A.S. Collins writes: In their recent paper Khan et al. 1997 succinctly discussed geochemistry of the Kohistan island arc of northern Pakistan. They presented excellent new data from metavolcanic lithologies throughout the arc and convincingly argued that Pb and Sr isotopic compositions suggested formation above the DUPAL isotopic anomaly, presently some 2000 km south of the present position of Kohistan. However,Khan et al. 1997 interpreted trace element geochemical data to imply a southerly dipping subduction zone beneath Kohistan during its formation as an intra-oceanic island arc. Below, I present a critique of this interpretation which suggests that southerly dipping subduction in not supported by available data. I also include structural evidence from a recent field study of the Shyok suture which upholds an alternative tectonic interpretation of northerly directed subduction (Pudsey 1986; Petterson & Windley 1991; Robertson & Degnan 1991). This subduction zone consumed an ocean that separated Kohistan and Eurasia in during the Cretaceous.
(i) Interpretation of geochemical data. Fundamental to an interpretation of geochemical data in terms of lateral changes in tectonic environment, is showing that sampled rocks formed at the same time at sites separated horizontally from each other. The rocks in Kohistan are characterized by a sub-vertical foliation parallel to boundaries between the major rock units. Coupled with this is a progressive, and well-documented, increase in metamorphic grade southwards from chloritoid slates in the Hunza valley (developed in the Yasin formation that lies interbedded with the topmost Chalt Volcanics; Pudsey 1986) through greenschist-facies rocks (Jaglot group,