Wednesday, August 5, 2009 - 2:50 PM

OOS 28-5: How, when and where is it good to be small?

Roberto Salguero-Gomez and Brenda B. Casper. The University of Pennsylvania


Projection matrix models are very powerful tools for demographic studies. The latest advancements in elasticity analyses have allowed plant demographers to compare status, competitive strategies, and responses to external agents (e.g. fire, drought, herbivory) among plant species based on three demographic processes: growth, stasis and reproduction. Surprisingly, the possibility of reduction in size (which is implicitly modeled in matrices as retrogression transition probabilities), an innate feature of plants due to their modular design and phenotypic plasticity, has been historically overlooked. This process is of interest because it may represent a side-effect of plant senescence, a little understood phenomenon in plants or because it could, alternatively, be an adaptive strategy to cope with adverse environmental conditions.

We asked how often shrinkage occurs in natural populations as a function of taxonomic family, habitat and growth form, therefore identifying its demographic importance. Our database includes over 380 plant species with different plant life history strategies (annuals, pseudoannuals, herbaceous perennials, succulents, lianas, shrubs and trees). For two species, we further decomposed their demographic dynamics using loop analyses to test our hypothesis about the causes of shrinkage: the result of senescence versus an environmentally induced response.


We found that shrinkage occurs in over 70% of the species in natural populations, yet in a very low percentage of those cases do authors acknowledge its occurrence. Our results are suggestive of both shrinkage that is a result of convergent evolution among unrelated species, and of shrinkage that may have resulted due to phylogenetic inertia in highly modular growth forms. We also raise an important concern in demographic studies employing traditional projection matrices. We found a correlation between (1) the presence/absence and (2) elasticity value of shrinkage and the number and type of classes utilized by the authors, especially in very long-lived species such as shrubs and trees.