<p>The importance of populations inhabiting ecologically marginal areas has been extensively debated and shown to depend largely on their adaptive value. These populations may harbour unique genetic combinations essential for long-term resilience. This study investigates the adaptive value of marginal populations through a multidimensional genomic analysis that integrates selective signals, neutral processes, and temporal scales. Using targeted exome sequencing, we searched for genomic signatures of divergent selection between environmentally marginal and core populations of the Mediterranean alpine species <i>Silene ciliata</i>. We also analysed genetic population structure, demographic history, and selective sweeps at different levels. Populations from core and marginal environments showed similar values of nucleotide diversity, inbreeding, and relatedness. Demographic reconstructions revealed that marginal, low-elevation populations functioned as stable historical rear-edges (refugia) during post-glacial upslope colonization, retaining high effective population sizes despite current demographic decline. We detected putative signals of diversifying selection involving 11 SNPs in nine genes and five gene pathways (e.g., RNA transport and stress-response metabolism), suggesting that local adaptation persists despite high gene flow. Our findings demonstrate that current environmental suitability is a poor predictor of genetic diversity, which is instead primarily shaped by historical stability. These results underscore the critical value of marginal populations as evolutionary reservoirs that can provide the genetic variation necessary for adaptive responses to climate change through gene flow.</p>