Many nuclear physics experiments require cryogenic or highpressure gas targets with thin windows. To obtain the best data, i.e. to reduce background signals, a target window would ideally have zero thickness and zero deflection. Since this is not possible a trade off is made among target geometry, window geometry, window material, and window thickness. Measurements of deflections of a variety of window materials and geometries were performed at room temperature and in liquid nitrogen to aid in designing windows in the future. A simple graphical method for predicting the rupture pressure for several typical window materials is described. The results of the deflection measurement are compared to a finite element computer program MEMBRAN which predicts window deflection and material stresses for a uniform applied pressure.