ASTRONOMY IN A WORLD AT WAR—DOUGLAS 157 
II 
Turning to the subject of time measurement, it is worthy of note 
that during these war years an accuracy never before dreamed of 
has been attained. It was in April 1938 that Essen described before 
the Royal Astronomical Society the researches at the National Physi- 
cal Laboratory which had resulted in the new quartz clock, of which 
so much was hoped. This clock makes use of the properties of the 
crystal oscillator, one of the most reliable and perfect mechanical 
systems known to man. Essen describes quartz clocks briefly as “con- 
sisting of phonic motors controlled via frequency dividers by vibrating 
quartz crystals.” In a paper presented to the Royal Astronomical 
Society last June, Greaves and Symms record the intercomparisons 
of three Greenwich free pendulum Shortt clocks, two National Physi- 
cal Laboratory quartz clocks, and three quartz clocks at the Post 
Office Radio Branch Laboratories. 
They analyze clock errors into three classes: (a) erratic varia- 
tions in phase, (b) erratic variations in rate, (c) a combination of 
phase and rate variations, producing a cumulative effect. They show 
that two Shortt clocks and two quartz clocks may indicate approxi- 
mately the same mean absolute second differences of relative clock 
error, but the distribution of errors between the three classes is differ- 
ent—the quartz clocks show very little error of (b) and (c) relative to 
Shortt clocks, and errors of class (a) do not affect the long-period 
performance of a clock. 
The famous Shortt clocks are now known to be incapable of giving 
the precision demanded, but the Astronomer Royal hastened to pay 
them a deserved tribute: 
Twenty years ago we had several papers dealing with the performance of 
the Shortt clocks, then looked upon with great expectations. In this clock was 
achieved in a simple and beautiful manner what horologists had been striving 
after for years, namely, a pendulum designed solely for the purpose of beating 
time whilst being called upon to perform no mechanical work. But if the subse- 
quent performance of this type of clock did not fully come up to our high expec- 
tations, the Shortt Free Pendulum has one thing to its everlasting credit—it 
forced the astronomers to adopt the use of Mean Sidereal Time where formerly 
True Sidereal Time had been adequate. During the intervening 20 years since 
this type of clock was installed in many observatories, new requirements have 
sprung up. In the past the main purpose of a time service was to provide absolute 
time with an accuracy sufficient for navigational and surveying requirements. 
But the new use of frequency standards has raised a demand for 24-hour intervals 
correct to the very high accuracy of a millisecond. 
It will be seen then that as absolute standards at Greenwich, Shortt clocks 
have become obsolete. Our long-range predictions are now based entirely on 
quartz clocks, free pendulum clocks being used only for extrapolation over an 
interval of 24 hours. 
