Pigmentation has emerged as a premier model for understanding the genetic basis of phenotypic evolution, and a growing catalog of color loci is starting to reveal biases in the mutations, genes, and genetic architectures underlying color variation in the wild. However, existing studies have sampled a limited subset of taxa, color traits, and developmental stages. To expand the existing sample of color loci, we performed QTL mapping analyses on two types of larval pigmentation traits that vary among populations of the redheaded pine sawfly (): carotenoid-based yellow body color and melanin-based spotting pattern. For both traits, our QTL models explained a substantial proportion of phenotypic variation and suggested a genetic architecture that is neither monogenic nor highly polygenic. Additionally, we used our linkage map to anchor the current genome assembly. With these data, we identified promising candidate genes underlying (1) a loss of yellow pigmentation in populations in the mid-Atlantic/northeastern United States [C locus-associated membrane protein homologous to a mammalian HDL receptor-2 gene () and lipid transfer particle apolipoproteins II and I gene ()], and (2) a pronounced reduction in black spotting in Great Lakes populations [members of the gene family, tyrosine hydroxylase gene (), and dopamine -acetyltransferase gene ()]. Several of these genes also contribute to color variation in other wild and domesticated taxa. Overall, our findings are consistent with the hypothesis that predictable genes of large effect contribute to color evolution in nature.