The study's purpose was to investigate the differential effects on the achievement of second-grade students that could be attributed to three instructional sequences for the learning of the addition and subtraction algorithms. One sequence presented the addition algorithm first (AS), the second presented the subtraction algorithm first (SA), and the third was an integrated presentation of the two algorithms (I). Students were randomly assigned from IQ blocks to the three treatment groups. The performance of 49 subjects was measured prior to instruction, during instruction, immediately following six weeks of instruction, and two weeks after instruction; posttest and retention test included both timed and power test variables. Data were analyzed using multivaraite analysis of covariance. Findings showed: (1) the instructional sequence SA produced generally lower performance than either of the other two treatments; the tendency of all groups was to learn addition first, (2) the amount of learning and retention measured by the timed variables was significantly greater for group AS than for either of the other two groups, and (3) on the retention test, except for addition facts and addition without regrouping, the overall performance of group I for the timed variables was significantly greater than that of group SA. (DT)