A series of 394 women underwent 399 stereotactically guided preoperative hook-wire localization of non-palpable breast lesions. 250 lesions were localized without checking the position of the needle tip with a further stereo film. Later the position of the localization needle was checked by means of a further stereo film in 149 lesions. If the needle tip was beside the lesion (failed x- or y-coordinates), we took a new stereo film. Errors in needle depth (z-coordinate) produced particularly subtle signs. Therefore a polystyrene phantom and a metallic object was used to ascertain how to recognize needle depth error on the stereotactic film. In 51 lesions readjustment of the needle depth was necessary. Measured by our criteria, successful and acceptable localization procedures with or without a check stereo-image film showed no significant differences. Unsuccessful localizations without the check film were significantly higher (16.0%) than with the check film, without readjusting the needle depth (7.1%) or after readjusting the needle depth (9.8%). These unsuccessful localizations were mostly architectural distortions or microcalcifications. Our results suggest that success of the localization procedure is better by using the stereo check film as without it, even in lesions which are difficult to resolve and target on films. However, readjustment of the needle depth did not always lead to successful localization.