We present paleomagnetic results obtained from samples collected in the summer of 1989 along a 1000 km traverse between Yecheng (Xinjiang) and Shiquanhe (Tibet) in the western Qinghai-Xizang Plateau (along ~80¿ E longitude). During this field trip, 480 paleomagnetic cores in Jurassic to Tertiary limestones and sandstones were drilled at 49 sites; 400 were measured. Isothermal remanent magnetization and analysis indicates the existence of high coercivity minerals, and thermal, rather than alternating field demagnetization, was found to be more efficient in isolating paleomagnetic directions. The characteristic magnetization has rather low unblocking temperatures (250--560 ¿C) in most limestone samples and higher unblocking temperatures (400 ¿C--680 ¿C) in the sandstone samples. About half of the measured samples show the present geomagnetic field or aberrant directions. The remainder display stable multicomponent magnetizations. The stable characteristic magnetizations indicate (1) a probable Jurassic remagnetization; (2) a Cretaceous (120--80 Ma) magnetization which passes a fold test, yielding a mean pole at 66.2¿N, 245.0¿E (A95=5.1¿, N=14 sites); and (3) a large scatter of Tertiary directions, which may be due to tectonic problems and/or remagnetization in the sampling zone. The Cretaceous paleodirections from Aksaichin and Longmuco (11 sites on the Qiangtang (North Tibet) block> are very similar to those from Shiquanhe (three sites on the Lhasa (South Tibet) block>, separated by 300 km.

This suggests that the North and South Tibet blocks formed a single unit at least as early as the Cretaceous. Comparing our results with previous paleomagnetic work on the Tibetian plateau, it is found that the Paleolatitude of 10.6¿¿5.1¿ obtained from this study for a reference at Domar (33.75¿N, 80.4¿E) is close to that (9.6¿¿2.8¿) obtained by Achache et al. (1984) and Lin and Watts (1988) from the Takena formation (110--100 Ma) in the Lhasa region (30¿N, 91¿E). In addition, it appears that the southern margin of the Tibetan blocks (i.e., Eurasia) occupied an E-W trending position at about 7¿¿6¿N latitude. Paleodeclination differences between the western and eastern Tibetan blocks suggests internal deformations of the plateau at different scales (from 1 to 1000 km). According to differences in paleomagnetic declinations, the arcuate shape of the blocks would have been partly acquired after the collision. ¿ American Geophysical Union 1993