How do neural circuits evolve to produce new behaviors? Several decades of research in the field of evolution and development have shown us how gene regulatory networks can evolve to produce different animal morphologies. However, when the output of a gene network is behavior, rather than anatomy, we are almost completely in the dark. What kinds of genes and mutations contribute to behavioral evolution? Does change occur through a small number of genes of large effect, or through many quantitative loci? Does it occur at the level of developmental transcription factors capable of re-wiring entire neural networks, or by subtly altering the activity of conserved neural circuits? Innate male courtship in fruit flies is relatively simple, stereotyped, controlled by a defined set of neurons and evolves rapidly, making it an ideal system to study this problem. In my seminar, I will discuss how we are leveraging the increasing power of the whole Drosophila subgroup as a genetically tractable system to understand how the male courtship neural network has evolved using a variety of approaches, such as transferring candidate loci between species, quantitative genetic mapping and functional assays of neural activity in multiple species.