355

GREENWICH TIME AND ITS TELEGRAPHIC
DISTRIBUTION.

BY WILLIAM ELLIS, F.R.A.S., SUPERINTENDENT OF THE TIME DEPARTMENT, ROYAL OBSERVATORY, GREENWICH.

THE

THE object of our present paper is to describe that system by which Greenwich time, as found by astronomical observation at the Royal Observatory at Greenwich, is daily transmitted, by telegraphic aid, to distant parts of the kingdom; a subsidiary use of the telegraph which, although not directly contemplated in its establishment, yields practical advantages of no small value.

But before directly proceeding to consider its utilitarian applications we must give some account of the manner of reckoning and determining time. The astronomical considerations involved are, however, so fully treated in works on astronomy that we need only concern ourselves here with the more practical aspect of the subject.

When the sun reaches its greatest daily altitude in the heavens we call the time noon, and the interval which elapses between one noon and the next we call a solar day. But the natural solar day thus measured is (owing to the varying motion of the earth in its elliptic orbit, and the inclination of its axis of revolution to the same orbit) to a slight extent variable. Its length oscillates between certain small limits, which renders the ordinary use of such a day for many reasons inconvenient. The inequality is fortunately small as compared with the length of the day, so that its use in practice is avoided by assuming the existence of an artificial solar day-one of uniform length, and consequently better adapted to the wants of mankind. It is known as the mean solar day. Natural solar time (that shown by a sun-dial and variously called "true" or "apparent" solar time) is sometimes rather before and sometimes rather after mean solar time (that shown by a clock). Four times in each year they are together. The difference usually existing between them, which amounts to as much as 16 minutes in

the month of November, is the "equation of time" of our almanacs. Its amount for each day at noon to the nearest second of time is contained in common almanacs usually under the heading either of "clock before sun" or "clock after sun :” for greater accuracy reference must be made to the Nautical Almanac.

Having shown the relation existing between apparent or sun-dial time and mean solar or clock time, we see how it is that, taking time from a sun-dial and allowing for the equation of time, ordinary clock time is obtained. But a sun-dial is useless for any accurate determination; and of other instruments and methods, giving something more of accuracy, space will not allow us to speak. We must hasten to describe that special instrument, the "transit instrument," which is always employed in fixed astronomical observatories. This instrument consists of a telescope fixed at its centre to a cross axis supported at the extremities on bearings firmly fixed in an east and west position, so that on turning the telescope on its axis it points successively to all parts of the meridian (that imaginary great circle in the heavens which corresponds to the brazen meridian of a celestial globe, and at which the heavenly bodies attain, between rising and setting, their greatest altitude). In order that it may do this precisely, the line of sight of the telescope must be at right angles exactly to the cross axis, and the axis itself must be truly level and also precisely east and west; but no instrument, if placed in exact position, will long remain so. It is therefore usual to register its small deviations, and apply corrections as necessary to the observations. The instrument made use of at Greenwich is the one meridian instrument of the Observatory, the noble transit-circle (designed by the present Astronomer Royal). Such an instrument is used for many purposes besides the determination of time, but it is with this use of it only that we have to do here. On looking into its telescope we see a number of delicate vertical threads across which objects must pass in their transit through the field. The centre thread represents the meridian, the others being uniformly distributed, an equal number on each side. The time at which, by the sidereal clock (always that employed in an observatory), the object to be observed is upon each thread being noted, the mean of the observed times gives a more accurate value of the meridian transit. Usually an observer counts the beats (seconds) of his clock, and estimates the time at which the object is on each thread; but at Greenwich this method is no longer pursued, for by means of the chronograph (brought into use in the year 1854) all transits are registered by galvanism. Of this instrument we cannot here attempt description further than to say, that by its means

the sidereal clock is made to register its seconds by punctures on paper fixed on a cylinder which revolves uniformly, and on which the observer at the transit-circle is able similarly to register any transits he may make. He has only to press a finger-piece attached to the transit-circle to effect the necessary registration. The punctures made by the clock form a scale by which the times corresponding to the punctures made by the observer are easily ascertained. The times for any transit being extracted from the register, and the mean taken, it is further corrected as necessary for the small deviations of the instrument, and finally for "personal equation," that slight constant difference found to exist between even the best observers, by taking account of which observations are reduced to one standard. By this treatment we obtain the clock times of transits such as would have been found had the transit-circle been in perfect adjustment and all observations been made by one person.

Now it is of course possible to observe the sun at noon with the transit instrument and mean-time clock (taking the mean of the transits of the preceding and following borders), and, by aid of the equation of time, infer the error of the mean-time clock. But this is not the way an astronomer proceeds: he refers to the stars. Time can be thereby more accurately determined, and stars may be seen at some part of most nights, whilst the sun will often be invisible at noon for many days together. But the sidereal day differs from the solar day. The length of the solar day depends on the revolution of the earth on its axis and its advance in its orbit round the sun; that of the sidereal day on the revolution of the earth on its axis alone. The consequence is that the sidereal day is shorter than the solar day by nearly 4 minutes of time; and therefore a sidereal clock, or one that completes 24 hours in a sidereal day, must be used. The sidereal day commences when the "first point of Aries" (on a celestial globe one of the points in which the ecliptic cuts the equator) is on the meridian. Mean solar and sidereal time coincide once only in each year, on some certain day in spring. At other times they differ; for as the stars shift once round in a year as regards the sun or solar day, so the sidereal clock in the same period shifts once through the 24 hours as respects the mean solar clock, the relation between the two being very exactly known. The position of any star is known by its right ascension and declination. In a general sense these correspond in the heavens to longitude and latitude as measured on the earth. In determining time we have to do with its right ascension, which is reckoned from the celestial meridian of the first point of Aries. And although the stars are, as it were, "fixed," it is still matter of calculation

to obtain their places for any given time, principally on account of small changes of the celestial planes of reference which astronomers are compelled to use. Suffice it therefore

to say, that from the Nautical Almanac may be obtained the places of a great number of stars for every day of the year. An observation of any suitable star of this list being made with the transit-circle, and treated as before described, the result compared with the Nautical Almanac right ascension gives the error of the sidereal clock, knowing which the error of the mean-time clock can be easily found.

It was always necessary that time should be regularly determined at the Royal Observatory for its own special purposes. In the year 1833, however, an attempt was successfully made to give time by signal to the outer world. A pole, carrying a black ball about 5 feet in diameter, was then fixed on one of the turrets of the ancient portion of the Observatory. Being raised on the pole very shortly before 1 h. (half-way up at 5 m. before 1 h., and full up at 3 m. before 1 h.), the ball is dropped precisely at the instant of 1 h. Greenwich mean time. Its fall is at first rapid (this start is the proper instant to note); afterwards a piston, attached to a rod extending from the ball downwards, entering an open-topped cylinder, the gradual escape of the compressed air so checks its fall that it is terminated quite gently. The ball was for many years dropped by hand (by pressure on a trigger which released the piston), but since the year 1852 it has been dropped by automatic galvanic action, as will be hereafter described.

Before proceeding to speak of the wider distribution of Greenwich time, we must say a few words in reference to local time. This is merely time as determined at any place by astronomical observation. Places north or south of a given place have the same local time; places east or west differ. Suppose, for instance, two clocks, one at Greenwich and one at Bristol, are set right by astronomical observation. Could one then be transported to the place of the other, the Greenwich clock would be found to be about 10 minutes fast of the Bristol clock. Or the sun arrives at the Greenwich meridian 10 minutes before it arrives at the Bristol meridian. It will be thus understood how, when railways began to grow up, it became necessary to employ, for safe regulation of the traffic, one uniform time. Greenwich time, now long known as "railway time," came to be adopted, and its use in the country is now universal. In course of time the want of some accurate and convenient standard of reference seems to have been felt, not alone for the service of railways, but also for that of the accompanying telegraphs, which so rapidly sprung up. The Astronomer Royal also, viewing the gradual rise of the telegraph system, and

especially the laying of the first submarine cable, became early desirous of placing the Royal Observatory in communication therewith, to be prepared to meet the scientific demands likely then to arise, as well as to be ready to supply any possible public demand for Greenwich time. How at last the timesignal system came to be proposed the Astronomer Royal himself is scarcely prepared to say: it was, as he expresses it, "partly in conversation, partly in other ways; but to Mr. C. V. Walker, Mr. Edwin Clark, Mr. Latimer Clark, and afterwards Mr. C. F. Varley, is the existence of the system due." After some correspondence, the Astronomer Royal, in the year 1852, obtained permission from the South-Eastern Railway Company (on the representation of Mr. C. V. Walker, their telegraphic engineer) to erect wires on their railway, for the purpose of obtaining communication with London; and special apparatus being provided, the system of signalling time from the Observatory was commenced. We shall not, however, further follow the subject historically, but proceed to describe the system as it now exists.

As respects, now, the time-distributing apparatus, the clock specially erected at Greenwich claims first attention. This clock, the normal mean-time clock of the Observatory (erected in the year 1852), is kept adjusted as nearly as possible to Greenwich mean time. It is maintained in action (on Shepherd's plan) by galvanic power alone. When its pendulum (a seconds pendulum) swings to the right, a galvanic circuit is closed, which causes an electro-magnet to raise a small weight. This being discharged on the pendulum in its swing to the left gives it a small impulse, which, repeated at each swing to the left, suffices to maintain it in action. Other galvanic circuits, closed, one as the pendulum swings to the right, another as it swings to the left, allow galvanic currents, alternately positive and negative, to pass to a pair of electro-magnets placed above it. These currents cause the electro-magnets to attract and repel alternately certain bar-magnets, giving thereby a reciprocating motion to the axis which carries them. An anchor on the axis gives forward motion to a wheel carrying the seconds hand, from which, by a simple train of wheels, motion is communicated to the minute and hour hands. So far as concerns the normal clock proper. But if the wire which passes from the pendulum to work the hands is afterwards led (before being returned to the battery) to other electromagnets in different parts of the building, each pair similarly working hands on a dial, the hands on all will advance together, their forward movement depending entirely on the galvanic current let off at each second by the one pendulum, which consequently governs the whole system. There are within the