At the tria , the plaintiff proved the hand-writing of Cornelius Wyckoff, one of the subscribing witnesses, and his death; that another witness had been dead many years; and that the third went to England about 30 years ago, and, had not been heard of since. There was no evidence offered to prove the handwriting of the testator, or that of Patience Titus, or John *224Crawley. This evidence, per se, is not sufficient to prove the execution of the will. It is necessary, in every case, to establish the fact, that the testator executed in the presence of three witnesses; or, in cases where such proof does not exist, to give other evidence from which it may be presumed. This may be done by calling one or more of the witnesses, to prove the execution agreeably to the statute ; or, if the witnesses are dead, proof of their handwriting and that of the testator, are proper to be left to the jury, upon the question, whether, under such circumstances, it may not be presumed that the formalities of the statute were observed. (Phil. Ev. 383.) It is decided in Hands v. James, (Com. Rep. 531,) that when three witnesses have set their names, it must be intended they did it regularly. If, however, (the witnesses being all dead,) the hand-writing of one or two of them is established, it raises no other presumption, but that so far as related to their attestation, the will was well executed. They may have subscribed their names in the presence of the testator and the other witness; but the testator may have acknowledged his signing to the witnesses separately. It seems, there fore, to follow, that nothing short of proving the hand-wri ting of all will be sufficient. Proof of the hand-writing of one witness, raises no presumption that either of the others subscribed his name. Cases arising on other instruments, do not decide the case of a will. There it is enough, if the party executed; but here he must execute in a particular manner, or his will is utterly void.
The mero efflux of time has never, I apprehend, been adjudged sufficient to admit a will to be read without proof. The note to Doe v. Brabant, (4 T. R. 709,) is very short and unsatisfactory. It is said that the Master of the Rolls had decided that a will above 30 years old should be read without proof. I do not find that this doctrine is supported by the cases; and therefore think it can only be correct, when the antiquity of the instrument is connected with possession, or other circumstances. A doctrine repugnant to this, and resting merely on the existence of the instrument for more than 30 years, has never been ju *225didaily recognized in this state; and, if introduced, would be an experiment of very dangerous tendency.
The case of M’Kenire v. Fraser, (9 Tes. 5,) was relied on. Upon examination, it is far from being satisfactory. The hand-writing of two of the witnesses was proved; and no account could be given of the third. The will being above 30 years old, and the testator having been dead for twenty years, an objection was made to the proof of the will. The master of the rolls said, he could not see any distinction, in this respect, between a will and a deed, except that the former, not having effect till the death, wants a kind of authentication which the other has; that is, from the nature of the subject; “ but, in this case,” he added, “ I think the proof sufficient; for, in a late case, in the Court of King’s Bench, Cunliffe v. Sefton, (2 East, 183,) inquiry of the same kind was held sufficient.” The case referred to in 2 East, was an action on a bond, attested by two subscribing witnesses; and, after diligent inquiry, no account could be obtained of one of them. Proof was then received of the hand-writing of the other witness, one of the plaintiffs. If the rule applied, in the present case, there was this defect: here the witnesses were known; they are now dead ; no attempt was made to prove the hand-writing of two of the witnesses. There was then a want of diligence. Where a witness cannot be found, nor any account of him obtained, nothing more can be done. but when he is ascertained, and the reason of his non-production is satisfactory, then not only due diligence, but the rule that the best evidence that the case admits of shall be adduced, demands the proof of his hand-writing, if practicable.
The case of Jackson v. Laroway, (3 John. Cas. 283,) decides that a will concerning real property may, under certain circumstances, be given in evidence, as an ancient deed; and it is laid down as a general rule, that a deed appearing to be of the age of thirty years, may be given in evidence without proof of its execution, if possession be shown to have accompanied it; or where no possession has accompanied it, if such account be given of the deed as may be reasonably expected, under all the circumstan*226ces of the case ; and will afford the presumption that it is genuine. The rule is founded on the necessity of admitting other proof, as a substitute for the production of witnesses, who cannot be supposed any longer to exist. Where no possession appears, other circumstances are admitted to account for it; and raise a legal presumption in its favor. In that case, the admission of the will of Leonard Lewis was on the latter ground ; for there had been no actual possession under it by the plaintiff. It is true that the peculiar situation of the property afforded an explanation of the want of possession. Yet, had possession under the will been deemed the only test, it is manifest that the Court would not have allowed the will to be read. The decision then is put, not on the ground of possession, but other facts proven, which raised the presumption that it was genuine. These appear to have been the endorsements on it, and proof of the hand-writing of the clerks, and one of the judges, who certified. They were not received as proof of the due execution of the will, but with a view to show the antiquity of the instrument; and that it existed at the periods when those certificates bear date. The law of this case has never been overruled : and so far as applicable, must govern the present.
The paper offered as the will of Aert Middagh, bears date January 7th, 1777. By it, he devised his property to his five children, in different proportions.
I will first examine how stands the question of possession under the will. The testator devised to his son John, one acre and a half of land, joining to the south side of his land; and the remainder, which joined the acre and a half, he devised to his daughters, Margaretta and Magdalena.
By the evidence stated in the case, it appears that John Middagh owned a house and lot in his father’s lifetime; and never claimed more; and that the acre and a half devised, adjoined his house and lot on the south; that Aert Middagh died possessed of all the property described in the map made by Jeremiah.Lot, and annexed to the case, except the lot where John Middagh’s house stood. John Read testified, that he never heard, or knew, that John Middagh exercised any control over any part of the land south of his lot, unti *227after his father’s death, when he built four houses, next to each other, on the ground south of the place, where his own house stood. I think the possession thus taken by John Middagh, after his father’s death, must be considered as taken under the will; for he disclaimed any right previously. It is not stated when his possession commenced, or when the houses were built; but it may be inferred, that it was soon after his father’s death." There is no evidence to show how long this possession continued. If it had been for thirty years, that fact Avould have been sufficient to admit the will to be read.
I consider, then, the evidence thus far, as establishing the proposition that a possession, under the will, in one of the devisees, is made out.
It cannot, I think, be shown, that under the rule of thirty years possession, is meant that where there are ten devisees of separate parcels of a testator’s estate, to ten individuals, a possession in each devisee of more than thirty years, must be made out, before the will can be read as an ancient will.(a) When Margaretta and Magdalena entered on the premises under the will, does not appear; but the deed from the former to the latter, dated the 18th October, 1793, is satisfactory evidence that it was as early as that day. This was nearly 30 years previous to the trial. 1 admit it is short of the period entitling a party to rest on length of possession; yet it is a circumstance connected with the other proofs, going to support the presumption of the antiquity and genuineness of the will. I cannot forbear remarking it is matter of surprise, that evidence more full and conclusive, as to the holding and length of possession under the will, by the different devisees, had not been laid before the Court. I incline, however, to think, that enough is shown to warrant the reading of it in evidence. My opinion is not founded on a possession of thirty years under the. will: for to that extent it is not made out; but on the other alternative, that there are circumstances Avhich raise the presumption, that it is genuine. They áre the following: The hand Avriting of one of' the subscribing witnesses, long since dead is satisfactorily proved. *228The evidence of the possession of John Middagh, and the acts of Margaretta and Magdalena, tend to show the antiquity of the instrument, and its probable authenticity, as fully as they were shown in Jackson v. Laroway. I am therefore, of opinion, that, according to the doctrine laid down in that case, the will ought to be received in evidence on the footing of an ancient deed.
The next question is, whether the rights of the plaintiff are affected by the partition between Margaretta and Magdalena ? This will depend on the construction of the words in the will: “ I give unto my other two daughters, Margaretta and Magdalena, the remainder of my land which joins the acre and a half, to be equally divided between them, share and share alike ; and to be to them for and during their natural life; and after their death, then to be to their and each of their children ; and to be divided between them share and share alike.” In the construction of wills, I adhere to the rule laid down by Lord Kenyon, in Hay v. The Earl of Coventry, (3 T. R. 85.) 11 The general rule which is laid down in the books, and on which Courts alone can, with any safety, proceed in the decision of questions of this kind, is, to collect the testator’s intention from the words which he has used in his will; and not from conjecture. It is not necessary that any technical, or artificial form of words should be used in a will; but we must collect the meaning of the testator from the words which he has used: and cannot add words which he has not used.” Whenever words have received a judicial determination, the security of titles requires that such construction be adhered to. The words, “ equally to be divided,” in wills, go to the quality of the estate, and not to the limitation of it. They create a tenancy in common. So, also, in deeds, which receive their operation from the statute of uses; but not in common law conveyances. (1 P. Wms. 14. 1 Atk. 493, 494. 2 Atk. 122. 3 Atk. 525, 731. 1 Ves. 115. 2 Ves. 252, 257. Willes, 341. 1 Salk. 391, 392.)
If the construction put on these words be correct, then it follows, that Margaretta and Magdalena, being tenants in common for life, had no other power or control over the property devised, than is given by iw to per *229sons holding such estates. They cannot claim that the words, “ equally to he divided,” are directory as to a partition. The testator meant nothing more than the signification which the law attaches to those words. If, however, they are considered as directory, that partition be made between the tenants in common, I perceive no cause for believing, that the testator intended that those in remainder should be affected by the acts of the tenants for life. I can not suppose that in that event, he had any thing more in view, than that Margaretta and Magdalena should divide the premises for and during their lives; so that each might have a separate possession. It is highly improbable, that the testator intended to sanction the division to be made between the daughters, as binding on their children. A division by consent might have been very unequal, at the time when it would be made. Circumstances may have greatly changed the relative value, at the decease of either of the tenants for life; when the children of the deceased tenant would, in my opinion, become entitled to their share. Why should the property devised to them be accepted, thus fettered and impaired in value, by the act of the tenant for life 1 I cannot assign a motive, to induce the testator so to subject the property devised to his grand children. They were equally objects of his bounty. The power now claimed, as derived from the will, was not necessary forth e beneficial enjoyment by the daughters. It was enough for them to divide for themselves, so long as their estates continued. I cannot, therefore, construe these words, as implying an intent that a division made by the testator’s daughters, should be binding and conclusive on their children.
But it is contended, that if no power was given to divide, so as to effect those in remainder, then the children of the daughters would take per capita, so that if Margaretta had one child, ana Magdalena ten, they would take equally. I think this is a mistaken view of the devise. The latter clause is, “ and after their death, then to be to their, and each of their children, and to he divided between them share and share alike.” The argument supposes that, by denying the power of the tenant for life to bind those in remainder, the *230devise is to receive such construction; and further, that as long as either of the daughters are alive, the grand children can take nothing. I consider the will as giving a tenancy in common, for life, to each of the daughters, with no other powers than such as are incident to that estate; and the words, “ after their death,” are to be understood, that upon the death of either, the quantity or portion of the estate of which the daughter, so dying was seised as a tenant in common for life, became vested in her children, as tenants in common in fee.
What, then, were the powers of the daughters, as to a partition ? As a general proposition, it may be laid down, that partition, so as to affect the inheritance, is not one of the incidents to a tenancy for life. (2 Bl. Com. 122.) The estate devised to the grand children, is a vested remainder, though to be enjoyed in futuro, which nothing can defeat or set aside. (2 Bl. Com. 169.) It would, therefore, seem to be inconsistent with the nature of this estate, to turn that, which was a tenancy in common in all the land devised to Margaretta and Magdalena, to a tenancy in common in the divided share of the tenant for life.
Tenants in common are compellable, by statute, to make partition; which they were not, at common law. (2 Bl. Com. 193. Co. Lit. 198, b. sec. 318.) The partition between Margaretta and Magdalena, was voluntary, and undoubtedly valid as between themselves; but its operation was necessarily limited by the power incident to an estate for life, by the common law. The law is correctly laid down in Pledgard v. Lake, (Cro. Eliz. 718. 5 Bac. 862.) The case was this: A. tenant for life, remainder to B. in tail; he, in remainder, lets for years, to begin after the death of tenant for life; the tenant for life afterwards suffers a recovery, with voucher of him in remainder in tail, and dies. It was hold .that the lease shall take place; for the lessee may falsify the recovery, either by the common law or statutes; foi the tenant for life hath no power to bar him inremainder, without his assent or concurrence, as the tenant in tail in possession hath. (1 Co. 62. Poph. 6.) The principle of this case applies to the one before us. The tenant for life had *231no power to change or alter the tenancy in common, which those in remainder held in the premises devised to Margaretta and Magdalena. A partition does not change the claim, by confining it to a divided parcel. The partition was not made in pursuance of any statute; nor was it sanctioned by consent of those in remainder. In England, by the statute (31 Henry 8, ch. 1,) joint tenants, and tenants in common, of estates of inheritance, were compellable to make partition. The 32 Hen. 8, ch. 32, gave the like remedy to joint tenants and tenants in common, for life or years. (5 Bac. 699.)
The act of February 6th, 1788, to compel joint tenants and tenants in common to make partition, (2 Greenleaf. ed. of Laws N. Y. 13,) contains substantially, the same provisions. The second section of this act declares, that tenants in common for life, and where one has an estate with others that have estates of inheritance or freehold, may be compelled to make partition; but that no such partition shall be prejudicial to persons, other than such as be parties to the same. The tenants for life could not, by their individual act, prejudice the rights of the lessors of the plaintiff.
I am, on the whole, of opinion that a new trial be granted with costs to abide the event.
New trial granted.
Vid. Jackson v. Davis, ante, 123. S. P. as to a deed.