delivered the opinion of the Court:

At the time of the execution of the various instruments in question, the statute of Maryland relating to the execution of devises of real estate was in force in this District, and provided: “All devises and bequests of any lands or tenements devisable hy law shall be in writing and signed by the party so devising the same, or by some other person in his presence and by his express directions, and shall be attested and subscribed in the presence of said devisor by three or four credible witnesses, or •else they shall be utterly void and of no effect.” Thompson’s Digest, sec. 4, 345. This provision of the Maryland statute is •substantially the same as sec. 5 of chap. 3, 29 Charles II., known as the statute of frauds.

The sole question here presented is whether or not a will attested by two witnesses, followed by two codicils making specific reference to the will and attached thereto, attested by a third and a fourth witness, are, when taken together, sufficient to constitute a will attested by three witnesses, and can be construed as a revocation of a former will containing different devises of the same property. It is well settled, both by the English and American decisions, that the subscribing witnesses to a will need mot know the contents of the document; they may attest it without the presence of each other; they, or any of them, need not see *417the testator sign the will, provided he acknowledge the signature to each of the witnesses; and they need not even know that the document they 'have witnessed is a will. This rule as to what may constitute a valid attestation is substantially as laid down by Lord Mansfield in Windham v. Chetwynd, 1 Burr. 421, and it has been followed and approved by the modern decisions.

The ancient English case of Lee v. Libb, 1 Shower, 69, is decisive of the exact point here under consideration. In that case the will was executed in 1678, two years after the passage of the statute of frauds. The will was attested by two witnesses, who subscribed their names in the presence of the testator. One year later the testator made a codicil changing in some minor respects the devises made in the will, and expressly confirming the will. The codicil was attested by two witnesses, one of these being a witness to the will.

The question presented was Avhether the third witness to the codicil could be regarded as a witness to the will. The court held that there were two distinct writings, and there were not three witnesses to either. This case has been approved and followed since, both in England and the United States. It is directly in point. Here, there are three distinct instruments, no one of which is attested by three witnesses.

Another case in almost all particulars the same as the one at bar is Dunlap v. Dunlap, 4 Desauss. Eq. 305. There the will was attested by two witnesses. The draftsman was present, but did not sign as a witness. A codicil was executed in the presence of the same draftsman and two attesting witnesses, one of whom had not attested the will. The codicil, as in the present case, referred to the execution of the will. The draftsman testified that he wrote both the will and the codicil at the request of the testator; that he saw both signed by the testator; that he saw the witnesses attest both instruments in the presence of the testator and in the presence of each other; and that he attached the codicil to the will in the presence of the testator. Upon this state of facts, the court, referring to the draftsman, said: “Mr. Taylor was present and saw the execution of the will, and might *418have become a subscribing and attesting witness under tbe statute ; but tbat, in point of fact, he did not become so; be did not subscribe tbe will at all. Now tbe statute is so clear and peremptory on tbis point tbat it is impossible to get over it. Nay, if Mr. Taylor bad, after the execution of tbe will in bis presence, actually subscribed tbe will as a witness, but in another room, and out of tbe presence of the testator, tbis subscription would not be sufficient. Tbe will, so imperfectly executed, cannot pass real estate. Such is tbe law, and I must be governed by it.”

In tbis jurisdiction, in tbe case of Re Porter, 9 Mackey, 503, tbe court, considering a codicil signed by two witnesses wbo saw tbe testator sign it, and by a servant wbo did not see it signed or know what tbe document was, but signed it by command of tbe testator, said: “We accept as authoritative tbe line of cases which we have cited. They show tbat, within tbe meaning of tbe statute of 29 Charles II., attestation is properly made, either when tbe witness has seen tbe testator sign tbe instrument, or has beard him after it has been signed, acknowledge that the act was bis. They show also tbat tbis acknowledgment may be conveyed either by words or by acts which indicate to tbe witness tbat be is called upon to act in tbe capacity of witness and to fittest something which is treated by the testator as bis act. They show, too, tbat it is not necessary to tbe effectiveness of tbis acknowledgment, or of tbe attestation, tbat tbe witness should know tbe nature of the instrument produced and submitted for bis attestation.”

Counsel for defendants in bis brief has displayed a patient and able research, and cited many cases showing tbe great length to which courts will go in upholding tbe execution of a will in order to carry into effect tbe evident intention of tbe testator, but be has produced no approved case where tbe statutory number of witnesses has been supplied, either by considering a person present at tbe signing, but not formally attesting, a witness, or by considering each of tbe several witnesses to a will and its codicil or codicils as witnesses to tbe entire instrument. It is well settled, both in England and tbe United States, tbat a re-*419execution of a will requires the performance of the same formalities as are necessary in its original execution; and a will revoking a former will, or a codicil changing the terms of the will to which it is attached or refers, that is not attested by the requisite number of witnesses, is void. 1 Underhill, Wills, sec. 211.

In the present , case it is sought to furnish the requisite number of witnesses to the will by parol evidence of the presence of Moorehead when the will was originally executed, or by having his attestation of the codicils relate back to the attestation of the will. It is unnecessary for us to consider this contention, since the codicils, being illegally executed for lack of the requisite number of attesting witnesses, cannot be relied upon as a republication to make either of them a part of the will. St. John’s Parish v. Bostwick, 8 App. D. C. 452. The will and the two codicils, each being defectively executed, it is not clear to us how they can be so put together as to be construed into a legal instrument for the conveyance of real estate. To accomplish this, it would not only be necessary to make the attestation of Moorehead to the codicils relate back to the attestation of the will, but, in order to have the requisite number of witnesses to the codicils, it would be necessary to make the two attesting witnesses to the will relate forward to the first codicil, and one of them relate forward to the second codicil, though neither of them, presumably, ever saw the codicils, or even knew of their execution by the testator. The mere statement of the proposition is sufficient answer to defendants’ contention.

We have been cited-to no ease where a court has assumed to supply an attesting witness by parol evidence. Such a ruling would open wide the door to all sorts of abuses. It would be an invitation to persons having an imaginary interest in the estate of the testator to cure a defective will by fraud and perjury. A valid attestation can only exist when the signatures of the number of witnesses required by statute appear attached to the instrument in such a manner as to indicate that they were placed there with the knowledge ánd consent of the testator, cither at *420his request or by his command. The only case called to our attention where a different holding seems to have been made appears in Bacon’s Abridgment, Vol. 1, marginal page 309, referring to Lea v. Libb, 3 Mod. 262, where it is stated that it was “adjudged, though it was objected, the will and codicil make but one will, and the circumstance of three witnesses wanting to complete the will was perfected by the codicil.” This case is purported to have been decided in 1688, twelve years after the enactment of the statute of frauds. If the reference is authentic, the reasons for the judgment have not been followed by the subsequent English decisions, and it must be presumed that they were overruled and discredited.

The will in question not having been legally executed, it cannot either operate as a conveyance of the real estate here in controversy, or a revocation of the former will. The judgment is affirmed, with costs, and it is so ordered. Affirmed.