Towards an optimal sampling effort for paternity analysis in forest trees: what do the raw numbers tell us?

The ever-increasing use of paternity analysis to estimate the dispersal capability of forest trees calls for a quantitative evaluation of potential errors due to sampling design. Previous studies on optimal sampling strategies for seed trapping experiments suggested a link between sampling effort and error rate in the reconstruction of the seed dispersal kernel. We considered 92 papers on paternity analysis to quantitatively assess the sampling strategy used to study the characteristics of pollen dispersal patterns (pollen immigration rate, distribution of male reproductive success and estimates of pollen dispersal kernel parameters). For each studied stand we report data on the sampling effort (the total number of sampled seeds, the number of mother trees and the number of seeds per mother tree) and additional information on the studied species and characteristics of the sampling areas. The reviewed papers used a median of 8 mother trees (acting as pollen traps in paternity analysis studies), a median of 29 seeds per mother tree and a median of 240 total sampled seeds. These are values (especially the number of mother trees) lower than usually found in classical seed trapping studies, for which accuracy and precision of seed dispersal estimates had already been assessed. These findings underline the need of evaluating the consequences of realistic sampling efforts on the estimation of parameters describing the pollen dispersal pattern to provide the basis for meaningful guidelines to paternity analysis.