The performer begins by filling a clean glass and the tannin glass with water from the pitcher. The clean glass he offers to the audience to taste, or sips of it himself. Then he pours the contents of both glasses into the pitcher.

Now if he fills the clean glasses and the iron glasses from the pitcher, they will appear to be filled alternately with water and wine. Pouring the contents back and immediately refilling the glasses, they are all filled with wine.

Before filling the twelfth glass, however, the performer picks up an acid glass, covering it with his hand so as to conceal its contents, and, filling it, remarks: "All wine. No, here is water, and here again"—filling the twelfth glass—"is wine."

When the contents of the glasses are poured into the pitcher this time, the action of the acid will bleach the solution completely, and on filling the glasses this time there appears to be nothing but water.

Again, however, the performer picks up another glass, the one with the ammonia, and filling this, it appears to contain wine, but of a lighter color than before.

This time, on emptying the glasses into the pitcher, it will appear to be filled again with wine, the ammonia counteracting the effect of the acid.

Once more the glasses are filled with wine, all but one, the second acid glass, which the performer has filled.

For the last time the pitcher is again filled, and when the contents are poured out there is nothing but water. "For," as the performer remarks, "having begun with water, it is only right we should end with it."

As the strength of the various ingredients varies a great deal, the amateur will do well to experiment for the proper proportions of tannin, acid, and ammonia before attempting to exhibit the trick, bearing in mind that the smaller the quantity used the better it will be.

Mr. Kellar, who is rather an exhibitor of stage illusions than a sleight-of-hand performer, since his "feats of prestidigitation," to quote the language of the show-bill, are of the most simple character, has made quite a name by offering what he is pleased to call "mental phenomena." In one of the most surprising of these he almost instantly tells the day of the week on which any date of the present century falls.