Modular plants are structurally stable systems. Perturbations in the genome that arise during growth (somatic mutations) are, it is conjectured, detected and eliminated by diplontic selection at the apical meristem. This requires accurate control, which, it is suggested, results in the accuracy of module arrangement, or phyllotaxy. During differentiation vegetative modules tend to retain autonomy and totipotency; the differentiated state is maintained by a constant flux of signals between parts. This enables damage to be repaired and differentiation to be adjusted to the availability of resources. The differentiation of modules to form flowers, on the other hand, is achieved by loss of totipotency, by hierarchical organization of the genotype, and by tissue-specific signals between parts. Elaborate, but fixed-function, structures can be produced in this way. The physiology of modular plants is best described in terms of cooperation, not competition, between modules. The theory of multicomponent systems predicts that as plants increase in size, structural stability of growth will be lost unless the connectance between modules is kept below a critical value. Experiments confirm that the exchanges of assimilate between modules are limited, but not fixed (the system can adapt to damage). The distribution system is vulnerable to exchanges that might benefit individual modules but that would reduce the inclusive fitness of the genome. Such exchanges are controlled by the organized senescence of branches, leaves, fruit and ovules.