A fundamental tenet of all recent theories of comprehension, problem solving and decision making is that success in such cognitive arenas depends on the activation and appropriate application of relevant pre-existing knowledge (Spiro et al., 1987; Cited in Richards and Renandya, 2002). In this sense, understanding texts is based on the reader’s general knowledge of the world (schemata) and the extent to which the relevant pre-existing knowledge is activated during the actual process of reading. But few empirical data are available about the processes by which schemata are evoked (Carrell, 1987).This study is an attempt to see how much of the schema is activated in L2 reading by adult ESL students who are proficient readers in their native language. An experiment is designed to investigate to see the ‘how' of schema activation to see whether the relevant part of the schema or the whole schema is activated. Sixty third semester students (30 males and 30 females) studying ESL at Isfahan teacher training centers (Teachers’ university), participated in the experiment. The subjects were first told to read a list of ten sentences (Read list) about the kitchen in English. To forget what they had already read, they were given some math problems. Then, the subjects were asked to search in the Test list (consisting of the ten sentences in the Read list and other sentences not existing in the Read list concerning both the kitchen or other parts of the house) and check the sentences they had already read in the Read list. The mistakes they made in checking the ten sentences were of two kinds: Relevant (about the kitchen but not existing in the Read list) and irrelevant (about other parts of the house). The results of t-value computation indicated that most subjects activated the relevant part of the schema (the kitchen). It also showed that significantly more male subjects activated the relevant part of the schema than females.