Two kinds of melange extend meridionally parallel to the Jinsha River-Red River deep fracture and are distributed on its opposing sides. The region to the west of this deep fracture belongs to the Sankiang (i.e. Three-Rivers) eugeosynclinal fold belt, along which consolidation shifted from north toward south during the Noric stage. The Sungpan-Ganze eugeosynclinal fold belt extends over the region to the east of the deep fracture, and the eugeosyncline was developing until the end of Triassic. During the Triassic when two plates converged on which laid both eugeosynclines mentioned above, there occurred ophiolitie melange and olistostromic melange distributed on the under-thrusting (western) plate and on the overthrusting (eastern) plate respectively. It is considered that the so-called subduction zone here seems to be a deep fracture, namely, a plate suture line. The ophiolitic melange is characterized by the matrix of a suite of ophiolite containing many exotic blocks of various rocks, such as omphacite-eclogite, rodingite and fossiliferous limestone. However the matrix of the olistostromic melange dominated by argillaceous and arenaceous rocks containing fossils all younger in age than those in the exotic blocks. All the characteristics of the two kinds of melanges concerned herein may serve as a reference for elucidating their genesis. -It can be supposed that in the case of subduction of one plate beneath the other one the ophiolitic melange may be considered as a direct product under an intensive compressive stress acting in the upper part of the lithosphere. On the other hand, olistostromic melange can be regarded as the sedimentary unit formed under the horizontal stress in the water bodies nearby. After briefly reviewing previous works in this respect and preliminarily examining the studies on mechanism of melanges made by some geologists, the authors believe that the present concept as proposed here of the ophiolitic melange and the olistostromic melange could eventually get general approval, because it gives not only a reasonably genetic classification, but also a criterion easily grasped in field observations. It should also be noted that the results of research on melange reached by K. J. Hsu and A. Gansser are important contributions, but a certain amount of their revision still remains to be done.