Tomonaga's ingenious method of the description of collective surface motion is extended to the case of three dimensional nuclei. In this extension some difficulty appears which was not the case with Tomonaga's two dimensional example, and a method to keep out from such a difficulty is given in this paper. Applying this method a Hamiltonian is derived which contains terms standing in one to one correspondence with those of Bohr's together with other correction terms. Brief calculation of the nuclear surface energy is also performed. The method of collective description mentioned here may be useful for many other problems where the collective motion takes place.