"Their failure to give anything of a proof must have frustrated Sierpinski, for it appears that twenty years later he still did not know how to divide by three."