Data from: Is a larger refuge always better? Dispersal and dose in pesticide resistance evolution

The evolution of resistance against pesticides is an important problem of modern agriculture. The high-dose/refuge strategy, which divides the landscape into treated and non-treated (refuge) patches, has proven effective at delaying resistance evolution. However, theoretical understanding is still incomplete, especially for combinations of limited dispersal and partially recessive resistance. We reformulate a two-patch model based on the Comins model and derive a simple quadratic approximation to analyze the effects of limited dispersal, refuge size and dominance for high efficacy treatments on the rate of evolution. When a small but substantial number of heterozygotes can survive in the treated patch, a larger refuge always reduces the rate of resistance evolution. However, when dominance is small enough, the evolutionary dynamics in the refuge population, which is indirectly driven by migrants from the treated patch, mainly describes the resistance evolution in the landscape. In this case, for small refuges, increasing the refuge size will increase the rate of resistance evolution. Our analysis distils major driving forces from the model, and can provide a framework for understanding directional selection in source-sink environments.农药抗性演化是现代农业面临的重要课题。高剂量-庇护所（high-dose/refuge）策略将农田景观划分为施药斑块与非施药（庇护所）斑块，已被证实可有效延缓抗性演化进程。然而目前的理论认知仍不完善，尤其是针对有限扩散与部分隐性抗性共同存在的情形。我们基于康因斯模型（Comins model）重构了双斑块模型，并推导出简洁的二次近似表达式，用以分析有限扩散、庇护所规模以及抗性显性程度对高效施药方案下的抗性演化速率的影响。当施药斑块中有少量但可观数量的杂合子能够存活时，更大的庇护所规模总能降低抗性演化速率。但当显性程度足够低时，庇护所种群的演化动力学，其动力间接来自施药斑块的迁移个体，主要反映了整个农田景观的抗性演化情况。在此情形下，对于规模较小的庇护所，增大其规模反而会提升抗性演化速率。本研究从模型中提炼出核心驱动因素，可为理解源-汇生境中的定向选择提供理论框架。