The insect wing is a key evolutionary innovation that was essential for insect diversification. Yet despite its importance, there is still debate about its evolutionary origins. Two main hypotheses have been proposed: the paranotal hypothesis, which suggests that wings evolved as an extension of the dorsal thorax, and the gill-exite hypothesis, which proposes that wings were derived from a modification of a pre-existing branch at the dorsal base (subcoxa) of the leg. Here, we address this question by studying how wing fates are initially specified during Drosophila embryogenesis, by characterizing a cis-regulatory module (CRM) from the snail (sna) gene, sna-DP (for dorsal primordia). sna-DP specifically marks the early primordia for both the wing and haltere, collectively referred to as the DP. We found that the inputs that activate sna-DP are distinct from those that activate Distalless, a marker for leg fates. Further, in genetic backgrounds in which the leg primordia are absent, the DP are still partially specified. However, lineage-tracing experiments demonstrate that cells from the early leg primordia contribute to both ventral and dorsal appendage fates. Together, these results suggest that the wings of Drosophila have a dual developmental origin: two groups of cells, one ventral and one more dorsal, give rise to the mature wing. We suggest that the dual developmental origins of the wing may be a molecular remnant of the evolutionary history of this appendage, in which cells of the subcoxa of the leg coalesced with dorsal outgrowths to evolve a dorsal appendage with motor control.