The present research tested the hypothesis that the reading of science text can create new misconceptions in students with incongruent prior knowledge, and that these new misconceptions will be similar to the fragmented and synthetic conceptions obtained in prior developmental research. Ninety-nine third- and fifth-grade children read and recalled one of two texts that provided scientific or phenomenal explanations of the day/night cycle. All the participants gave explanations of the phenomenon in question prior to reading one of the texts and after they read it. The results showed that the participants who provided explanations of the day/night cycle at pretest incongruent with the scientific explanation recalled less information and generated more invalid inferences. An analysis of the participants' posttest explanations indicated that these readers formed new misconceptions similar to the fragmented and synthetic conceptions obtained in developmental research. The implications of the above for text comprehension and science education research are discussed.