since no method for calculating gene flow is infallible, conclusions should, whenever possible, be based on estimates that have been obtained using two or more different methods. If these are more or less in agreement then we can be reasonably confident that our findings are correct; however, discrepant results are not uncommon and there are several possible reasons for this. We have noted already that indirect estimates may not reflect the actual extent of gene flow because the calculation of Nem is based on an unrealistic model. Although assignment tests are becoming an increasingly popular alternative, there is still a need for rigorous testing of different models against various combinations of sample sizes, marker polymorphisms, and demographic parameters. At the same time, direct estimates often suffer from their inability to detect more than a very small proportion of dispersing individuals, for example dispersal estimates based on mark--recapture will be flawed if attempts at recapture are made at the wrong time or in the wrong place. An example of how mark--recapture can lead to underestimates of dispersal was provided by a study of yellow-bellied marmots (Marmota flaviventris), in which many of the dispersal events (65 per cent of males and 42 per cent of females) that were picked up by radio tracking were missed completely by mark-recapture (Koenig, van Vuren and Hooge, 1996). Parentage studies will also remain inconclusive unless the number of potential parents is very small. Table 4.4 summarizes some of the attributes of the different methods for estimating dispersal and gene flow.
On a final note, it is worth pointing out that the estimation of dispersal, as opposed to gene flow, from direct methods and assignment tests is not necessarily a drawback because dispersing individuals can have an impact on an invaded ecosystem even if subsequent reproduction does not occur. Parasites and microbes, for example, can decimate a population through disease even if they do not subsequently reproduce. We could argue, therefore, that estimates of both dispersal and gene flow are needed before we can have a truly comprehensive picture of population dynamics.
Was this article helpful?