Snow (1931)
showed that the fate of a branch of a plant that had suffered local herbivory could be determined by correlative effects of other branches. This neglected work was continued in the current study, herbivory being simulated by the removal of leaves of different ages from pea plants with two branches. A damaged branch was suppressed when an undamaged alternative branch was present; otherwise the damage never prevented continued development. The removal of mature leaves had a smaller effect than the removal of immature, expanding leaves. When leaves were removed from both branches it was the branches that suffered less damage to their immature leaves that continued shoot development. Branches from which all photosynthetic leaves were repeatedly removed developed only when they retained their immature leaves and remained dominant, inhibiting the development-but not the photosynthesis -of the other branch on the same plant. Accounting for these results requires mechanisms that compare the different branches of the plant and 'select' for development the ones that have the greatest potential for future, rather than present, photosynthesis. It is concluded that a compartmentalized or sectorial distribution of essential substrates can be modified by correlative relations that are probably mediated by hormones.