Abstract
The position effects on the abortion of flowers, fruits, and ovules were examined in relation to the resource limitation hypothesis using a temperate legume species, Vicia cracca. Fifty plants were randomly sampled from a natural population located at the Belmont Conservation Site in Belmont, MA, USA. Due to indeterminant growth and acropetal flowering of this species, the position of flower or fruit on the plant was able to be coded by the combination of architectural levels such as plant half (lower and upper half) and inflorescence half (lower and upper half) within each plant half. Overall, only 17.3% of the flowers on an inflorescence began to develop into fruits, while overall 36.0% of the fruits initiated failed to mature. Consequently, 11.5% of the flowers successfully matured into fruits. The mean flower number per inflorescence differed significantly among plants as well as between plant halves. Although the probability of fruit initiation was higher in the lower halves of both plant and inflorescence than in the upper halves of the latter, fruit abortion rate differed significantly only between plant halves. The overall mean seed set was 33.6%, implying that about two out of six ovules in a fruit matured into seeds. The seed set was independent of both plant identity and architectural level unlike other reproductive characters examined. When the ovule developmental stage was examined on data pooled over all fruits, the proportion of ovules in varying developmental stages decreased in order of early abortion, seed maturation, late abortion, and seed predation (48.4, 34.9, 12.2, and 4.5%, respectively). A within-fruit ovule position was also used as a class variable for the analysis of position effects on ovule development. All architectural levels considered exerted singificant influence over the ovule development. In particular, ovule development was strongly affected by the within-fruit ovule position. Ovules in both extreme ends within a fruit tended to abort early, while those in the middle position were more likely to mature into seeds. The strong position effects detected from the flowering to seed maturation stage were interpreted as an indication of competitive interaction among reproductive characters which are largely constrained by plant architecture.