and may have a similar effect on the ma-ternal allele of FWA. A study reported in a recent issue of Developmental Cell (15) proposes that imprinting of MEDEA is controlled by antagonism between the two DNA-modifying enzymes, MET1 and DEMETER (15). Thus, imprinting of MEDEA and FWA in endosperm may rely on the same mechanism. Initially, both parental copies are silenced, then DEME-TER removes methylation of FWA 5 ′ re-peats but only for the maternal allele of the central cell, triggering endosperm-specific expression of FWA. Unlike FWA, which is expressed only during formation