幼兒在Hood（1995）的彎管作業，易根據直線（而非彎管形狀）判斷物體落點，此現象稱為重力偏誤。實驗一以三種不同入射角入口（0°、45°及90°）修訂Hood的彎管作業，探討起始運動方向與投球方式對幼兒判斷落點的影響。在所有角度組別中，幼兒於前半段嘗試次觀看實驗者投球，接著找尋球落下的位置，後半段嘗試次由幼兒自己投球並找球。結果發現：幼兒的錯誤最易發生於投入口垂直正下方的位置，特別當球以90°入射角進入彎管時，作為投球者比作為旁觀者更易產生重力偏誤。實驗二以自行投球先於實驗者投球的順序重複90°入射角入口，結果發現作為投球者與旁觀者產生的重力偏誤沒有顯著差異。本研究結果顯示作業物件提供的知覺線索、投球方式的資源損耗及順序排程等，皆可能影響幼兒抑制物體垂直落下的重力期待。 When a ball drops down into one curved tube, preschoolers tend to search for the ball in the location directly below the entrance to the tube, even if they are not connected to each other. The search error has been interpreted as a gravity bias (Hood, 1995). Based on Hood's tubes task, Experiment 1 investigated whether initial direction of movement and drop-ball procedure influenced the gravity bias. Children observed a ball drop down the designated tube in one of three conditions that differed in entrance angle to the tube (0°, 45°, or 90°). They observed an experimenter drop the ball in the first block of trials, and then dropped the ball themselves in the second block. The results showed that most children failed to find the ball and tended to search the vertical aligned box. The gravity bias was worse when children dropped the ball themselves in the 90° condition. In Experiment 2, we reversed the block sequence and found that the block difference was not observed as previously shown in the 90° condition. Thus, the expectation that objects fall in a straight line seems to be modulated not only by perceptual information about the task but also by cognitive resources recruited by the drop-ball procedures.