In the late nineteenth century it was found that the rotation of the Earth (i.e. the length of the day) was both irregular on short time scales, and was slowing down on longer time scales. In fact, observing the position of the Moon, Sun and planets and comparing this with their ephemerides was a better way to determine the time.
Using the ephemerides based on the theory of the apparent motion of the Sun by Simon Newcomb (1898), the SI second was defined in 1960 as:
Caesium atomic clocks became operational in 1955, and quickly made it evident that the rotation of the earth fluctuated randomly. This confirmed the utter unsuitability of the mean solar second of Universal Time as a measure of time interval. After three years of comparisons with lunar observations it was determined that the ephemeris second corresponded to 9,192,631,770 cycles of the caesium resonance. Between 1960 and 1984 the length of the SI second was defined to be equal to the ephemeris second.
The difference between ET and UT is called ΔT; it changes irregularly, but the long-term trend is parabolic, decreasing from ancient times until the nineteenth century, and has been increasing at about 0.7 seconds per year since (see leap seconds). International Atomic Time (TAI) was set equal to UT2 at 1 January 1958 0:00:00 . At that time, ΔT was already about 32.18 seconds. The difference between Terrestrial Time (TT) (the successor to ephemeris time) and atomic time was later defined as follows:
This difference may be assumed constant—the rates of TT and TAI are designed to be identical.