With regard to the relation between climatic oscillation and terrace formation, some ideas that I have collected in the course of studying accelerated (man-made) erosion, may be useful in a geo-morphological setting. Huntington’s principle and the idea I used in my paper “The Ice Age in West China” are not necessarily opposed, if one limits the application of Huntington’s principle to semi-arid regions which are just on the margin of being able to support a vegetative cover. In such regions a further decrease in precipitation, leading to serious reduction in the protection afforded by vegetation, would accelerate erosion on hillsides and steep valley floors, and this in turn would lead to deposition of spoil on more gently sloping valley floors lower down. A further oscillation of the climate, in the direction of increasing rainfall, would restore the vegetative cover, reduce erosion, reduce the load in the rivers, and allow downcutting to continue. But a still further increase in precipitation might be expected to increase erosion of the type associated with landslips and flooded streams and rivers; once more increased load will be swept down from the steeper hills and stream-beds to the lower and gentler slopes of the rivers, where it will be deposited. In other words, one may say that there is a level of precipitation, probably a moderately low one, at which (apart from human interference) there is a close vegetative cover and geological erosion is at a minimum; any appreciable departure from this level of precipitation in either direction will cause increased geological erosion in the upper courses of streams, and hence increased deposition in their lower courses.