SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. References

In this issue of Molecular Ecology, authors Robledo-Arnuncio & Garcia present a compelling approach for quantifying seed dispersal in plant populations. Building upon methods previously used for quantification of pollen dispersal, the authors not only examine the behaviour of the model with respect to sample sizes, dispersal distance, and the kurtosis of the dispersal function but also provide an empirical example using Prunus mahaleb.

The distribution of genetic variation in plant populations is the result of two separate processes: the dispersal of pollen from the pollen donor to the recipient followed by the subsequent dispersal of seeds away from the maternal individual. The relative distance which genetic material is moved during each step has asymmetric effects on population genetic structure. If pollen dispersal is panmictic and seed dispersal is limited, then the formation of spatial autocorrelative genetic structure will accumulate because proximate individuals will have an increased probability of sharing at least one parent. Conversely, if pollen dispersal is limited and seed dispersal is panmictic, then we expect little spatial genetic structure as full- and half-sibs are being randomly distributed across the landscape. Despite the importance of both dispersal vectors for overall population genetic structure, the vast majority of recent emphasis based upon genetic analyses has focused on the pollen mechanism.

The Robledo-Arnuncio & Garcia model presented in this Molecular Ecology issue brings the analysis of seed dispersal up to par with current approaches for quantifying pollen dispersal processes. In fact, their model is a reduced parameter case of the more general maximum-likelihood approach used to estimate two-dimensional isotropic pollen dispersal kernels (e.g. Adams & Birkes 1991; Burczyk et al. 2006). Due in part to the longevity of this approach, the behaviour of these models have been extensively investigated as have the statistical mechanisms required for accurate parameter estimation. It is generally accepted that the density function describing the probability of propagule dispersal away from the parent is highly skewed with the vast majority of dispersal distances being relatively local. However, these distributions have a long tail that contains a nontrivial percentage of dispersing propagules that travel considerable distances (e.g. Okubo & Levin 1989). Estimation of a dispersal kernel allows for a general characterization of the density function for dispersal distances and provides a robust tool for comparison among species, populations, or other biologically meaningful strata. As we look forward to how models such as this one can be utilized in natural populations, there are several challenges to be confronted as well as exciting opportunities to be explored with regard to seed dispersal.

First, the estimation of dispersal kernels has recently become a topic of great interest, especially in the context of pollen-mediated gene escape for transgenic crops (e.g. Lavigne et al. 1996; Baker & Preston 2003). The primary focus has been on the behaviour of the tail. The long tail present in these dispersal distributions can have significant consequences for gene escape. Empirically, characterization of the tail has proven to be logistically difficult due to sampling requirements to capture rare long distance migrants. While mostly discussed in the context of pollen escape, Williams et al. (2006) showed that the same concerns are present in seed-mediated gene escape. Continued theoretical and empirical work is needed to gain a better understanding of the behaviour of these long dispersal tails.

Another challenge to be met is that this model implicitly assumes perfect identification of maternal individuals. With the use of diploid maternal tissue such as the endocarp or wing tissue on seeds (Ziegenhagen et al. 2003; Grivet et al. 2005), identification of maternal individuals from multilocus genotypes is much more tractable than then attempting to identify both parents from only offspring genotypes. While protocols for the development of highly polymorphic microsatellites markers continue to be refined, it is still a nontrivial process in terms of time and resources to develop a robust set of markers for less-studied taxa.

The largest challenge we face as we utilize these models is the underlying assumption that the dispersal of propagules proceeds in a purely diffusive fashion that can be adequately described by a two-dimensional density function that is both continuous and symmetric in all directions (e.g. Levin & Kerster 1974; Devlin et al. 1988; Erickson & Adams 1989; Austerlitz & Smouse 2001 Robledo-Arnuncio & Garcia 2007). Under simulation conditions, continuous symmetric distributions are easily approximated, facilitating easy characterization of possible bias in model parameters. However, in natural populations, there are several biotic and abiotic factors that may influence the movement of genes across a landscape in a fashion that deviates from continuity and/or symmetry. For example, Grivet et al. (2005) showed that the acorn woodpecker (Melanerpes formicivorus L.) dispersed the seeds of Quercus lobata into discrete granaries in such a manner that pairs of granaries had a very low probability (< 1.5%) of containing seeds from the same maternal individual. It is difficult to determine if such discrete and localized placement of seeds across the landscape causes a systematic bias in parameter estimation in models that assume a continuous density function. At present, it is largely unknown the extent to which heterogeneity in the behaviour of dispersing animals influences the estimation of dispersal kernel parameters, especially if the behaviour is species-specific or modified by the particular ecological context.

Even for species within which propagules are dispersed passively by wind, symmetry in dispersal away from the source may not exist. In the analysis of pollen dispersal in a stand of Pinus sylvestris, Robledo-Arnuncio & Gil (2005) showed significant deviations from uniform distribution of pollen away from dispersing individuals. While wind-dispersed pollen may respond differentially to wind, Nathan et al. (2002) showed that variation in thermal uplift can significantly influence pollen dispersal distance for several forest tree species. Given heterogeneity in canopy structure across the landscape and its effects on pollen movement (e.g. Dyer & Sork 2001), it would be easy to argue that the net effect of forest heterogeneity would be asymmetric seed dispersal shadows. Other factors that could cause deviations from isotropy include microclimatic heterogeneity, prevailing winds, and the idiosyncratic behavioural responses of dispersing animals.

A consequence of both noncontinuous and/or asymmetric dispersal of seeds is that estimates of the dispersal kernel will not adequately capture the dynamics of the real dispersal process. Clearly, the dispersal kernel is highly informative for comparing dispersal potential across taxa or specific sampling strata. However, in many cases the deviation from isotropy within sites may be of interest, especially when attempting to couple dispersal processes with ecologically relevant information. The Robledo-Arnuncio & Garcia model represents a step forward in the tool set available to examine the functional dispersal mechanisms that Mayr (1963) refers to as the ‘internal cohesion of the gene pool’. The challenge that remains is for us to go out and apply it to natural systems.

References

  1. Top of page
  2. Abstract
  3. References
  • Adams WT, Birkes DS (1991) Estimating mating patterns in forest tree populations. In: Biochemical Markers in the Population Genetics of Forest Trees (eds FineschiS, MalvoltiME, CannataF, HattemerHH), pp. 157172. SPB Academic Publishing, The Hague, The Netherlands.
  • Austerlitz F, Smouse PE (2001) Two-generation analysis of pollen flow across a landscape III: impact of adult population structure. Genetical Research, 78, 271280.
  • Baker J, Preston C (2003) Predicting the spread of herbicide resistance in Australian canola fields. Transgenic Research, 6, 731737.
  • Burczyk J, Adams WT, Birkes DS, Chybicki IJ (2006) Using genetic markers to directly estimate gene flow and reproductive success parameters in plants on the basis of naturally regenerated seedlings. Genetics, 173, 363372.
  • Devlin B, Roeder K, Ellstrand NC (1988) Fractional paternity assignment: theoretical development. Theoretical and Applied Genetics, 76, 369380.
  • Dyer RJ, Sork VL (2001) Pollen pool heterogeneity in shortleaf pine, Pinus echinata. Molecular Ecology, 10, 859866.
  • Erickson VJ, Adams WT (1989) Mating success in a coastal Douglas-fir seed orchards as affected by distance and floral phenology. Canadian Journal of Forest Research, 19, 12481255.
  • Grivet D, Smouse PE, Sork VL (2005) A novel approach to an old problem: tracking dispersed seeds. Molecular Ecology, 14, 35853595.
  • Lavigne C, Godelle B, Reboud X, Gouyon PH (1996) A method to determine the mean pollen dispersal of individual plants growing within a large pollen source. Theoretical and Applied Genetics, 93, 13191326.
  • Levin DA, Kerster HW (1974) Gene flow in seed plants. Evolutionary Biology, 7, 139220.
  • Mayr E (1963) Animal Species and Evolution. Harvard University Press, Cambridge, Massachusetts.
  • Nathan R, Katul GG, Horn HS, Thomas SM, Oren R, Avissar R, Pacala SW, Levin SA (2002) Mechanisms of long-distance dispersal of seeds by wind. Nature, 418, 409413.
  • Okubo A, Levin SA (1989) A theoretical framework for data analysis of wind dispersal of seeds and pollen. Ecology, 70, 329338.
  • Robledo-Arnuncio JJ, Garcia C (2007) Estimation of the seed dispersal kernel from exact identification of source plants. Molecular Ecology, 16, 50985109.
  • Robledo-Arnuncio JJ, Gil L (2005) Patterns of pollen dispersal in a small population of Pinus sylvestris L. revealed by total-exclusion paternity analysis. Heredity, 94, 1322.
  • Williams CG, LaDeau SL, Oren R, Katul GG (2006) Modeling seed dispersal distances: implications for transgenic Pinus taeda. Ecological Applications, 16, 117124.
  • Ziegenhagen B, Liepelt S, Kuhlenkamp V, Fladung M (2003) Molecular identification of individual oak and fir trees from maternal tissues of their fruits or seeds. Trees Structure and Function, 17, 345350.

Dr Rodney Dyer's research focuses on understanding how genes are dispersed in both space and time.