Abstract
1. The effect of various patterns of flower removal on pod setting was investigated in Lupinus luteus L. Four-fifths, three-fifths, or two-fifths of the flowers of the main inflorescence were removed according to ten different patterns. 2. All flowers could produce pods but later ones were less efficient in doing so. Developing pods had an abscission-inducing effect on later flowers, which became increasingly effective towards the apical part of the inflorescence. More pods were retained when flowers on each consecutive whorl were arranged in a spiral than when the same number was arranged vertically. Pod setting was incomplete when the number of flowers per inflorescence was reduced well below the total number of pods normally present. 3. The number of ovules in consecutive flowers gradually decreased from an average of 5.7 at the base to 4.3 at the top of the inflorescence. The ratio of seeds to ovules fluctuated irregularly between 65 and 94 per cent, and did not indicate a general trend in embryo abortion. 4. The growth-rate of pods at the top of the inflorescence was much slower than at the bottom. Vascular differentiation was almost absent at the top of the inflorescence when the flowers were fertilized, and further vascular tissue was produced only when flowers produced pods.