Wednesday, August 5, 2009

PS 54-139: Cost of reproduction and artificial defoliation in Aplectrum hyemale (Orchidaceae)

Hetal Rawal and Thomas P. Rooney. Wright State University

Background/Question/Methods

Reproductive costs represent a common life history trade-off in plants.  When combined with other stressors like deer herbivory, reproduction can lead to declines in growth, future reproduction, and survivorship.  We examine the cost of herbivory and reproduction in Aplectrum hyemale, perennial deciduous forest herb with a unique phenology.  A single leaf emerges in early fall, photosynthesizes through the winter, and senesces in spring as the canopy closes.  Populations are often small and highly fragmented, making them vulnerable to stochastic processes.  We conducted a defoliation experiment to determine how tissue loss and reproduction affect growth rates. We predicted that defoliated plants would exhibit lower growth rates than control plants, with plants defoliated shortly after shoot emergence exhibiting the greatest declines in growth. We also predicted that reproductive plants would have lower growth rates than non reproductive plants. We tagged plants and assigned them to four treatments: control, defoliation shortly after emergence, defoliation midway through the season, and defoliation prior to leaf senescence.  To examine the cost of reproduction we grouped plants into two groups: flowering and non-flowering.  We used ANCOVA to analyze how defoliation and reproduction affect relative growth rate (RGR) of shoots.  Leaf area was our covariate.

Results/Conclusions

We found that defoliation did not influence growth rates, but reproduction depressed future growth rates. Non-flowering plants exhibited a growth rate 41% greater than that of flowering plants (ANCOVA; df = 1, 144; F = 8.74; P = 0.003), after controlling for plant size.  We conclude that defoliation had no effect on growth but flowering plants incurred a significant cost of reproduction. Defoliation does not seem to reduce RGR over short time scales, but we caution against extrapolating these results to longer time scales.