Sixty undergraduates were given one of five programmed mathematics texts. The texts were either completely programmed, programmed only for definitions, programmed only for calculations, programmed only for conceptual information, or not programmed at all. Time spent reading each section of the eight-unit text was recorded and a 24-question posttest was administered. The results showed that although programming increased the time spent reading the programmed sections, it did not enhance posttest performance on the material in those sections. An explanation of programmed instruction based on a simple time-spent model is discussed. (Author/MK)