The test-development process used to develop criterion-referenced tests in reading for a large-city school district's system-wide testing program is described. Two types of instruments were developed: (a) a measure of students' ability to read text-books at their assigned grade levels and (b) tests of enabling skills emphasized in the basal reading program. Item construction and selection techniques are described as well as the strategy underlying the use of the two types of tests. The paper is especially relevant to situations where criterion-referenced testing is applied outside the context of instructional development models. (Author)