Previous research into children’s understanding of line of sight has led to differing conclusions as to when and how children become able to appreciate that their view of an object will be different from another persons’ view of the same object. This is probably due to the diversity of response methods required from the children as well as different types of tasks and settings being used between the experiments. The aim of the present thesis is to investigate systematically how children will fare across various settings and whether their comprehension of line of sight can be biased by the task’s setting. The first experiment assessed children’s understanding of line of sight through a tube that was bent to varying degrees of curvature and whether their response pattern would change when feedback was provided. Results showed that children have great difficulty performing correctly on this task, especially when the degree of curvature is small. The older children corrected their response pattern when feedback was provided but the younger children tended to persevere in their response pattern regardless of contradictory feedback. The second experiment looked at children’s performance when walls were used – half the walls were smooth gradual curves while the other half was walls made up of two segments that met to form an angle. Again the children were asked to predict if two dolls placed at opposite ends of each wall would be able to see each other. Results showed that though even young children have no trouble in performing correctly on the “angled” walls, performance on the curved walls was significantly poorer with the older children performing better than the younger children. The third experiment sought to quantify the point at which children deemed line of sight became possible. To do this we used a single “U” shaped trench with the children being asked if one doll could see another in various configurations. The results showed a strong bias towards over estimating visibility. The fourth experiment repeated the second experiment but used wooden trenches instead of walls but also sought to quantify the “switchover” point at which the children deem vision becomes possible between the two dolls. The difference between angles and curves was once again replicated as was the age difference. The fifth experiment compared children’s appreciation of line of sight through/along tubes, trenches and walls. This performance level varied strongly depending on the type of task the child was asked to perform upon with the tube proving to be the most difficult and the angled trench the easiest. The overall findings of the experiment pointed to a context-dependent performance, implying a piece-meal development of childrens’ comprehension of line of sight.