Abstract
Plant species differ in the amounts of energy allocated to different reproductive tissues, driving differences in their ecology and energy flows within ecosystems. While it is widely agreed that energy allocation is key to reproductive outcomes, few studies have estimated how reproductive effort (RE) is partitioned among different pools, for multiple species in a community. In plants, RE can be partitioned in several meaningful ways: seed versus non-seed tissues; into flowers that form seeds and those that fail to develop; into pre- versus post-pollination tissues, and into successful versus aborted ovules. Evolutionary theory suggests several hypotheses about how these tissues should be coordinated across species. To quantify variation in allocation to different reproductive tissues, we collected detailed RE measurements for a year from 14 perennial species in a recurrent-fire coastal heath community in eastern Australia. Overall we found that total accessory costs - the proportion of RE not directly invested in provisioning the seed - were very large, varying from 95.8% to 99.8% across the study species. These results suggest that studies using seed or fruit production as measures of RE may underestimate it by 10- to 500-fold. We propose a suitable alternative that well-approximates true RE. When comparing species, we found strong support for three evolutionary trade-offs that are predicted to arise when a given energy pool is divided into different tissue masses and counts across species: 1) between successful pollen-attraction costs and mature ovule count, 2) between total reproductive costs and seed count, and 3) between seedset and relative investment in pollen-attraction costs. As a result of these trade-offs, species were also predicted to show coordinated shifts in the amounts invested in floral construction, in seedset and seed size. These shifts in investment were indeed observed, with the amount allocated to discarded tissues increasing with seed size and the amount allocated to pollen-attraction decreasing with seed size. It is already well-established that the seed size axis aligns with the colonization-competition life history spectrum; here we show that relative construction costs of pollen-attraction versus provisioning tissues and seedset are also part of this trajectory, expanding our understanding of the relatives sizes of floral and fruiting structures observed across angiosperms.
