The paper discusses the limitations of using indirect measures, such as FST, to estimate gene flow and migration in populations. While FST is a useful measure of genetic differentiation, it is based on assumptions that are often not met in real populations. The island model, which underlies many FST-based estimates, assumes a large number of populations, random migration, no selection or mutation, and equilibrium between migration and genetic drift. However, real populations often deviate from these assumptions, leading to inaccurate estimates of migration rates. The paper highlights several factors that can affect FST estimates, including selection, mutation, spatial structure, and population dynamics. For example, selection can increase or decrease FST, and spatial structure can lead to non-random migration patterns. Additionally, population size and migration rates can vary across populations, which can bias FST estimates. The paper also notes that FST is a nonlinear function of migration rate, making it difficult to accurately estimate migration rates from FST values. The paper argues that while FST is a valuable tool for studying genetic structure, it should not be used to directly estimate migration rates. Instead, more direct methods of measuring migration, such as mark and recapture, should be used when possible. The paper also emphasizes the importance of considering the limitations of FST-based estimates and the need for more realistic models that account for the complexities of real populations. Overall, the paper concludes that while FST is a useful tool, it should be used with caution and in conjunction with other methods to accurately estimate migration rates.The paper discusses the limitations of using indirect measures, such as FST, to estimate gene flow and migration in populations. While FST is a useful measure of genetic differentiation, it is based on assumptions that are often not met in real populations. The island model, which underlies many FST-based estimates, assumes a large number of populations, random migration, no selection or mutation, and equilibrium between migration and genetic drift. However, real populations often deviate from these assumptions, leading to inaccurate estimates of migration rates. The paper highlights several factors that can affect FST estimates, including selection, mutation, spatial structure, and population dynamics. For example, selection can increase or decrease FST, and spatial structure can lead to non-random migration patterns. Additionally, population size and migration rates can vary across populations, which can bias FST estimates. The paper also notes that FST is a nonlinear function of migration rate, making it difficult to accurately estimate migration rates from FST values. The paper argues that while FST is a valuable tool for studying genetic structure, it should not be used to directly estimate migration rates. Instead, more direct methods of measuring migration, such as mark and recapture, should be used when possible. The paper also emphasizes the importance of considering the limitations of FST-based estimates and the need for more realistic models that account for the complexities of real populations. Overall, the paper concludes that while FST is a useful tool, it should be used with caution and in conjunction with other methods to accurately estimate migration rates.