Since an author does not provide all critical information, a text is never fully explicit; it demands that the reader/listener supply missing elements in order to form a coherent and logical whole. A study was conducted to explore both the types of responses urban first grade students made when asked to predict a narrative ending and differences in ability between the fall and spring semesters. Two first grade classes in each of five different reading programs served as subjects. The reading programs were (1) Lippincott (1975), (2) Language Experience, (3) Houghton-Mifflin (1979), (4) Open Court (1979), and (5) Ginn 720 (1976). These programs included analytic and synthetic, language experience, and linguistic approaches. In the fall, each child was read part of a story and asked how he or she thought the story would end. Each response was tape recorded, transcribed, and analyzed. The same procedure was followed in the spring. Qualitative measures were concerned with the convergent or divergent nature of the predictions as well as the changes that occurred from fall to spring. Results failed to show a relationship between the reading program used and the type of response given. Overall, the study found that students were capable of responding with both convergent and divergent predictions at the beginning and end of first grade. (FL)