Single crystal films prepared by evaporation of silver, gold, copper, and nickel on heated cleaved rocksalt crystals and polycrystalline films of silver, gold, copper, platinum, aluminum, and nickel have been studied by transmission electron microscopy and electron diffraction. By means of electron microscope dark-field image and selected area diffraction technique the existence of small growth twins in the films is shown. Extra spots observed in the electron diffraction diagram can be explained partly as spots of the four twin orientations, partly as arising from double diffraction of the (100) matrix orientation and of a twin orientation. In the case of polycrystalline films the double diffraction mechanism results in additional diffraction rings seeming not to belong to the normal face-centred cubic lattice. The assumption of an hexagonal phase or of stacking faults or of periodic lattice defects cannot explain all the extra spots and additional rings. The changing diffraction contrast of thick and thin microtwins obviously depends on the different reflection conditions in the electron microscope. It is shown that it is very complicated to distinguish stacking faults from microtwins, but some criteria of distinction are given. By means of a heating stage the single crystals have been heated in the electron microscope up to more than 1000 °C. The thermal behaviour of the twins has been studied in detail. It can be observed that up to 95% of the twins become invisible by transformation from the twin orientation to the (100) matrix orientation. It is concluded from the experiments that the transformation nucleates at the top or bottom of the thin films. Nevertheless more than 1013 twins and stacking faults per cm3 remain unchanged inspite of heating to temperatures near the melting point of the bulk metal.