To better understand the India‐Asia collisional process and intracontinental deformation within Asia, a paleomagnetic study has been conducted on the Qushenla Formation lava flows dated at ~132–120 Ma from the Yanhu area in the western Lhasa terrane. Stepwise thermal demagnetization isolates stable characteristic remanent magnetizations, which include dual polarity and pass fold tests at a 99% confidence level, indicating primary magnetizations. The tilt‐corrected site‐mean direction for 51 sites is D = 28.2°, I = 34.5°, and k = 74.3° with α₉₅ = 2.3°, corresponding to a paleopole at 61.4°N, 192.9°E (A₉₅ = 2.1°). Our new paleomagnetic data, combined with previous Cretaceous volcanic paleomagnetic data from the Lhasa terrane, show that the precollisional southern margin of Asia was at ~16.8°N with a relatively E‐W alignment and that the Lhasa terrane did not experience significantly discrepant north‐south movement although local vertical axis rotations did take place after the Cretaceous. Comparison with the apparent polar wander paths (APWPs) of India and the Cretaceous‐Paleocene paleopoles of the Tethyan Himalaya shows that the India‐Asia collision was at ~54.3 Ma and the Greater India during the Paleocene had a large northern extent of ~2000 km (~18.1°) beyond its present northern margin. Comparison with the Cretaceous Eurasian APWP indicates that a latitudinal convergence of ~1000 km has taken place between the Lhasa terrane and Eurasia since the India‐Asia collision. The amount of north‐south shortening deduced from Cretaceous paleomagnetic data is consistent with that accommodated by the Cenozoic fold and thrust belts between the Lhasa terrane and the Hexi corridor.

The timing of the north‐south collision between two terranes can be determined by the overlap of their paleolatitudes or the change of their convergence rate. For example, the overlapping paleolatitudes of the Lhasa terrane and Tethyan Himalaya and the dramatic decrease in the velocity of the Indian plate are usually ascribed to the India‐Asia collision at ~55 Ma. However, little is known about the paleolatitudinal evolution and velocity change of the Lhasa terrane resulting from the Lhasa‐Qiangtang collision during the Jurassic‐Cretaceous period. To better constrain the velocity change of the Lhasa terrane during this period, to constrain when and where the Lhasa‐Qiangtang collision occurred, and to assess the distribution of land and sea in the Tethyan realm, we provide a high‐quality Cretaceous paleomagnetic pole obtained from the limestone from the western part of the Lhasa terrane, which yields a paleolatitude of ~16.8° ± 1.9°N for the sampling area (32.2°N, 80.8°E) during the time interval of ~113–72 Ma. We compile existing paleomagnetic results from the Lhasa terrane, Qiangtang terrane, Tethyan Himalaya, and India and reveal that the Lhasa‐Qiangtang collision most likely occurred at or near the J/K boundary at ~19°N for the reference point at (32°N, 88°E) and that the Neo‐Tethys reached its maximum width (≥~7450 km) during this period.

The paleogeographic relationship between South China and Gondwana is critical for understanding the dispersion of Gondwana, accretion of Asia, and evolution of the Paleo‐Tethys. However, the lack of robust Devonian paleomagnetic data prevents a confirmative reconstruction of South China's connection to Gondwana and its subsequent separation during the Paleozoic. Here we report a new paleopole (33.6°N, 236.4°E; A₉₅ = 3°) from the Givetian red beds (~385 Ma) in central South China. Fitting apparent polar wander paths between South China and Gondwana suggests that South China was connected to East Gondwana from the earliest Cambrian to Early Devonian, with its position closed to NW Australia. Thereafter, South China separated from Gondwana during ~400–385 Ma, as evidenced by their decoupled apparent polar wander paths. The paleomagnetic data suggest that the Paleo‐Tethys Ocean between South China and East Gondwana had been up to ~1,600 km latitudinally wide by ~360 Ma.

To better constrain the origin and drift history of the North Qiangtang terrane (NQT), we report a well‐dated paleomagnetic pole from the Late Permian volcanics of the NQT that appears to average out secular variation. Our new results yield a paleolatitude of −7.6 ± 5.6°N at ~259 Ma for our sampling area, which confirms the NQT drifted northward during the Permian and Triassic periods. The equatorial paleolatitude of the NQT is similar to that of the coeval South China block, demonstrating that they were in close proximity. Combined with palaeontological and magmatic evidence, paleomagnetic constraints on the drift of the NQT in the Permian indicate that the NQT moved northward together with the South China block at this time. The paleolatitude evolution of the NQT implies that the NQT rifted from the northern margin of the Gondwana in the Devonian, which is earlier than the departure time of the South Qiangtang terrane.

The breakup of eastern Gondwana is among the hottest topics in the Earth sciences because of its effect on global climate during the Jurassic‐Cretaceous, its influence on the evolution of life, and its importance to paleogeographic reconstruction. To better constrain the Jurassic and Cretaceous paleogeographic position of the Tethyan Himalaya and the breakup of eastern Gondwana, a combined paleomagnetic and geochronological study was performed on the Zhela and Weimei Formations lava flows, dated at ~138–135 Ma, in the Luozha area of the eastern Tethyan Himalaya. Both positive fold and reversal tests together with a maximum grouping at 100% unfolding indicate that the characteristic remanent magnetization directions are primary magnetizations acquired before folding. The tilt‐corrected directions yielded a paleopole at 0.9°N, 293.4°E with A₉₅ = 7.0° and a corresponding paleolatitude of 53.5°S ± 7.0°S for the Luozha sampling area (28.9°N, 91.3°E), validating that the original erupted position of the Zhela and Weimei Formations lava flows was located in the center of the Kerguelen mantle plume. Our new results, together with the published paleomagnetic, geochronological, and geochemical results, demonstrate that the Comei‐Bunbury large igneous province originated from the Kerguelen mantle plume. The temporal and spatial relationships between the Comei‐Bunbury large igneous province and the Kerguelen mantle plume indicate that eastern Gondwana initially rifted at ~147 Ma and that the Indian Plate fully separated from the Australian‐Antarctic Plate before ~124 Ma.

To better constrain the Lhasa-Qiangtang collision, a combined palaeomagnetic and geochronological study of the far western Lhasa terrane was conducted on the Duoai Formation lava flows (~113–116 Ma), as well as on the Early Cretaceous Jiega Formation limestone. Following detailed rock magnetic, petrographical, and palaeomagnetic experiments, characteristic remanent magnetisation directions were successfully isolated from most samples using principal component analysis. The tilt-corrected direction groups yielded a palaeopole at 69.1°N, 319.8°E with A₉₅ = 4.8° (N = 19). A primary origin for the magnetisation is consistent with positive fold tests. Our results from the Early Cretaceous units, combined with published palaeomagnetic data obtained from Cretaceous strata from the Lhasa and western Qiangtang terranes, show that these two terranes had already collided by the Early Cretaceous, the Lhasa terrane had a relatively east-west alignment, and it remained at a relatively stable palaeolatitude during the entire Cretaceous. Comparing the Cretaceous palaeolatitude calculated for the western Lhasa terrane with those from Eurasia and Mongolia suggests a latitudinal convergence of ~1400 ± 290 km and ~1800 ± 300 km, respectively, since the Early Cretaceous.