"In considering the structure of the universe... Einstein assumed... the irregularities in the distribution of matter can be neglected [i.e., the universe is homogeneous]. ...he noted that the largest velocities then assigned to the stars and nebulae were very small compared with the velocity of light. In 1917... the immense red-shifts of the farther nebulae were unknown. Einstein, therefore, considered a model of the universe in which matter was distributed in a uniform and continuous manner, the relative motion of the various parts being negligible [i.e., a static universe]. Following Seeliger, he found it impossible to regard the system as filling the whole of [infinite] Euclidean space. Also he could not regard the universe as an island in infinite space. For, in applying a well-known theorem of Boltzmann relating the densities at various points of space in which a distribution of particles is moving at random, he showed that zero density at the boundary would necessitate zero density at all points inside. Thus it appeared to Einstein that the universe as a whole could be neither infinite nor have a finite boundary. Hence space as a whole could not be Euclidean. ...In the classical picture time and space are distinct, time being infinite in duration and space Euclidean. In devising an alternative model Einstein retained this world-wide separation of time and space, despite the fusion of the two concepts in General Relativity, but he assumed that space as a whole was of the type known as spherical."