Abstract: Four experimental conditions including orhographic problem solving followed by isometric and orthographic recognition, isometric problem solving followed by recognition of isometric views and orthographic recognition, were used to study how performances on spatial problem solving and recognition were influenced by the types of projection of visual objects used in each task. Besides the fact that had further supported the claim that orthographic problem solving was accomplished by constructing a mental representation of three-dimensional object and ruled out the alterative explanation of adequacy of encoding of two-dimensional projection of objects, the results also indicated: ( 1) Mental representation of three-dimensional object was constructed in the course of problem solving based on two-dimensional information, rather than in recognition of isometric view of object; (2) The complexity of orthographic problem solving was significantly higher than that of isometric problem solving; furthermore, a constructed mental representation was advanatageous only for the sequent recognition of isometric views of three-dimensional objects, but not for orthographic recognition; (3) The constructed mental representation had the properties that made spatial information more easily organized and stored and less fragile than two-dimensional infromational representation. The findings would be of significance not only to further study on the characters of spatial representation in problem solving and recognition, but also in finding further evidence concerning implicit memory for object structure.

Key words: projection view, solving spatial problem, recognition, construction