The sky is divided into 24 segments or jiéqì (节气) based on the seasons of the year. The earliest calendars assumed that the motion of the sun was constant and divided the year into 24 segments with equal numbers of days. This method is called píngqì (平气). Because the motion of the sun is not consistent, this was found to be inaccurate. The calendar then changed to a method where the ecliptic (the path of the sun as seen from earth) was divided into 24 equal parts of 15 degrees. This method is called dìngqì (定气). Calendars from the Warring States Period through the Ming Dynasty used the pingqi method in their designs and only changed to the more accurate dingqi method during the Qing Dynasty.

The average lunar month is a little more than 29 and 1/2 days. It is the half day that is a problem. The solution was to have both short (29 days) and long months (30 days). The procedure works for a short time but eventually gets out of step with the new moon because the lunar month is slightly longer than 29.5 days. To correct that, periodically a double long month would be added to the calendar to bring the beginning of the month back to the new moon. The leap concept was also used to bring the calendar back to agreement with the year. An average of 29.5 days in 12 months only gives you 354 days in a year. Every few years, an extra month called an intercalary month would be added to make up the days needed to keep the solstices and equinoxes consistent with the seasons.

During the Eastern Han Dynasty (25-220 AD), Jiǎ Kuí 贾逵 observed that the motion of the moon was actually inconsistent, but he continued to use the pingqi method of calculation in his Sì Fèn calendar (四分历). In 223 AD Kàn Zé (阚泽) used the dingqi method to predict eclipses but still based his Qian Xiang calendar on the pingqi method. Subsequently, several calendars were proposed (from the Yuan Jia calendar in 445 AD through the Wu Yin calendar in 619 AD) that incorporated the dingqi method. In each case the imperial court refused the innovation because the procedure would include more than two long months in a sequence.

It is amazing that a culture that regularly required its astrologers to rewrite the calendar would hold on so tightly to the simple rule of the pingqi calendars, "Never more than two long months in a row," but they did. The relationship between our view of stability and the mythology we create around time is deep seated. We happily deal with uneven months and leap years, but if it were seriously proposed that we straighten the whole thing out, there would be riots in the streets as there were when the Gregorian calendar was introduced. It is not logical that we cling to something that is just a concept, but we do. As a consequence, calendars could be "reset", but they had to use the same rules.

Although future astronomers continued to use the dingqi calculation to develop their calendars and to make observations after the Qian Xian calendar, it wasn't until the Qing Dynasty that it was actually built into the calendar. All of the earlier calendars were based on the pingqi method.

Below is a table of the 24 seasonal segments shown in the picture above. They are based on the idea of predicting the seasonal variation using the moon's motion to measure the time it will take for the earth to return to a certain position (the solstices and equinoxes) in relation to the sun. Thus, the Chinese calendar is a lunisolar model.

Below that is a partial list of some of the calendars in Chinese history. A few will be discussed in greater depth on subsequent pages.