There is lively debate on whether Asian plate material was involved in southward flow of mid-lower crust in a ductile channel beneath southern Tibet. One argument against such involvement is the apparent absence of material derived from Asian lithosphere within the High Himalayan Series (Indian plate) that could represent the putative channel. A north-south–trending mid-Miocene dike swarm that intrudes the Tethyan sedimentary cover of the Sakya gneiss dome (Indian plate) yields new Sr-Nd isotopic data (87Sr/86Sr = 0.7071–0.7079; ϵNd −4 to −6) indicating that these melts share the same source as Miocene dacitic dikes from north of the Indus-Tsangpo suture. Moreover, dikes on both sides of this suture represent crustal melts derived largely from mid-lower crust of the Asian plate, exposed today as the Nyainqentanglha gneisses that underlie the Gangdese batholith. We infer that melting of the Asian lithosphere extended south of the surface trace of the suture, requiring southward propagation of anatectic Asian middle crustal material during the Miocene. The emplacement ages of the southern dike swarm (12–9 Ma) thus delimit the timing of active southward ductile flow of Asian material.