In some spirit photographs produced by double exposure there is a double background, as occurred in my own photograph referred to above. There may be either two different backgrounds, or a double outline of the same background; in either case the “spirit’s background” is usually fainter than the “sitter’s background,” and shows through the darker parts of the sitter. Sometimes attempts are made to retouch these appearances on the negative, and many spirit photographs show clumsy brush or pencil work, which must immediately stamp them as frauds.
Attempts are sometimes made to obliterate other tell-tale marks, such as a piece of a spirit’s hat or collar, which has accidentally got on to the plate. Other mediums, however, are less particular, especially in America, and produce their spirits with ordinary hats, collars and ties. But as a rule only spirit robes are permitted, apparently made of butter muslin not quite in focus. Hands are often present: I have seen a case in which the position of a spirit hand would have necessitated a many-jointed arm about four feet long; but perhaps spirit arms _are_ like this. One spirit extra I have seen has two hands, but both appear to be left hands--evidently a left-handed spirit.
Frequently, again, careful examination shows that spirit extras are not photographs at all, but resemble wash drawings. This gives the clue to their origin, for several of the methods described in a preceding section produce a result of this kind. It has been several times pointed out that spirit extras in some cases show the characteristic dots produced by the half-tone newspaper illustration process; if the medium cannot obtain a real photograph of the required spirit, he has to copy a newspaper reproduction. If he is clever, he can eliminate these process marks by printing in his spirit slightly out of focus; but very often he does not take the trouble.
In many, perhaps in the majority, of spirit photographs produced by professional or semi-professional mediums, a critical observer with practical photographic experience can point out some such definite evidence of fraudulent manipulation. In many other cases, where no one particular point can be singled out as indicative of fraud, minor points of suspicion are noticeable, which taken together leave little doubt of the nature of the picture. But photographs _can_ be prepared by purely mechanical means, especially if no kind of test conditions are employed, which will contain no internal evidence whatever of manipulation. By carefully combining enlarged positives, for instance, and re-photographing the whole, results can be produced which will defy the most critical examination. But such photographs are seldom produced, even when the medium is given practically a free hand.