In the past, science textbooks have been the target of criticism by educators and feminists in studies of sex bias in education. The purpose of this study was to develop a useful tool that could be used in formative or summative evaluation of the level of sexism in science textbooks and as a research tool to study the effect of sexist instructional materials on the course choices made by secondary school students. Seven popular grade 10 biology textbooks were examined for their representation of women in illustrations of scientific activity. Three variables were examined in each of the biology textbooks; gender represented by the image; the presence/absence of dominance between persons in an illustration; and the traditional/nontraditional nature of the activity of the individuals in the illustration. It was expected that these factors would help to explain the variation in the impression left with a person by these texts. Using multiple linear regression methods, it was found that these three factors accounted for a significant portion of the variance among the textbooks. Another factor, intensity, the number of illustrations which carry these socializing messages in a given text, has emerged from the study of this data. (Lists of 39 references and the 7 textbooks are provided.) (Author)