Seed interventions are the major agricultural response during emergency and recovery phases of humanitarian relief. They are implemented by diverse agencies, and widely promoted: for instance the FAO alone managed 400 such projects between 2003 and 2005. However, seed aid suffers from a lack of critical attention, perpetuating widespread myths among practitioners, policymakers, and the larger humanitarian community. This paper challenges five predominant myths about seed aid: (1) seed aid is needed whenever food aid is; (2) seed aid can do no harm; (3) disasters wipe out seed systems; (4) effective implementation is a straightforward logistical exercise, and; (5) improved seed is the best form of aid. These myths are juxtaposed with recent empirical work across a range of countries, particularly in Eastern and Southern Africa. The perpetuation of such myths highlights a serious absence of scrutiny of emergency seed aid, and helps explain why such aid is repeated year after year in many sites, with little apparent positive effect. The paper argues that the invisibility of seed aid is a major cause for the lack of oversight and concludes that donors and farmer beneficiaries must become centrally involved in seed aid governance.