The Matuyama-Brunhes (M/B) boundary, termination of Jaramillo normal subchron (J/M) as well as three young geomagnetic excursions has been revealed in the Weinan loess section <Zhu et al., 1994, 1998, 1999>. Recently, we carried out further paleomagnetic directional studies on the same section and recovered the onset of the Jaramillo normal subchron (M/J) and four geomagnetic excursions, two in the Brunhes and two in the late Matuyama. The fact that all seven excursions found in this section are in loess units and expected excursions in paleosols are not found suggests that short-lived geomagnetic signals could be only recorded in loess units, and may have been fully smoothed or wiped out in paleosol units by strong pedogenesis. Constrained by a new astronomical timescale (HLD00), ages of the geomagnetic excursions and reversal boundaries have been estimated. We tentatively correlated these excursions with previous results, indicating that a large lock-in model may not be correct for the whole loess profile, although some apparently big positive age discrepancies were observed for the main boundaries.