A parallel study of red cell antibodies was undertaken using two wellknown AutoAnalyzer techniques: the bromelin-methylcellulose (BMC) method as described by Marsh, and the low-ionic strength-Polybrene (LISP) technique of Lalezari. In the present study, minor modifications were made to both methods, without changing the basic principle of antibody detection. The results obtained in this parallel study were compared to standard manual techniques. Screening of 13,135 sera gave 8.7% positive reactions. Of these, 22% were identified manually as being caused by red cell antibodies. Further antibodies could be identified by the machine [described elsewhere]. The sensitivity of the automated methods was generally higher than manual techniques, about 12 times for the BMC and 70 times for the LISP method. Anti-s, -Fy^a, -Jk^a, -Jk^h, and -K were less sensitively detected by the BMC than the most appropriate manual method in a test of 14 commercial typing sera. In no case was the LISP less sensitive than manual techniques. However, hemolyzing antibodies can be missed in rare instances. Anti-D was found at a lower sensitivity level of 20-50 ng/ml with the indirect antiglobulin technique, 5-15 ng/ml with the twostage papain technique, 2-4 ng/ml with the BMC method and 0.2-0.4 ng/ml with the LISP method. Some weak anti-D’s (5%) reacted more strongly in the BMC than LISP. When anti-D preparations used for Rh prophylaxis were quantitated with both methods, the results agreed well.