The Double Dawn in the Time of King Yi

Recently the powerful computers at the Jet Propulsion Laboratory in Pasadena, California, were turned to an unusual application: putting events in ancient China in terms of the contemporary calendar.

The long journey by which questions from antiquity reached high-tech machines better known for directing space flights began with observations made and recorded more than 3,000 years ago. The observers could not be called Chinese; then China as we know it did not exist. The scribes who recorded what they saw in the great Yellow River valley would have considered themselves subjects of one or another of three near-legendary dynasties, Zhou, Shang, or Xia.

In fact, the founding of China led to the loss of almost all the bamboo strips on which the deeds and observations of those predecessor dynasties had been recorded. Emperor Shi Huang Di (221-209 B.C.), first ruler of all China, ordered the archives' destruction. His eager minions overlooked only a few chronicles, including some interred in a royal tomb. They were discovered in the reign of a more scholarly emperor centuries later, and have been preserved ever since.

These Bamboo Annals are a study subject of Hung-hsiang Chou, a historian at the University of California at Los Angeles. According to Archaeology magazine, they have also entered the work of CalTech astronomer Kevin Pang--and the JPL computers.

Translated from the Bamboo Annals, the key passage for this team was: "During the first year of the reign of King Yi, in the first month of spring, the sun rose twice at Zheng." Most probably, this described a total eclipse of the sun occurring just at sunrise. The location of Zheng city is known, so the problem became one of ascertaining when a total eclipse at dawn could have been seen there.

Since an eclipse depends on the positions of the earth and the moon relative to one another and to the sun, for a computer to calculate when it might have been seen at a known location may seem easy--except that those relative positions change over time. The most significant change is that Earth's rotation is slowing, but its rate of slowing is not constant. It can be affected in complex ways by a host of factors, including things like shifting sea levels and melting ice sheets.

Even though the slowing of planetary spin is tiny, a matter of a full rotation taking thousandths of a second longer when compared across centuries, it's enough to put a glitch in the calculated place and time of ancient eclipses. Pang and Chou set about matching the Chinese historical records of eclipses at sunrise and sunset with computer simulations of movements of the earth and moon. Painstakingly, they corrected the computer when they could rely on the written record.

With equal care, they searched the historical record to refine the time span during which selected events took place. For example, the records tell that King Yi ruled "two generations" before the first accurately known date in Chinese history, 841 B.C. From inscriptions on bronzes and bones, they deduced that a generation was about 30 years; King Yi began his reign sometime near 901 B.C.

With that approximate date in the computers, they found only one candidate for the dawn eclipse at Zheng. Factoring in their calculated difference in the rotation rate, they made the day 42 thousandths of a second shorter than it now is--and came up with an eclipse at sunrise in Zheng on April 21, 899 B.C.

Without that imperceptibly shorter day, incidentally, the computer assured them that the eclipse would have occurred nearly six hours later, and hundreds of miles to the west.

The team has continued working on other reports, finding exact dates for even earlier events (the oldest, an unusual alignment of the planets, provides the beginning of the Xia dynasty: February 23, 1953 B.C.). From the work, astronomer Pang has gained a nice record of changing day length; historian Chou has gained some remarkable precision in setting dates for important happenings in long vanished kingdoms. Even the computers gained something--a more accurate program for calculating ancient eclipses.