LAUSR

The Matuyama‐Gauss (M‐G) magnetic polarity reversal is regarded as a fundamental time marker in the stratigraphic division of the Quaternary‐Neogene. However, previous paleomagnetic studies have shown that the M‐G is mainly recorded in the Chinese loess unit L33—a glacial stage (corresponding to marine isotope stage 104, i.e., MIS 104)—which is asynchronous with the timing recorded in marine sediments. Here, we solve this long‐standing debate by exploiting a method to extract reproducible records of paleomagnetic field intensity from Xifeng and Lantian loess profiles with meteoric 10Be. The results showed that for both loess profiles, the 10Be‐derived M‐G boundary is located in paleosol S32 ca. 2,589 ± 3 ka, which corresponds to MIS 103. This is synchronous with that seen in marine sediments, though it is, on average, ∼19 ka younger than the boundary inferred from paleomagnetic measurements from the two profiles, which demonstrates that magnetic overprinting has occurred.