.

is really very simple,” replied Thorndyke. “Let us consider what would happen in the two alternative cases. Take first that of the seal engraver. He has handed to him one or more of the wax impressions from the boxes and is asked to engrave a seal which shall be an exact copy of the seal which made the impressions. What does he do? If the wax impression were absolutely perfect, he would simply copy it in intaglio. But a seal impression never is perfect unless it is made with quite extraordinary care. But the wax impressions on the boxes were just ordinary impressions, hastily made with no attempt at precision, and almost certainly not a perfect one among them. The engraver, then, would not rigorously copy a particular impression, but, eliminating its individual and accidental imperfections, he would aim at producing a seal which should be a faithful copy of the original seal, without any imperfections at all.

      “Now take the case of the electrotype. This is a mechanical reproduction of a particular impression. Whatever accidental marks or imperfections there may be in that impression will be faithfully reproduced. In short, an engraved seal would be a copy of the original seal; an electrotype would be a copy of a particular impression of that seal.”

      Mr. Penfield nodded approvingly. “An excellent point and very clearly argued. But what is its bearing on the case?”

      “It is this: since an electrotype seal is a mechanical copy of a particular wax impression, including any accidental marks or imperfections in it, it follows that every impression made on wax with such a seal will exhibit the accidental marks or imperfections of the original wax impression, in addition to any defects of its own. So that, if a series of such impressions were examined, although each would probably have its own distinctive peculiarities, yet all of them would be found to agree in displaying the accidental marks or imperfections of the original impression.”

      “Yes, I see that,” said Mr. Penfield with a slightly interrogative inflexion.

      “Well, that is what I have found in the series of seal-impressions from Mr. Hollis’s boxes. They are of all degrees of badness, but in every one of the series two particular defects occur; which, as the series consists of over thirty impressions, is utterly outside the limits of probability.”

      “Might those imperfections not have been in the seal itself?

      “No. I took, with the most elaborate care, two impressions from the original seal, and those impressions are, I think, as perfect as is possible. At any rate, they are free from these, or any other visible defects. I will show them to you.”

      He took from a drawer a portfolio and an envelope. From the latter he produced one of the two impressions that he had made with Mr. Hollis’s seal and from the former a half-plate photograph.

      “Here,” he said, handing them to Mr. Penfield, “is one of the seal impressions taken by me, and here is a magnified photograph of it. You can see that every part of the design is perfectly clear and distinct and the background quite free from indentations. Keep that photograph for comparison with these others, which show a series of thirty-two impressions from the boxes, magnified four diameters. In every one of them you will find two defects. First the projecting fore-legs of the left-hand horse are blurred and faint; second, there is, just in front of the chariot and above the back of the near horse, a minute pit in the back ground. It is hardly visible to the naked eye in the wax impressions, but the photographs show it plainly. It was probably produced by a tiny bubble of air between the seal and the wax.

      “Now, neither of these defects is to be seen in Mr. Hollis’s seal. Either of them might have occurred accidentally in one or two impressions. But since they both occur in every case, whether the impressions are relatively good or bad, it is practically certain that they existed in the matrix or seal with which the impressions were made. And this conclusion is confirmed by the fact that, in some cases, the defect in the horse’s fore legs is inconsistent with other defects in the same impression.”

      “How inconsistent?” Mr. Penfield demanded.

      “I mean that the faint impression of the horse’s legs is due to insufficient pressure of the left side of the seal; the seal has not been put down quite vertically. But here—in number 23, for instance—the impression of the chariot and driver on the right-hand side is quite faint and shallow. In that case, the left-hand side of the impression should have been deep and distinct. But both sides are faint, whereas the middle is deep.”

      “Might not the seal have been rocked from side to side?”

      “No, that would not explain the appearances; for if the seal were rocked from side to side, both sides would be deep, though the middle might be shallow. It is impossible to imagine any kind of pressure which would give an impression shallow on both sides and deep in the middle. The only possible explanation is that the matrix, itself, was shallow on one side.”

      Mr. Penfield reflected, helping his cogitations with a pinch of snuff.

      “Yes,” he agreed. “Incredible as the thing appears, I think you have made out your case. But doesn’t it strike you as rather odd that this ingenious rascal should not have taken more care to secure a good impression from which to make his false seal?”

      “I imagine that he had no choice,” replied Thorndyke. “On each box were six seals; three on the paper wrapping, two in the recesses by the keyhole, and one on the knot of the string. Now, as the paper had to be preserved, the seals could not be torn or cut from that. It would be impossible to get them out of the recesses. There remained only the seals on the knots. These were, of course, much the least perfect, though the string was little more than thread and the knots quite small. But they were the only ones that it was possible to remove, and our friend was lucky to have got as good an impression as he did.”

      Mr. Penfield nodded. “Yes,” said he, “you have an answer to every objection. By the way, if the paper had to be preserved so carefully, how do you suppose he got the parcels open? He would have had to break the seals.”

      “I think not. I assume that he melted the seals by holding a hot iron close to them and then gently opened the packets while the wax was soft.”

      Mr. Penfield chuckled. “Yes,” he admitted, “it is all very complete and consistent. And now to go on to the next point. You say that there is evidence that these boxes were opened by some person other than Hollis himself; a person connected in some way with Woodstock’s office. Further that they were opened, not in the office itself, but in some other place to which they had been taken. I should like to hear that evidence; especially if it should happen to be connected with those mysterious floor-sweepings.”

      “As a matter of fact, it is,” Thorndyke replied, with a smile. “But the floor-sweeping was not the first stage. The investigation began with Mr. Hollis’s boxes, from which I extracted every particle of dust that I could obtain; and this dust I examined minutely and exhaustively. The results were unexpectedly illuminating. For instance, from every one of the untouched boxes I obtained one or more moustache hairs.”

      “Really! But isn’t that very remarkable?”

      “Perhaps it is. But moustache hairs are shed very freely. If you look at the dust from a desk used by a man with a moustache, you will usually see in it quite a number of moustache hairs.”

      “I have not noticed that,” said Mr. Penfield, “having no moustache myself. And what else did you obtain by your curious researches?”

      “The other result was really very remarkable indeed. From every one of the boxes I obtained particles—in some cases only one or two, in others quite a number—of the very characteristic dust which is shed by worm-eaten furniture.”

      “Dear me!” exclaimed Mr. Penfield. “And you were actually able to identify it! Astonishing! Now, I suppose—you must excuse me,” he interpolated with an apologetic smile, “but I am walking in an enchanted land and am ready to expect and believe in any marvels—I suppose you were not able to infer the character of the piece of furniture?

      “Not with anything approaching certainty,” replied Thorndyke. “I formed certain opinions; but they are necessarily speculative, and we are dealing