Some misconceptions concerning Babinet's principle are pointed out. The Fraunhofer diffraction patterns of complementary screens are not similar unless the detail in the screens obeys certain conditions, which are not easily fulfilled. Some examples of the diffraction patterns of complementary screens are given, illustrating the extent to which Babinet's principle is valid in practice. A way in which the principle can be used to give the diffraction patterns of three-dimensional crystal-structure models is indicated.