The new unified context for crystallization of linear macromolecules from the melt, established by recent work, is expounded. The morphology of spherulites shows that they form because of a short-range force operative at the branch points of dominant lamellae that causes them to diverge at noncrystallographic angles of about 20°. Work on monodispersen-alkanes has confirmed the identification of this short-range force with the pressure from dynamic cilia during growth. Accordingly, spherulites, like chain folding, are a direct consequence of molecular length. It is suggested that the tendency to form coarse spherulites even for extended-chain growth at lower temperatures may result from the increasing difference between the lengths of nucleus and molecule. Crystallization on linear nuclei has been used to maximize the concentration of segregants at the growth front and to demonstrate cellulation in undoped polymers for the first time. The behavior of branched polyethylenes differs from the uniform growth of the linear polymer in coarsening and developing protuberances at the growth front, all the while slowing continuously toward an asymptotic steady state; differences of detail may be useful in distinguishing polymers of different catalytic origin and branch content. Spherulitie growth is also nonlinear for these polymers, but is always faster than for rows. When there is sufficient segregation, spherulites themselves cellulate, increasingly so for higher branch content. Cellulation is thus an uncommon and secondary process may be superposed on regular spherulitie growth beyond a certain distance. Cell dimensions do not scale with the diffusion length; in so doing, the phenomenon displays new physics.