Children with autism may struggle in developing conditional discrimination repertoires. Saunders and Spradlin (1989, 1990, 1993) arranged "blocked" teaching trials in which they presented the same sample stimulus repeatedly across trials (in lieu of randomly alternating targets across trials) and then faded the number of trials in each block. We replicated the effects of this blocked-trials procedure in teaching identity matching to a child with autism and evaluated the necessity of fading. Arranging blocked trials facilitated the acquisition of identity matching, but fading the block size was not necessary to maintain discriminated performance. (Contains 1 figure.)