Procure a substantial-looking muzzle-loading pistol, the larger the better, and get a tinman to make a tube that fits neither tightly nor loosely inside. One end of the tube must be closed, and the open end be furnished with a turn-over rim, which you colour or polish, so as to resemble as nearly as possible the muzzle of the pistol. If it comes easier, by all means reverse the process, and make the muzzle of the pistol resemble the muzzle of the tube. The tube should be at least two-thirds the length of the interior of the pistol barrel, and not about an inch and a half only. Now make a wooden ramrod, which will fit rather tightly into the tube—tightly enough to bring it away from the interior of the pistol when withdrawn after being rammed into it. The exterior of the tube should match the ramrod in colour and appearance, so that the fact of the one being within the other when held in the hand in that condition will not be detected.
Advance to the audience with the pistol, ramrod, powder, caps, paper, and bullets. The tube is concealed somewhere about you. Give round one or more bullets to be marked by the audience, and, at the same time, give the pistol into the hands of someone for examination. If the pistol takes to pieces, so much the better, as it is highly advantageous for the audience to be quite convinced that there is no mechanical preparation of the performer's own devising connected with the implement. Have a fair charge of powder put in, and give the ramrod for ramming the charge. Whilst this is being done, get the tube into the hand, and, when the ramrod is returned to you, pass it into the tube, and then at once say earnestly, "I hardly think you rammed it sufficiently, sir." Without taking the pistol yourself, hold the muzzle in the fingers of the left hand, and ram with the right. The left hand must pinch the rim of the tube, which will thereby become disengaged from the ramrod, and remain inside the pistol. By this means the tube has become inserted in the pistol without any occasion for the pistol leaving the possession of the person assisting. The bullet or bullets are then dropped in by such of the audience as have marked them, the pistol being handed round by the volunteer assistant, who is also asked to put a piece of paper in as well. This piece of paper is very small, and is rolled up by the performer himself, who will, of course, take particular care that it is not bulky enough to jam in the tube. The performer now, without taking the pistol, puts in the ramrod and presses down the bullets—apparently. He, however, takes good care to avoid pressing them at all, but brings the hand holding the ramrod over the muzzle of the pistol, and so secures the tube. There is nothing in the action to excite the least suspicion; still, it is best to have the ramrod of such length that only an inch or so of it protrudes from the muzzle. If this particular be observed, then the hand must approach the muzzle on every occasion of ramming down a charge, and so no suspicion can possibly be excited by the action. The imaginary ramming is, of course, continued, the tube being moved up and down inside the pistol vehemently. Withdraw the ramrod and tube, carefully concealing the junction of the two, and have the pistol capped. Having secured the bullets in the tube, the next thing is to get them into the hand. As it is impossible to withdraw the rod from the tube in full view of the audience (it can certainly be done behind the table; but the unavoidably lengthened absence of both hands for the purpose would be fatally suspicious), the performer must necessarily retire behind the scenes. The best excuse is that of requiring a plate. Whilst fetching this, the rod is withdrawn from the tube, and the bullet or bullets shaken out. The paper ought not to stick, but it might; and so it is always as well to have a piece of wire with a sharp hook at one end in readiness. This abstraction of the bullets could be managed whilst the performer retires up the stage, and it would be advisable to do so if the bullets were certain to drop out, which, however, they are not. When the performer has a stage assistant, as he always should have when possible, the matter becomes much easier. The assistant holds the box of caps on or near the stage, and the performer goes to him for them, and gives him the wand with the tube on it in an offhand manner. The assistant then manages all the abstraction whilst the capping of the pistol is going on, and stands, just removed from sight, with the plate in one hand and the bullets in the other. The performer then has only to stretch his arm behind to receive the bullets first and the plate next, both in one hand. The bullets are held under the fingers, which are on the inside of the plate. The performer now holds the plate either in front of his face or at arm's length, and requests that the pistol may be fired at it. When this is done, the bullets are allowed to roll on the plate, which is brought simultaneously into a horizontal position, a kind of "grab" being made with it, as if catching the bullets in the air. I like this method better than catching the bullets in the mouth, as the performer can at once run forward with them on the plate, and the audience will be thus enabled to see for themselves that no substitution has taken place during transit.
The difficulty of extracting the bullet or bullets I finally mastered, after various trials, by using a tolerably thick piece of cork for the purpose of closing the one end of the tube. When the performer retires up the stage, he holds the closed end in his hand and presses the ramrod against his body, thus forcing out cork, bullet, and paper with one vigorous push. The operation does not occupy a second, and, when the performer turns facing the audience again, the ramrod is in his hand as before. It will naturally occur to everybody that, if the ramrod fits tightly into the tube, the whole arrangement will be neither more nor less than a popgun, and the cork will be blown out as soon as the ramrod is inserted into the tube. To avoid such a startling result, the ramrod must, although fitting the tube at certain places, be made out of truth, so as to admit of an escape of air, or else the cork must have a hole burnt through it by means of red-hot wire. Piercing a hole through the cork will not suffice, as the nature of the material will speedily cause the opening to close up again. If this cork arrangement be used, then the performer need never leave the stage.
I have given the learner what I consider to be the best method for performing the trick. A method which differs only slightly from mine, the principle being the same, is to use a tube barely two inches in length. This tube is dropped into the pistol by the performer, who takes the pistol entirely in his hands, for the purpose of ramming down the powder charge. As it lies secreted a long way down the pistol, it, of course, can never be seen. In pushing the ramrod down, it becomes fixed in the tube, and brings it away, the ramrod being the least bit tapered, to insure its going into the tube. It should be tapered at each end, so that there may be no bother about looking to see which end is the graduating one. The rest of the trick is performed exactly as in the other method. My objections to the use of the short tube are that, in order to get it into the pistol, the performer must secure actual possession of the firearm, and he must repeat the manoeuvre when he wishes to take it out again. Again, it is very difficult to conceal the point of junction of a short tube on a long ramrod. If the performer does not use the utmost caution, the tube will be noticed sticking on the rod as they are withdrawn from the pistol. The rim of the long tube is covered with the hand before the abstraction is made, and added to this is the fact that the pistol never leaves the hand of the volunteer assistant, which is a great feature in the trick. The only objection to the long tube I found to be the difficulty sometimes arising in abstracting the bullets; but my cork has now removed that difficulty. I have given both methods that the learner may choose for himself. Although I have said "bullet or bullets," because some conjurors employ one only and some three or four, I should, myself, never think of using more than one. The effect is the same, whilst the trick is made immeasurably easier to the performer. People may, perhaps, be a little more satisfied at seeing several bullets marked by different people, but it is just as easy, and quite as effective, to have one bullet thus treated. It will not matter in the least to the performer how many people mark it.
Another method for performing the trick is to substitute for the marked bullet a blacklead one. This, on account of its lightness, the performer must himself drop into the pistol and see that it is so rammed that it is broken up. I would never advise anyone to adopt this method, although Houdin caused consternation amongst the Arabs by allowing himself to be fired at from a distance of a few paces. With his usual completeness he finished the trick by firing another pistol at a whitewashed wall, on which appeared a large splash of blood. This was managed by means of a ball of black wax, the inside of which was filled with blood. The Arabs were duly impressed.
There are some pistols made with spring openings in the barrel, through which the bullet falls into the hand of the performer; but, in this case, the pistol cannot be examined, which fact is quite sufficient to taboo the method. Unless the pistol is given to be examined, and left in the hands of the spectators the whole time, there is, in my opinion, nothing in the trick.