Soil evaporation (Es) is a non-productive loss of water during crop production and can account for a substantial fraction of total water loss. While there have been many field determinations of Es in individual crops, Es is seldom quantified at the annual cropping system scale, a case in point being the extensive rice-wheat (RW) system of north-west India. The sustainability of this system is threatened by groundwater depletion, so methods to reduce Es are needed. Rice straw mulch reduces Es from wheat, but whether the residual mulch affects Es during the subsequent fallow and rice crop is not known. The effects of zero tillage (ZT) for wheat and rice on Es are also unknown. Therefore, given the interest in changing to conservation agriculture RW systems, the effects of tillage and mulch on Es were determined in a dry-seeded RW system field experiment conducted over two years in Punjab, India. Total Es from the ZT dry seeded RW system was 600−700 mm year−1. The majority (56–66 %) of this loss occurred during the rice phase, and 22 % and 12–22 % during the wheat and fallow phases, respectively. Almost half the loss during the rice phase occurred during establishment of the dry seeded rice (DSR). Tillage method had no effect on Es from the mulched wheat crop, however conventional tillage (CT) significantly increased Es in non-mulched wheat, probably because the latter required two more irrigations than the non-mulched ZT crop. Mulching of wheat with rice straw decreased Es of CT wheat by 50 mm, and of ZT wheat by 32−34 mm but did not affect Es during the fallow period after wheat harvest. Methods for reducing Es from DSR, especially during the establishment phase, should be given high priority given that around 30 % of the total system Es occurred during this phase.