General hypotheses for why seedling dynamics are so important for determining patterns of abundance and diversity in plant communities

Background/Question/Methods

Plant community ecologists generally regard early life-cycle stages (seeds and seedlings), and non-random transitions between them, as especially important for governing the assembly, composition and structure of the mature stage in communities.  Supporting evidence for this view is accumulating, but little attention has been focused on why non-random departures from community drift should be greatest among early life-cycle stages.  Here we present two heuristic models, the ‘Turnover’ and ‘Niche’ models, that serve as alternative hypotheses for why the early seedling stage could be disproportionately important in determining patterns of diversity and relative species abundance.  Both are predicated on a hump-shaped relationship between the proportion of species dying non-randomly (P) and mortality as it accumulates over time, applicable to all size classes of plants from seedlings to reproductively mature individuals.  In these models, a species dies non-randomly when the number of individuals dying over a specified period was significantly different to that expected if the per capita probability of mortality was random with respect to species identity.  In both models, the point of interest is the peak in P, resumed to occur when cumulative mortality is around 50%.

Results/Conclusions

In the Turnover Model, peak P reaches the same amplitude regardless of plant size.  In other words, large plants show the same propensity for non-random mortality as do seedlings and, given time, P among the largest size classes will be just as great as P among the seedlings of a community.  However, these early stages contribute more to community structure because seedlings turn over so much faster.  By contrast in the Niche Model, seedlings show a significantly greater propensity to nonrandom mortality, such that at any given point along the cumulative mortality axis, P among seedlings is significantly higher than that for larger size classes.  A test of these models using long-term (35 y) demographic data from a rainforest site in Australia shows support for both the Turnover and Niche Models, but it is unclear why seedlings show a greater propensity to nonrandom mortality.  One hypothesis is that interspecific variation in key functional traits that determine survival is significantly greater at the seedling stage than at larger stages, permitting a greater degree of nonrandom mortality.  Faster turnover in seedlings is probably commonplace, but our demographic data suggests that an additional ‘niche’-based non-randomness is also hyperactive in the earliest stages.