FIFTEEN DAILY DRILLS EMPHASIZING MASTERY OF BASIC NUMBER FACTS, SUCH AS ARITHMETIC OPERATIONS AND UNITS OF MEASUREMENT WERE CONSTRUCTED TO ENCOMPASS 7 PREVIOUSLY REPORTED ATTRIBUTES OF AN EFFECTIVE DRILL. ATTRIBUTES ARE MIXED DRILL, TIME LIMIT, INCREASINGLY DIFFICULT EXAMPLES, THOROUGH COVERAGE, FREQUENT AND SMALL AMOUNTS, VERBAL PROBLEMS, DIAGNOSIS FACILITATION. SHORT DRILLS AVERAGING 20 PROBLEMS REACHED THE 41 GRADE 4 PUPILS (AVERAGE IQ WAS 130) VIA A COMMERCIALLY AVAILABLE TELETYPE CONNECTED FROM THEIR CLASSROOM TO A STANFORD UNIVERSITY COMPUTER. LITTLE TEACHER PREPARATION WAS REQUIRED. WITH IMMEDIATE FEEDBACK AND A 10-SECOND TIME LIMIT PER RESPONSE, EACH STUDENT HAD 2 CHANCES TO ANSWER CORRECTLY. RANKING THE LESSONS BY PROPORTION OF CORRECT RESPONSES AND CLASSIFYING THEM BY TYPE OF PROBLEM SHOWED THAT LESSONS OF THE SAME TYPE TEND TO BE GROUPED IN ADJACENT RANKS. AMOUNT OF PREVIOUS PRACTICE, AND NUMBER OF PUPILS TAKING EACH LESSON CONFOUND THIS RELATIONSHIP, HOWEVER. THE RANK-ORDER CORRELATION BETWEEN THE AVERAGE PROPORTION OF ERRORS AND THE TIME TAKEN TO COMPLETE A LESSON WAS STATISTICALLY SIGNIFICANT. A COMPARISON OF 5 INDIVIDUAL STUDENTS ALSO SHOWS THIS RELATIONSHIP. DATA REPORTED HERE WERE INTENDED TO GIVE ONLY A SENSE OF METHODS FOR MORE RIGOROUS CLASSROOM EXPERIMENTATION. FUTURE EXPANSION OF THE PROJECT CALLS FOR 80 STUDENTS USING EACH OF 3 TELETYPES EVERY DAY, MORE EXTENSIVE REPORTS TO STUDENTS AND TEACHERS, PROGRAM BRANCHING CAPABILITIES. (LH)