CHINESE CHRONOLOGY. 73 
The tenth term exhausts the stems, and then the first stem 
is joined to the eleventh branch, and the second to the 
twelfth, making chid-hsii and yi-hai, the eleventh and twelfth 
terms. The twelfth term exhausts the branches; but the 
third stem is then prefixed to the first branch, making the 
thirteenth term, which is ping-tsze. And so the process goes 
on till we reach sixty, the least common multiple of ten and 
twelve. No more different combinations can be made with 
the two series of characters. The cycle is completed and 
a new one begins. To speak of sexagenaries instead of 
centuries sounds strange to us; but I would make little 
account of that, if we could tell where the inventors got the 
idea of its component parts,—the ten stems and the twelve 
branches, and how they were led to the employment of the 
characters,—for the most part hardly more complex in form 
than our figures, by which these characters are denoted. 
The Cycle at first intended for Days, not Years. 
16. Thus far, however, I have been baffled in my endeay- 
ours to discover light on these points. It is more important 
to observe that authorities agree that the object of the cycle 
at first was to keep a record of days and not of years. In 
the Shi and some of the other Ching, we find many such 
applications of it to days, but not a single instance of its 
application to years. We have seen that the cyclical names 
annexed to the years in Sze-ma Kwang’s history were not 
carried back to the time of Yao till our eleventh century ; 
and much in the same way it has been proved that the 
cyclical dates for the years in the Bamboo Annals were not 
in the tablets when they were disentombed in A.D. 279, but 
are a subsequent addition. Of the astonishing accuracy, how- 
ever, with which the cyclical terms were employed in the 
record of days, we have an example in the notices of the 
solar eclipses, which are recorded, as I said, in the Ch’un-ch’it 
of Confucius. Of the thirty-six eclipses mentioned in that 
computation, extending over 242 years, thirty-two have been 
verified by calculation. The year is always right, and the 
day; but the month is often wrong; the error in the months 
being explained by the trregularity with which the process of 
intercalation,* according to the Chinese method, was con- 
ducted. The very error is a strong confirmation of the 
genuineness of the history ;—it is a fine illustration on a con- 
* See this proved by Dr. Chalmers in the prolegomena to The Chinese 
Classics, vol. v, pt. i, pp. 93—97. 
a 2 
