Illustrated Explanation of the Record of Trades: Volume I

...zhi. The part located outside the spokes, near the linchpin The xia (轄) is the pin at the end of the axle that keeps the wheel from falling off. of the hub, is named the small opening. The source text reads "large opening" here, which appears to be a transcription error as it contradicts the logic of the following sentences; contextually, it must refer to the "zhi" or small opening. The circumference of the large opening is large, and the circumference of the small opening is small. Because the part of the axle inside the spokes bears the heavy weight of the carriage, it cannot be tapered sha: to shave down or reduce the diameter of the wood to fit a space or reduce weight. If that opening were made large, the hub would be weakened. Conversely, the part of the axle outside the spokes bears a lighter load, and thus it can be tapered. Making that opening small ensures the hub remains strong. Furthermore, tapering the axle serves to limit the hub’s movement, preventing it from shifting too far inward.

The Record The Kao Gong Ji (Record of Trades). states: "Divide the length of the hub into five parts: remove one part to determine the xian; remove three parts to determine the zhi." The xian is the metal bushing of the large opening; the zhi is the metal bushing of the small opening. See the "Explanation of Metal" for details. If we follow this, the circumference of the xian would be two feet, three inches, three minutes, and seven thousandths Approximately 2.337 chi. In the context of a hub's length, this would result in a hole so wide the surrounding wood would be paper-thin.. At such a size, the hub would be far too thin and the opening far too large—this is logically impossible.

Therefore, Master Zheng Kangcheng Zheng Xuan (127–200 AD), the preeminent Han dynasty scholar whose commentaries are foundational to understanding these ancient texts. stated: "The 'large opening' is excessively large; the text must be in error." In truth, the large opening should be "five parts of the hub length, remove two." After repeatedly examining this theory, it is indeed the most reliable basis; for if one removes two parts from five, the circumference is one foot, seven inches, and six minutes. With this circumference, the hub is not excessively large, the wood remains thick, and it will not easily break.

The error in the text was changing "remove two" into "remove one."