Reflections on the deep Moho of the Tethyan Himalayas have been obtain
ed from vertical seismics at the time predicted by wide-angle reflecti
on interpretations, supporting the previous inference of a low average
crustal velocity. A mostly felsic composition is then required throug
h the 70 km of the thickened crust. The Moho appears as a heterogeneou
s transition layer more than 10 km thick. The high velocity measured b
eneath the Moho from Pn refractions cannot be attributed any more to E
-W-oriented olivine since SKS observations do not confirm such an anis
otropy. Instead, eclogite just under the seismic Moho could account fo
r the observations. These three results from independent parts of the
seismic data set support a model of the Himalayan convergence in which
the Indian crust has undergone phase transformation in the granulite-
eclogite system upon burial. The intermediate and basic composition ro
cks of the middle and lower layers of an initial crust of average plat
form type can be transformed to such high velocities that they are los
t across the Himalayas into the level of the Moho or just under the Mo
ho defined by seismic velocity. Mostly upper-crustal-type material the
n remains to form the thickened crust above, consistent with the obser
ved low velocity. The variation in Moho depth across the Himalayas as
suggested from wide-angle seismics accounts for the first-order variat
ion in teleseismic transit times. Variations in the velocity-depth str
ucture of the mantle hence do not dominate, but a change in the orient
ation of SKS splitting anisotropy would suggest deformation within the
lithosphere to change across the High Himalayas and the Indus-Tsangpo
suture.