We think that the decree appealed from should be affirmed in so far as it admits to probate the original will made by the deceased, and should be reversed in so far as it admits to probate the two alleged codicils thereto.
Jonathan M. Dake died in February, 1900, being a resident of •Livingston county. He was over eighty years old at death and seems to have developed a quite well-defined predilection for writing wills and codicils and other communications to be published post mortem. He had been married four times. He was survived by his last wife and one child of that union, and also by descendants by each of his previous marriages. At and for some time prior to his death he resided with his last wife and child and apart from his children by former marriages. Up to a short time before his death he had been possessed of considerable property. After his death it appeared that he had executed various purported transfers of a large portion of the same to his widow. Shortly after his death said widow proposed for probate by the surrogate of Livingston county a purported last will and testament dated February 5, 1894, and also two purported codicils thereto dated respectively February 20, 1898, and November J, 1898. The son by her marriage with the deceased joined in the efforts to secure probate of these instruments. The children and grandchildren of the deceased by his former marriages opposed the same. Originally various reasons, including lack of testamentary capacity and undue influence, were urged by contestants why said purported last will and testament and codicils, which favored the widow and her son, should not be admitted to probate. Upon the hearing, however, and upon this appeal the objections were and are in substance reduced to the following:
*4061. That said original will contained material interlineations and alterations which avoided it.
2. That after its execution, and in 1896, another will was executed by the testator which operated as a revocation and annulment of the will proposed.
3. That the first proposed codicil was void and ineffectual because of failure upon the part of the testator to properly and sufficiently observe the statute governing the execution of wills (2 R. S. 63, § 40.)
4. That said second purported codicil was not in any proper and sufficient way so connected with or made applicable to the original will as to be admitted as a codicil thereto.
We shall consider the objections in the order stated.-
The original will did contain two or more material alterations by way of interlineation. These interlineations, however, like the body of the will are in the handwriting of the testator. An attempt ■ was made by the contestants to demonstrate by the evidence of an expert witness that the ink of the clauses inserted was so different from the ink used in the body of the will as to indicate that the alterations were made at a different and subsequent time from that at which the will itself was written. The proponents, upon the other hand, gave evidence tending to show that the body of the will and the inserted clauses were all written at the same time. The surrogate, both by his findings and by his refusals to find, by implication at least, adopted the theory of the proponents and refused to accept that of the contestants that the said interlineations were made after the will had been executed. In view of his findings and refusals to find, and under all of the circumstances of this particular case, we are unwilling to hold or presume that said changes were made in said will after the same had been executed. We think, upon the other hand, that it is more reasonable to assume that the same were made, by the testator in the course of the preparation of the instrument and before its execution. (Matter of Potter, 33 N. Y. St. Repr. 936; Crossman v. Crossman, 95 N. Y. 145.)
In support of their claim that the proposed will had been revoked by the acts of the testator, contestants gave evidence tending to prove that he.had made another and later will some time in the *407latter part of the year 1896. The evidence as to the date and details of the execution of this alleged later will was given by the subscribing witnesses thereto. It was not entirely clear and definite in all respects, especially as to the time of execution. It may probably he assumed with safety, however, that such will was made. The acting surrogate, while refusing to find as fully as requested by the contestants upon this subject, has still found somewhat in favor of their claim. This later will was not produced upon the hearing and no evidence whatever was given as to its contents, except that the witnesses stated that when the testator asked them to witness the will he stated in substance that it revoked his other will or wills.
Assuming that execution of this last will was proved, we do not think that all of the evidence bearing upon that subject is sufficient to sustain a finding that thereby the proposed will was revoked.
The statute bearing upon that subject (2 R. S. 64, § 42) provides: “Ho will in writing except in the cases herein after mentioned nor any part thereof shall be revoked or altered otherwise than by some other will in writing or some other writing of the testator declaring such revocation or alteration, and executed with the same formalities with which the will itself was required bylaw to-be executed,” etc.
It may also be accepted that a former will may be revoked or annulled by a subsequent one which is thoroughly and radically contradictory of or inconsistent with the provisions of the first one.
Both of these rules, impose upon the party seeking to establish such a revocation of a former will the burden of showing that the latter instrument, either by its express terms or by necessary implication, operated as a revocation. That evidénce is wanting in this case. The testimony of the witnesses Gurnee and O’Oonnell that the testator declared that the will of 1896 revoked his former wills, even if it had been accepted and believed by the acting surrogate, did not furnish sufficient evidence of the contents of .the later will to comply with the requirements of the rules adverted to, Even if it should be held that the statement of the testator as to the contents of his will was proper proof, we should hold that the declaration in this case claimed to have been made was nothing more than the expression by the testator of his opinion as to the legal effect of his *408later act. (1 Jarm. Wills [Rand & Tal. 5th Arn. ed.], 338; Matter of Williams, 34 Misc. Rep. 748.)
We pass finally to the consideration of the two purported codicils to the foregoing will, and in the admission of which to probate the learned acting surrogate in our judgment committed error.
These two alleged instruments, as before stated, are dated respectively February 20, 1898, and November 7, 1898. They are both, in the handwriting of the testator. As presented to the court, they were-part, of a parcel of papers pasted together, and consisting, in the order named, of a communication covering two sheets of foolscap addressed by the testator apparently to his son Moses; the later of the purported codicils covering one page of foolscap; the earlier purported codicil written upon two pages of distinct sheets of fools- . cap; the will in question, upon which are found certain detached declarations in the handwriting of the testator in substance that he had been free from any undue influence, etc.
It is urged against the codicil of February twentieth that. it is invalid because it does not comply with those provisions of the statr ute which require a will to be subscribed by the testator at the end thereof. This claim is substantially based upon the fact that as the codicil was presented in the bunch of papers above-described, the final sheet containing the signature of the testator and the witnessing clause comes first, while the alleged other and first sheet containing material provisions comes last. If we were to assume that the tes-, tator designedly and in.some effective way executed this purported! • codicil in the form -now presented, with his signature at.the commencement instead of at the end thereof, it would be a violation of the requirements of the statute. Upon the other hand, without holding it, we may grant for the purposes of this case that if a will were presented which beyond question had been executed in a valid manner and in which by inadvertence the final sheet containing the testator’s signature had been transposed to some other place in the order of pages, the court would find it possible to avoid the results of this accident.
" The objection to this proposed codicil involves more than a mere inadvertent displacement of pages. The question in our opinion presented is whether, upon the facts appearing, we should say that the two sheets of paper were legally executed as one codicil, and *409then, by transposing their position, bring them within the provisions of the statute and give them the indorsement of regularity and validity.
It will be well to consider briefly some of the facts which bear most potently upon the answer to be given to this question.
In favor of an affirmative answer thereto we have the evidence, somewhat indefinite to be sure, that a codicil or will was executed by the testator at the date of this alleged one; that both pages are in the handwriting of the testator, and that there is nothing in the substance of the provisions thereof which prevents' them from being-read together with intelligible, consistent results; also that the pages are numbered in such order as conforms with proponent’s theory, and that by or in behalf of the parties interested in proving the same, it is testified that they were found in the order presented to the court in a private box of the testator'after his death.
As against these considerations it is settled that the mere numbering of pages will not necessarily overcome the other objections. (Matter of Andrews, 162 N. Y. 1.)
There is nothing in the handwriting or physical appearance of the two sheets which indicates with any degree of certainty whatever that they were written at the same time or that they belong together or that taken together they constitute an entire instrument. All of the documents presented to us, although composed at different dates, are in the handwriting of the testator and, therefore, the fact that these two sheets are in his handwriting does not tend to prove that they were prepared at the same time. There is nothing in the condition of the paper or the ink which throws any light upon this subject. An inspection readily discloses that either sheet in the respect specified bears as close a resemblance to the main will executed years before as it does to its alleged companion sheet. While, as stated, there is nothing repugnant in the provisions upon the two sheets, there is, upon the other hand, no such connection in the contents thereof as binds the two together. There is no express reference or statement in either sheet which indicates that it belongs to the other. Upon the contrary, the only statement upon this subject is repellant of such idea or conclusion. In the attestation clause found upon the first sheet in order of physical arrangement is the expression “ the above instrument containing one sheet *410at- the date hereof signed and sealed,” etc. In addition, there is. a lack of connection between the final words upon what we are asked to regardas the first .part and the first words upon the alleged second page, which might be readily cured by an intervening sheet which has now been dropped out. Outside of the present attachment by mucilage, it appears by inspection that the two sheets are entirely distinct, separate and disconnected.
It may be conceded, as strenuously argued by the respondents, that there is no law which fixes the number of sheets upon which a will may be Written or which requires that at the time- of execution they be fastened together in any particular way or in any manner. It must, however, be regarded as well settled that if a testator desires to execute as a will or codicil several sheets of paper they must be all present at the time of execution. He cannot have a valid execution based upon the presence of one or two sheets which are thereafter taken elsewhere and attached to others. Heither if he legally executes several sheets as a complete and single instrument can he fail to destroy the validity of the whole if some of the sheets are subsequently dropped out and the remaining ones attached together in a purported entire instrument.
We have looked in vain through the testimony of the subscribing witnesses to this purported codicil and elsewhere for any evidence which should adequately support the burden imposed upon proponents of establishing by a fair preponderance of reliable evidence that these two sheets of paper were executed together as the entire component parts of one and the same instrument. As against a lack of desirable evidence upon this point and as against a lack of connection and' consistency between the two sheets, we are compelled, if at all, to base a decision that they were executed together as a complete and entire codicil very largely upon the mere fact that they are produced in court years after the date of the alleged execution, pasted together in. an unnatural order and as part of a bundle of miscellaneous documents. We are unwilling to do this. It is possible that they were exécuted as claimed and that the will of the testator is being defeated. ■ We think it better, however, that the purpose should fail of one who carelessly or ignorantly has omitted to comply with the very simple provisions covering the execution of wills, than that the door should be opened through which *411may enter conjecture and speculation and uncertainty and fraud in other cases to thwart the care and pervert the intentions of those who have with intelligence and pains observed the law upon this subject.
There is no difficulty in executing a will in such physical form and shape as will leave no doubt where it begins and ends and what constitutes it. The court should not be required to select from a medley of papers sheets which may or may not be related to each other, transpose and rearrange them and finally construct for the testator a valid instrument. If a testator chooses to execute his will upon loose, detached sheets, before probate is allowed either the sheets themselves by their contents or other manifest relationship, or some other competent .proof, should adequately establish that they were executed as, and constitute, a single, entire will.
It is urged against the validity of the purported codicil of November 7, 1898, that there is nothing in the codicil itself or in the evidence outside of its mere physical attachment to the will in question, to show that it had relation to such will. This is true. So far as the codicil discloses upon its face it might have had relation to the will probated or to any one of the several others which the testator is said to have executed, at least one of them coming in point of time between this codicil and the will now proposed. There is nothing in the evidence of the subscribing witnesses which establishes this connection, and, therefore, in the case of this instrument, the question is presented whether a sheet of paper executed as a codicil, in the absence of all other evidence, is to be held to relate to and change a prior will simply because when presented to the court it is one of the same bundle of papers as said will. We do not think we should so hold. We think it is not going too far to hold that a purported codicil to a will, either by reference in the body thereof to such will or by other convincing evidence, must be shown to relate to the will as against which it is offered in probate.
The conclusions reached by us as heretofore expressed render it unnecessary to consider the various objections to evidence urged by appellants.
These views lead us to the conclusion that the decree 'appealed from should be affirmed in so far as it admits to probate the will, of February 5, 1894, and should be reversed in so far as it admits to *412probate the two codicils of February 20, 1898, and November 7, 1898, respectively, It is possible that upon a new hearing as to saicl codicils further evidence may be produced bearing, upon, the objections now urged to their probate.
McLennan and Williams, JJ., concurred; Adams, P. J., not voting, and Spring, J., dissenting as to the reversal of that part' of the decree which admits the codicils to probate.
Decree of Surrogate’s Court affirmed in so far as it admits to probate the last .will and testament dated February 5, 1894, and reversed in so far as it admits to probate the two purported codicils dated respectively February 20, 1898, and Noveinber 7, 1898, and ■ the proceeding remitted to the Surrogate’s Court of Livingston county, for a rehearing as to such codicils and such further proceedings as may be necessary and proper,