Pullout tests were performed to examine the top bar effect and the anchorage strength of D10 reinforcing bars used as horizontal reinforcement in walls. Test variables include the embedment length of the horizontal bars and the vertical height from the surface of hardened the slab. To simulate an actual construction situation, a wall of 3 m in height was divided into three regions with 1 m height per each region and concrete of each region was poured at 1 hour intervals. Twenty-seven bars out of thirty-two bars with the embedment length of 0.4 and 0.6, where means the development length calculated by KCI code, failed by pullout and the remaining five bars and ten bars with embedment length of 1.0 failed by bar fracture or the tests were terminated without failure due to the safety. For the bars showing pullout, no significant correlation was observed between the anchorage strength and the bar location. Measured bar stresses were compared to the predictions by KCI code, Orangun et al.'s equation, and ACI 408 committee equation. The averages of [test]/[prediction] ratios were from 1.39 to 1.73 and the minimum values of the ratios were from 1.05 to 1.37, which implies that anchorage strengths of horizontal bars in walls are much higher than those of longitudinal bars in flexural members. Top bar effect of small-diameter bars is less than that of large-diameter bars because bleeding water can be easily removed by vibrating compaction. In addition, there is no flexural deformation in walls and, therefore, the horizontal bars in walls have higher anchorage strengths than the longitudinal bars in flexural members.