Functional analysis probes were conducted over an extended time period to identify the consequences that maintained infrequent self-injury (occurred 1-5 days monthly). Occurrences of respite care and episodes of physical illness were concurrently measured in order to assess their influence on response-reinforcer relations during functional analysis conditions. Functional analysis of immediate antecedents and consequences produced inconclusive results. A subsequent systematic manipulation of respite care demonstrated a relationship between the presence of self-injury and the occurrence of respite the previous evening. The intervention consisted of altering the respite situation, which resulted in the long-term elimination of self-injury.