The design of experiment can be defined as a method of sampling. It has been employed and developed in the Rothamsted Experimental Station for agriculture as a technique of agricultural experiment in the later nineteenth century and was elaborated and introduced by R.A. Fisher into the field of study where experiments are done. In earlier days, the random sampling method has exclusively been used in the Pearsonian descriptive statistical method but the achievements of Fisher made it clear that the discriptive method is very imperfect one. That is to say, in the Pearsonian system, the descriptive method holds so long as the universe were to be homogeneous. In the practical case, however, the universe we are always facing with is, in most cases, not homogeneous but heterogenous and under these circumstances we can hardly grasp the characteristics of the phenomena by means of descriptive method with any precision. Henceforth the method of inductive statistics was deviced and developed. The industive statistics has something in its theoretical construction quite different from that of descriptive statistics. As, in many practical cases, the universe is constituted from heterogeneous elements, the precision can be expected to be improved if we classify the heterogenous universe into a number of homogeneous strata, and select sampling from this group than to select it at random from the universe. On this account, we are obliged to be in need of the universe. When we classify the universe into a number of strata and the frequency of observation is apportioned to these strata, the number of samples selected is hoped to be proportionate to the size of the universe. The main theme of this paper is to examine the above-said technique of sample design in sampling method. The design of experiment of R. A. Fisher, in effect, is the core in the problems of modern statistic`s and Fisher`s achievements in this field is a monumental contribution to the systematication of statistics together with the analysis of variance and these theory of estimation. It constitutes the theoretical backbone of inductive statistics together with Neyman-Pearsonian Theory of testing hypothesis. Modern statistics (industive statistics) is originated from the classical works Gosset and in developed through E.S. Pearson J. Neyman to R.A. Fisher and the studies in the theory and application of its method is enriched and refined by the works of E.S. Pearson, J. Neyman and A. Wals. As the application of the design of experiment is to the field of the control of experiment and mass production, it is an inevitable outcome that Fisher`s design of experiment came to qualify the methodology of control.