Subjects were exposed to a three-dimensional model of a house and were asked to draw it using a raised-line drawing kit. Independent groups of ten each of sighted controls, early-blind, and late-blind subjects were told to identify the vantage point of tangible pictures of the model, including side views, ‘bird's-eye’ views from above, and views involving linear perspective. The ease or difficulty of picture interpretation depended upon the nature of the tangible drawing, with much better performance being recorded for side views. Performance was poor for foreshortened 3/4 views. Early-blind subjects were particularly unlikely to recognize views from above. In a control experiment with blindfolded sighted subjects the influence of prior information was examined: some subjects were told that the drawings could consist of side view or bird's-eye, top view, or 3/4 view drawings. This experiment showed that performance can be greatly improved through prior information about the nature of the tangible pictures.