Throughmultiple field surveys in the Yellow River’s Hetao reach of Inner Mongolia in recent years, OSL and ESR dating, and comparative study with the other reaches of the Yellow River and other rivers of China, such as the Yangtze River, following discoveries and understanding have been concluded.  (1) The Hetao area not only includes the faulting basin which was filled with several kilometers thick Late Cenozoic fluviolacustrine facies sediments, but also contains the nine levels (T1～T9, 300m or more above river level) of the Yellow River terraces along the south piedmont of the Yinshan Mountains on the north side of the Yellow River, especially along the north margin of the Erdos Plateau. The occurrences and characteristics of these terraces are briefly described as follows. T1 terrace with a height of 2～12 m high above river level is an internal superimposed accumulation terrace, and composed of sand and gravels distributingd along the Yellow River and its tributaries. T2 to T4 terraces with heights of 12～45 m，28～80 m and 60～115m high above river level, respectively, decrease gradually in height from downstream to upstream. They are mainly bedrockseated terraces. Most of them are distributed in Hetao paleolake, with bedrocks to be lacustrine deposits. The Hetao paleolake occurs usually as lacustrine platforms of 1080~1100 m high above sea level. Typical examples are seen southeast of Tuoketuo County where lacustrine deposits are covered with the Malan loess, but at their bottom there is even a layer of indistinct paleosol S1. Some terraces of T2～T4 on the north margin of the Erdos Plateau take the Hipparion red clay or Paleogene rock or even other rock as their base, while some terraces of T2～T4 took on the south piedmont of the Yinshan Mountains take Paleogene rock or even other rock as their base, but without the Hipparion red clay. T5 to T8 terraces have height of 75～160m，130～200m，170～260m and 250～295m high above river level, respectively, with height differences to decrease gradually from downstream to upstream, too. They are composed of sand and gravel layers intercalated with one interbed or more of loess and paleosol and take the Hipparion red clay or Paleogene rock or even other rock as their base. Occasionally in some terraces of T5～T8 are fluvial gravel layers, which have been partly well calcium cemented occurred under terrace sediments or the Hipparion red clay, suggesting that they are sediments of older Yellow River. T9 terrace was found only at two sites of the Helaigou profile, which is 290~315 m high above river level and has a base of the Hipparion red clay containing as many as 22 layers of white caliche nodules.  (2) The OSL and ESR dating results of the samples collected from the above mentioned terraces show that these terraces developed throughout the Quaternary. T1 formed in the midHolocene; T2~T4 formed in the late, the middle, and the earlytomid of the Late Pleistocene, respectively; the Hetao paleolake was filled in the early Late Pleistocene; T5~T9 formed in the late, the middle, and the early of the Middle Pleistocene and the late and early of the Early Pleistocene, respectively. On the other hand, the quartz thermalactive ESR ages of the fluvial gravel layer and the well calcium cemented gravels under T5~T8 and under the Hipparion red clay are as old as Miocene or Pliocene. (3) From the distribution, the sedimentary facies and ages of T1~T9 and the occurrence of the loesspaleosol sequences and the Hipparion red clay, we think that since the Neogene, the Yellow River has exited not only in the Hetao strongly depressing fault basin，but also has flowed between the south piedmont of the Yinshan Mountains and the northern margin of the Erdos Plateau. At the same time the Yinshan Mountains and the Erdos Plateau have continually risen.The Yellow R. could feed into one end of the lake, and flow out from the other end constantly, and the river course also could be flow through from one or two sides of the lake. That is the riverlake coexistence theory.  (4) Through the survey of the upmidstream and some downstream reaches of the Yellow River, it can be found and considered that the Yellow River is an ancient long river through the upstream and downstream since the earlymid Miocene, although the preserved condition and consecution of the ancient Yellow R. heritages are weaker than those of the Yellow R. in the Quaternary due to the destruction by tectonization. (5) Considering the coexistence relation of riverlake and the similarity of theHetao fault basins to fault basins in other reaches of the Yellow River, such as the Sanmen basin, the Guide basin and the Gonghe basin, we think that at least since the early period of the Neogene, the Yellow River has been a large river without interruption, but not a connection of local rivers segment by segment through their individual source erosion. Fault basins or ancient lakes cannot prevent the Yellow River from running into sea, without the emptying of lake water.  (6) The Yellow River has been formed since the earlymid Miocene, it can cut the mountains and plateaus which are in the uplifted area, shaped many gorges and river terraces, and it also can filled the basins and plains which are in subsidence areas, accumulated thick late Cenozoic strata. The Yellow R. was flowing there all through, and the roughly parabolic longitudinal profiles also could be reserved at different stages. (7) Since theNeogene, the Yellow River not only flows through the Hetao fault basin, but also swings between the south piedmont of Yinshan Mountains and the northern Erdos Plateau which are located in the uplift areas. It shows that the Yellow R. not only can flow through many mountains and plateaus in uplifting area, as well as basins and plains situated in subsidence areas, but also can swing between them. The Hetao area is a typical example of riverlake coexistence in multistages. (8) The authors made up these three ideas by preliminary study of the Yellow River,the Yangtze River and other main rivers in China. The first one is antiquity of the Chinese rivers. The Yellow River, the Yangtze River and other Chinese main rivers are all very old; most of them have been formed at the earlymid Miocene at least. However, their primary development situations are still not understood by limited to the difference of researchful profundity. The second is the “riverlake coexistence theory”. Some tectonic lakes or barrier lakes developed in the long history of the river development by endogenic and exogenic geological processes (tectonic movement, faulting, seismicity, collapse, landslide and so on), but the rivers can through these lakes or flow beside them. It should be open to question about the viewpoint of explanation for the formation process of the Yellow River and other Chinese main rivers. If we really want to know the long rivers of China whether be derived from a process of river capture, we should to study them from the Paleogene at least. The third, once a long river was formed, the erosional and depositional power of it is very strong and hard to be blocked. It can cut the mountains and plateaus which are in uplifted area, form many gorges and river terraces, and it also can fill the basins and plains which are in subsidence areas, accumulated thick alluvium. Even the river longitudinal profiles could be changed constantly in different tectonic blocks, each river could maintain roughly parabolic shape at most of time except for the small headwater or gully. Finally, it still needs further study about the old course location and evolution history of the ancientYellow River. And crustal vertical movement range and rate of the Hetao area also needs to be analyzed in our following work by comparative study between the late Cenozoic strata in the Hetao Basin and the ancient Yellow River sediments outside the basin where locations have been tectonized.