The theory of non-local field was the result of Yukawa's endeavour to overcome the point model of elementary particles. He believed firmly that the elementary particles are not mathematical point, and tried to introduce a finite extension to them. The motive was on the one hand to obtain a finite theory expecting this extension to work as a kind of cutoff factor, and on the other hand to construct a unified theory of elementary particles by identifying various eigenstates of the internal motion carried by the extended structure with the actual particles. After the completion of the meson theory he concentrated his effort to this problem. But unlike the brilliant success of the meson theory no remarkable success has been obtained at least until now. But I feel that the notion of finite extension still deserve serious study. The revival of the string model on the Planck mass level is one reason. Another reason is the fact that it leads very naturally to the concept of spinor.1> That is because parameter space describing the motion of an extended body is doubly connected. So in this case we have no reason to require that the wave function must be single valued. If we allow double valued wave function half-integer spin eigenvalue is possible, and in fact the eigenfunction of a rigid sphere, for example, belonging to the eigenvalue l =1/2 is just the spinor.2' In view of these it would be of interest to try to seek for a possibility to check the adequacy of Yukawa's idea more directly. For this purpose let us discuss in some detail how to describe the motion of an extended body. Usually it is understood that this can be done by specifying the orientation of a body fixed frame B --the origin of which is assumed to be placed at the center of mass of the body for definiteness-relative to a laboratory fixed frame L. But this is by no means trivial if we consider the possibility that our space-time is curved according to general relativity. This is because in this case we cannot define a line parallel to L through the origin of B. So it is in general impossible to specify the orientation of B relative to L. Therefore, in order to describe the motion of an extended body in conformity with the principle of general relativity, we must specify the orientation of B relative to a reference frame the origin of which is always placed the origin of B. This is kind of gauge philosophy, and leads to a conclusion that if the center of mass of the body is in a motion, L cannot be fixed to the laboratory but must make a parallel displacement along the path of the center of mass of the body in accordance with its motion.