The theory developed in 1960 by Wagner, Hamilton and Seidensticker (WHS-theory) to explain observed crystal growth phenomena in Ge is critically reviewed and shown to be capable of explaining preservation of ABC stacking order in two dimensions in fcc crystals of effectively spherical closed shell molecules. In order to preserve this stacking order in all directions, i.e. to explain isotropic three-dimensional fcc crystal growth, the WHS-model is extended to contain at least two (crossing) twin lamellae, rather than one. The implications for the atomic arrangement in the crossing region are examined, with the main result that local fivefold symmetry is to be expected. This is related to the frequent observation of multiply twinned particles (MTPs), exhibiting fivefold symmetry, of fcc materials whose preference for fcc over hcp cannot be explained satisfactorily. Accordingly, it is proposed to view the atomic arrangement in the crossing region not as a result of cross-twinning, but rather as its origin, i.e. to think of fcc crystal growth as a process starting with the coalescence or intergrowth of decahedra with fivefold symmetry. Experimental evidence in support of this model is given. A possible implication of the proposed growth model, i.e. that the observed crystal structure of a substance is not necessarily that of lowest free energy for the infinite crystal, but rather corresponds to a local minimum that is made accessible by the kinetics of a particular growth process, is discussed.