Classical cloze tests have been criticized as unsatisfactory operationalizations of the concept of reduced redundancy and the theory of expectancy grammar. Use of the C-Principle is proposed as an improvement over the classical cloze procedure. Such tests are constructed according to a technique that deletes the second half of every second word in authentic texts, beginning with the second word in the second sentence and skipping single-letter words. Four to six similarly treated texts are combined in a C-Test. Only entirely correct restorations by the examinee are acceptable. The advantage of this technique is that redundancy is reduced at the word level rather than at the sentence level. The texts chosen should be brief, authentic, and neutral in subject matter. Tests for beginning learners are to be constructed from simpler texts, with a generous amount of time allowed for completion. Because of the authenticity of texts, item analysis can be performed only at the text level; it is defined as the percentage of correct solutions. In several years of testing in several languages (native, second, and foreign), the reliability and validity ratings of the technique have ranged from acceptable to high. Investigations have also established its construct validity. C-Tests can be used for student placement and diagnosis, and as research instruments. (MSE)