We are interested in understanding natural variation in behavior at the genetic and neurological levels. Male courtship behavior in Drosophila is a model system for connecting genes, brain, and behavior using molecular genetics, but little is known about how these behaviors vary or evolve. We used a multi-parent panel of recombinant inbred lines to map natural variation in a key parameter of the Drosophila courtship song (inter-pulse interval), and found several loci that explain modest fractions of song variation. The challenge for this and similar studies is then to map these QTL to the gene level for behaviors which don’t have obvious candidate genes. We are using a two-step approach. First, we use visible markers to select for recombination in intervals of arbitrarily small size. Benign fluorescent markers can be inserted anywhere in the genome with CRISPR, making this a powerful approach. Second, once we have mapped a QTL to a potentially causal gene, we are using the reciprocal hemizygosity test to validate that gene and quantitatively estimate its effect on our trait. These fine-mapping and validation steps should have broad applicability in other systems and for validating candidate genes from all genome-wide methods.