• Home
• Brief History
• The Greenwich Meridian
•  
• Greenwich (1675–1958)
• Herstmonceux (1948–1990)
• Cambridge (1990–1998)
• Outstations (1822–1971)…
  • – Chingford (1822–1924)
  • – Deal (1864–1927)
  • – Abinger (1923–1957)
  • – Bristol & Bradford on Avon    (1939–1948)
  • – Bath (1939–1949)
  • – Hartland (1955–1967)
  • – Cape of Good Hope    (1959–1971)
•  
• Administration…
  • – Funding
  • – Governance
  • – Inventories
  • – Pay
  • – Regulations
  • – Royal Warrants
•  
• Contemporary Accounts
• People
• Publications
• Science
• Technology
• Telescopes
• Chronometers
• Clocks & Time
• Board of Longitude
•  
• Libraries & Archives
• Visit
•  
• Search

Clock errors and the regulation, rating and comparison of the Greenwich clocks

Page under construction

 

The aim of this page is to bring together sections from different parts of the website to help further understanding on the processes used during different periods at the Observatory.

 

From the page on the Transit Clock Graham 3: Comparing the observatory clocks

An inspection of the manuscript observations of Bradley indicates that from its arrival as the new transit clock in 1750, Graham 3 was normally wound along with the other two month going clocks by Graham on the first day of the month. The manuscript observations also indicate that at the time of winding, the difference between Graham 3 and one of both of the other Graham clocks was also noted. By 1755, Bradley was using a shorthand code to record this. None of this information about winding or comparing the clocks was transcribed into the first volume of Bradley's published observations, which covered the Transit Observations made between 1750 and the end of 1675, but it was patchily transcribed into the second. In the preface to the second volume (which covers the Transit Observations made between 1756 and 1765 by both Bradley and Bliss), the code was explained as follows:

'That the reader may fully understand some remarks, it may be necessary to inform him, that N. stands for the new Clock in the Transit Room [Graham 3], Q. for the Clock in the Quadrant Room [Graham 2], and G. for the Clock in the Great Room [Graham 1]. When the times by any two of these are compared, the comparison is expressed thus, N. Q. 4", meaning that the Transit Clock was 4" before the Quadrant Clock.'

An example from May 1763 can be seen here. No explanation was given however as to how the clocks were compared.

Maskelyne (who arrived as Astronomer Royal in 1765) continued to normally wind the clocks on the first day of the month. In the preface to his first volume of Observations:he explained his method of comparing the transit clock with that in the Great Room (the Octagon Room):

'In the year 1766, two assistant clocks, with a loud beat, were provided, and set up; one in the transit room and the other in the great room; principally designed to be used in windy weather, when the beats of the other clocks could not be well heard at a distance: in which case the assistant clock was first set to beat exactly with the principal or astronomical clock. Moreover, by the help of the bell of the assistant clock, which rings exactly every minute, when the second-hand comes to 60, and which may be heard from the great room to the transit room, when a window in each is left open, the astronomical clocks standing in these rooms are readily compared together, and their difference noted, in order to reduce the observations made in the great room to the time of the transit clock.'

In later years, other methods were used to compare the sidereal clocks notably, with a mean solar time chronometer and the method of beats - a practice that partially ended with the introduction of the chronograph in 1854.

Comparisons were also made between the transit clock and others set to mean solar time using a special device by Hardy which was described by Sabine in the write up of his pendulum experiments as follows:

'The comparison of the clock [Graham 3] with the Greenwich transit clock [Hardy] was effected by means of a machine constructed by HARDY for the purpose, it being capable of indicating 0".05 in time; and from the mean of 5 comparisons which was always employed, it is hoped the comparisons never err 0.03 from the truth these comparisons were made at or near the time the observations were making for the rate of the transit clock, on the accurate determination of which must rest the accuracy of the rate of the clock used in the experiment.'

 

From the page on the Transit Clock Hardy

Information on how the transit clock was rated appears just once in the published observations of Maskelyne and not at all in those of Pond. Maskelyne's observations were published in four volumes and between them cover the years 1765–1810. Only the first (published in 1776) includes a preface. On page iii Maskelyne gives the following information:

'The Observations made with the transit instrument are the most numerous of any; besides the transits of the planets, the transits of several fixed stars being observed every day that the weather permits, to be compared with those of the planets, in order to settle their right ascensions, and serving also to determine the clock's rate of going from day to day. The stars commonly made use of for these purposes are those contained in the Tenth of the annexed Tables: which, as they lie near the Equator, move with greater velocity over the Meridian than stars of greater declinations, and they are moreover sufficiently bright to be easily seen in the day-time, when the air is clear, and not liable to be obscured by a small haziness of the air, or the thinner sort of clouds, either by day or night, as smaller stars would be. …

In making computations from the Observations, I deduce the daily rate of the going of the transit clock, from a mean of the differences of transits of the same star at the same wire on different days; in doing of which I make equal use of all the wires. …' Read more.

A somewhat clearer explanation is given by Robert Woodhouse, the Lucasian Professor of Astronomy at Cambridge. In Volume 1 of his A treatise on Astronomy, Theoretical and Practical (1821) he describes the process as follows:

'The practical method of determining the clock's daily rate, that is, its gain or loss during two successive transits of a star, is to subtract the mean meridional passages of certain stars on one day (as shewn by the clock) from the passages of the same stars on the next, or on some following day. The sum of the differences divided by the number of days intervening between the observations, and by the number of stars, is the clock's mean daily rate; to which quotient, or result, should the clock gain, the sign + is affixed; should it lose, the sign -.'

He then goes on to give a practical example based on Maskelyne's observations made with Graham 3 on 23 & 25 January 1798 and the observations made with the same clock by Pond on 6, 7 & 8 August 1816. Although it had no impact on the methodology, it should be pointed out that not only did Woodhouse made transcription errors when copying the data (for example, the Pond observations were actually made on 5, 7 & 8 August), but he also failed to notice that Pond or his Assistants had made an error in one of the calculations of the clock's rate – an error which he then repeated. When using a transit telescope and clock to measure right ascensions, there are further complications due to instrumental errors. As well as discussing these in his Treatise, Woodhouse also discussed them in his 1825 paper: Some Account of the Transit Instrument Made by Mr. Dollond, and Lately Put up at the Cambridge Observatory.

Interestingly, the rate of the Greenwich transit clock as computed from individual 'clock' stars only began to be included in Maskelyne's published observations part way though volume 3 at the start of 1793. To aid the reader, the number of days that had elapsed since the star had been last observed, was also recorded in the column headed 'N^o. of days. The practice of including the daily rate was continued by Pond, who in 1829, also began publishing on a single sheet, a table showing the error and rate of the clock at 0^h sidereal time at different points throughout the year. Two are reproduced in the section below. How these rates were computed from the figures given in the transit observations is not stated.

Meanwhile, back in 1828 Pond's Assistant Thomas Glanville Taylor had supplied information about the rate of Hardy to Edward Sabine who was carrying out experiments at the Observatory on invariable pendulums. Sabine's paper, which was published in Philosophical Transactions the following year, included the actual calculations Taylor made to determine the rate. Based on groups of stars of similar right ascensions, the composite figures differ significantly from the figures of the daily rates based on individual stars that were subsequenly published in Greenwich Observations.

In his Introduction to practical astronomy Pearson produced the adjacent table showing the rates of three different Hardy regulators that he had extracted (without acknowledgement) from Volume 39 and Volume 42 of Transactions of the Society Instituted at London for the Encouragement of Arts, Manufactures, and Commerce. Produced with the aim of showing Hardy's clocks in a good light, it needs to be interpreted with caution as the methodology used to rate the clocks differed and becauset he did not provide any equivalent data for his own Hardy. At the start of 1820, the Greenwich clock was off site being 'cleaned' by Hardy. It was not put back up until 1 February, having been absent since 20 December. Its pendulum was adjusted on 9 February and again on 19 April in order to bring its rate closer to zero. At this point in time, Pond was only making typically two or three transit observations a day with the Mural Circle and he did not tabulate a daily rate. Instead, every few days he recorded how fast or slow the clock was. The following entry shows the format used: 'At 6h 15', Clock too fast 0' 5".7' sidereal Time.' Pond did not record how the sidereal time was obtained, but it was presumably derived from Graham 3 which at that time was being used with the Transit Telescope. Whoever produced the table in Volume 39 of the Transactions appears to have taken these entries to compute a rate for the Greenwich Hardy. How the figures for the Sandhurst clock were computed cannot be determined. Those for the Wilna clock are given against observations of particular clock stars and vary from star to star as well as from day to day. Pearson was unable to give the temperature in which the clocks were operating apart for the clock at Wilna, which suggest it was not exposed to sub-zero ^oC temperatures (was it in some kind of protected space?). By comparison, the observations published by Pond show that in December 1820 the Greenwich Hardy was frequently at temperatures below zero.

 

Published records of the rate of the Transit Clock

Bradley Vol 1 Page 116 (Link)

Bradley Vol 2 p 308 &  p413

Maskelyne from 1793 onwards Vol 3

Pond, likewise