Although an ancient hematite artifact from the Olmec era in Mexico dating to roughly 1000 BC indicates the possible use of the lodestone compass long before it was described in China, the Olmecs did not have iron which the Chinese would discover could be magnetised by contact with lodestone.  Descriptions of lodestone attracting iron were made in the Guanzi , Master Lu's Spring and Autumn Annals and Huainanzi .    The Chinese by the Han Dynasty (202 BC – 220 AD) began using north-south oriented lodestone ladle-and-bowl shaped compasses for divination and geomancy and not yet for navigation .    The Lunheng , written by Wang Chong (27 – c. 100 AD) stated in chapter 52: "This instrument resembles a spoon and when it is placed on a plate on the ground, the handle points to the south".   There are, however, another two references under chapter 47 of the same text to the attractive power of a magnet according to Needham (1986),  but Li Shu-hua (1954) considers it to be lodestone, and states that there is no explicit mention of a magnet in Lunheng .  Shen Kuo (1031–1095) of the Song Dynasty (960–1279) was the first to accurately describe both magnetic declination (in discerning true north ) and the magnetic needle compass in his Dream Pool Essays of 1088, while the author Zhu Yu (fl. 12th century) was the first to mention use of the compass specifically for navigation at sea in his book published in 1119.        Even before this, however, the Wujing Zongyao military manuscript compiled by 1044 described a thermoremanence compass of heated iron or steel shaped as a fish and placed in a bowl of water which produced a weak magnetic force via remanence and induction; the Wujing Zongyao recorded that it was used as a pathfinder along with the mechanical south-pointing chariot .