a pyrometer, that a rod of very dry, well seasoned white wood, four feet long, three-fourths of an inch broad, and one-fourth of an inch thick, when exposed in an oven to the temperature of 235°, had contracted. Being again put into the oven, where it was permitted to remain for a long time, till it became a little discoloured, with a view to dissipate the whole of the moisture, it was placed in the pyrometer, and allowed to remain till it reached the temperature of the room, or 49°, when it was found to have contracted 0.0205 of an inch with 186° of Fahrenheit, from which we obtain by proportion 0.0049 of an inch for the expansion of one foot with 180° difference of temperature. Thus,

\[ \frac{0.0205 \times 180}{186} = \frac{0.0198}{4} = 0.0049. \]

But for a general description of this pendulum, and a full account of the manner in which it is constructed, we must refer our readers to the inventor's own paper, Nichol. Jour. vol. xx. p. 214.

Besides the use of the pendulum in measuring time, it has also been suggested as a proper standard for measures of length. See Measure.