The inversion metamorphic zones, garnet zone to starolite, kyanite, and sillimanite zones in the High Himalayan crystalline rocks at the Nyalam profile have been reinvestigated. Except for the sillimanite zone, the other metamorphic zones are brought about by the solid state reactions under regional dynothermal metamorphism. The presence of sillimanite is not resulted from the prograde metamorphism of kyanite or starolite zone, i.e., sillimanitization has no relations with the Ky-St-Grt zonation. The so-called sillimanite zone here is not an independent metamorphic one and therefore cannot be included in the inversion metamorphism. In fact,Sillimanitization is closely related to the solution or melt after anatexis. Temporally the sequence is prograde metamorphism and zonation→anatexis→sillimanitization. The occurrence of sillimanite in the High Himalayan crystalline rocks marks the waning, not the beginning period of anatexis. Genesis of sillimanite is mainly associated with the decomposition of biotite and/or potassium feldspar and the subsequent migration of alkaline and alkaline earth metals in the deformation process. The step segregation is crucial in the formation of sillimanite in the migration of solution or melting through the fissure. The Al, Si components condensed at early stage formed sillimanite and quartz, and other components precipitation, successively. Contemporaneously with the coarsening of fibrolite to prismatic sillimanite, the myrmekite texture and albitic plagioclase rim occurred. Sillimanitization may be related with the doming and uplift of the anatexitic granitic gneisses.