A 500-km-long west-east wide-angle seismic profile from Selin Tso to Yaanduo in the northernmost Lhasa block of Tibet, acquired by the Sino-French joint seismic program in 1982, has been reinterpreted. We model the P and S wave velocity structure of the whole crust, while recognizing that in many places, uncertainties are large. A surprising but robust conclusion, consistent with previous interpretations of both this data set and other newer data from Tibet, is that Moho depth is about 60--65 km at 90¿E but 75--80 km depth at 92.5¿E. Our detailed interpretation uses multiple wide-angle P and S wave reflections from the crystalline basement and the Moho; but no Moho refractions are recognized. Along most of the profile, the crust may be crudely divided into an upper crust (~5--30 km depth, 5.0 < Vp < 6.4 km/s), a middle crust (~33--45 km depth, 6.5 < Vp < 6.8 km/s) and a lower crust (depths below ~48 km, 7.0 < Vp < 7.4 km/s). The west-to-east increase in crustal thickness is accomplished by a ~50% thickening of the middle and lower crust. Larger vertical velocity gradients separate these three layers and bound them above (surficial and sedimentary rocks) and below (Moho transition zone). The most notable low-velocity zone in the crust lies at the base of the upper crustal layer. S wave velocity structure is less well constrained but parallels the P wave structure except that Vp/Vs ratios may decrease from west to east in the lower crust. Our data suggest considerable variation in structure along-strike of the Tibetan Plateau and show that interpretations of Tibet as a purely two-dimensional orogen are overly simplistic. The west-east increase in crustal thickness may occur across the Karakorum-Jiali fault system and be an indicator of lateral tectonic escape of the Qiangtang terrane.