To discover what aspects of number conservation tasks prevent young children from conserving, two components of number conservation tasks were investigated. A test was made of the following two hypotheses: (1) the frequency of conservation responses varies significantly with the materials used, (2) the frequency of conservation responses varies significantly with the number of objects used. Seventy preschool children were tested individually with two materials, candy bars and wood bars, in rows containing three, five and seven objects. Half of the Subjects were randomly selected from three year old classes and the other half from four year old classes. Subjects were labeled as a conserver, variable conserver, or nonconserver, depending on their answers to the experimenter's questions. Variable conservers correctly answered between one and five of six question sets. Results of the 25 variable conservers were analyzed. Using the Wilcoxon Test, it was found that children exhibit significantly more conserving responses when candy is used. Using the Cochran Q Test, it was determined that the probabilities of conserving responses are significantly different with three, five and seven objects. These results suggest that large numbers of objects and irrelevant materials may prevent young children from conserving. Results are discussed in terms of conservation theory and implications for early childhood education. (Author/RH)