A mathematical analysis was made of the effect of misalinement among the components of rectangular and equiangular strain rosettes on the determination of the magnitudes and directions of principal strains. Misalinements of or - 2 deg among individual gages introduce no serious errors in the computation of principal strains from rosette data. Errors caused by a given misalinement are proportional to the maximum shear strain at the test point and the effect is most important in regions of pure shear. The equations derived in the analytical investigation provide corrections that are applicable only if the mislinements are known. These equations indicate qualitatively the limits of accuracy with which strain determinations can be made when the possibility of misalinement exists. Verification of the analysis was made by experimental investigation of the behavior of a rosette the misalinement of which was known.