A. J. WALKER, J.
F. O. A. Sherrod having died without ever having had the possession of the slaves, the contingency on which he was to share in the profits of their *542labor is not presented; consequently, that feature of the bequest is to be left out of view in the consideration of the case. The bequest to be construed is a legacy, in the first place unqualified, to the grand-son, but by a subsequent provision made subject to this qualification, that if the grand-son should die, leaving a wife, but no child, nor the descendant of any child, then over to the persons in the proportions named. The grand-son having died in infancy, leaving no wife, nor child, nor the descendant of any child, to whom does the personalty bequeathed go ? The chancellor decided, that it went to the grand-son’s administrator, to be distributed among his next of kin. The appellants insist that it goes to the testator’s three eldest sons, or their families. The construction producing that result is, that the grand-son took only a life-estate, with remainder to the three eldest sons, subject to be defeated in toto by the grand-son’s leaving children, or their descendants, and in part by the contingency of his leaving a wife but no descendant from him. Rejecting that, we adopt as the true exposition of the clauses of the will in question the construction which gives to the grand-son the entire interest, subject, however, to be defeated in the contingency specified.
There is, in the first place, a clear and unambiguous gift of the absolute and entire interest. The language is, “ I give and bequeath,” &c. If the will had stopped with the former of the two clauses, no argument could have been made in favor of the proposition that the legatee took a partial interest. The case here, then, is one where an absolute gift is made, and a contingency is afterwards provided, in which the estate is to be defeated. Now, the rule is, that where there are clear words of gift, the court will never permit that absolute gift to be defeated, unless it is perfectly clear that the very case has happened, in which it is declared that the interest shall cease. — Harrison v. Foreman, 5 Vesey, 207. It is not to be inferred that the absolute gift is infringed, further than is expressed. The testator has with clearness and distinctness appointed *543one contingency, upon which the interest given was to fail. Upon the death of the grand-son, leaving a wife, but no child, nor the descendant of any child, “in that case'1'’ there is a limitation over in defeasance of the estate. When the testator thus specifies a concurrence of circumstances, upon which the estate was to fail, how can it be said that the estate must fail upon the occurrence of only one of those circumstances ? The death of the grand-son does not make “that case,” in which the will directs the limitation over.
There is in the books a class of cases, where a bequest to one, in the event of the non-existence, or on the decease of another, has been held to indicate an intention to make the latter a prior object of the testator’s bounty ; as, for example, a devise to A, upon the death of B. In those cases, the antecedent estate is not given ; and there must, therefore, be an intestacy, unless the courts imply a bequest of it. A bequest has, therefore, been implied in many cases, to him upon whose death the subsequent estate is to arise. The implication must not rest upon conjecture ; it must be necessary. The inference must be so plain as to be irresistible to the mind. — Brummel v. Prothers, 3 Vesey, 111; 1 Jarman on Wills, 465 ; 2 Lomax on Executors, 19; Browne v. DeLoet, 4 Bro. C. C. 535, note a; 2 Roper on Legacies, 1497-1498; Chum v. Respass, 1 Monroe, 25.. If a devise be made to the heir, after the death of another, the inference is irresistible, that the testator intended the latter to take a life-estate ; for otherwise the heir would take by inheritance, before the event upon which the devise to him was to take effect. The authorities upon this subject are numerous, and sometimes conflicting, and it is not necessary here for us to go into them. The principle upon which they proceed, has no application here.
The cases collected upon the briefs of appellants’ counsel, on the subject of implication, are all referrible to the doctrine stated above. Thus, in Crowder v. Clowes, (2 Vesey, 449,) a life-estate in the testator’s niece was implied from a bequest over to the person who might be entitled to the *544real estate, if she should die unmarried. So, in the case of Wainwright v. Wainwright, (3 Vesey, 558,) a similar decision was made, in reference to a bequest over, in the event of the deatli of one attaining the age of twenty-one years. So, also, in Hingham v. Baker, (Cro. Eliz. 15,) a life-estate was implied in favor of the wife, from a devise after her death to the son. To the same effect is the case of Hutton v. Simpson, 2 Vernon, 723. A review of all the authorities referred to would but afford cumulative illustrations of the same principle. The application sought to be made of those cases, would involve a perversion of them. They do not afford any authority for the cutting down an estate into a life-estate. If it be conceded, that the grandson would take a life-estate by implication from a bequest over after his death, it is not at all a sequence, that an estate actually bequeathed will be' diminished into a life-estate, by implication from a bequest over, in the event of there being a wife, but no child, at the time of his death. Such an implication would sitnply add a condition not authorized by any thing the testator has said.
The wife was, upon the occurrence of the contingency, to receive the same interest to which the laws of the State of Alabama would entitle her, if the legatee were seized and possessed of the same in fee-simple. This provision is not sufficient to defeat the estate given. It does not authorize the inference that the testator designed to give his grand-son an estate for life. If that effect were allowed, it would not harmonize with the preceding clause. Perfect harmony is secured between that and the clause which makes the bequest, by attributing to it a design to meet the contingency of an alienation by the grand-son. If the grand-son should leave a wife, but no descendant, then the limitation over was to take effect, just as if he were at the time of his death seized and possessed of the same in fee-simple, whether he was then actually seized and possessed of the same or not. Or, it may be that the words, without observing their technical meaning, were used by way of distinction to an estate defeasible upon a subsequent *545condition. Under either view, that harmony with the other provisions of the will, for which courts always strive, is observed. See 1 Jarman on Wills, 415-416. But, aside from that consideration, it is impossible that those words can be regarded as manifesting, with that, degree of certainty and clearness which the law requires, the intent to defeat an estate clearly granted by a previous clause. — Harrison v. Foreman, supra.
If we concede, that the testator’s grand-son took only a life-estate, it would result, that there is an intestacy as to the remainder ; the contingency not having occurred, upon which the limitation over was to take effect. It would not be a necessary implication from anything contained in the will, that the testator designed his three eldest sons to take the remainder. We may find grounds in the will for conjecturing that the testator designed to follow the legacies to the death of the first taker, and see that they did not pass awa,y from the class of children to which the first taker belonged; and we might then conjecture that he had that design in reference to the legacy to his grand-son. But it would be mere conjecture at last. The testator has not said so, nor has he said anything from which it can be inferred. Certainly, the court can not resort to conjecture, when the terms of the will are of intelligible import. To do so would be to make a will, conforming to what it is supposed the testator intended- — not to search for the intention in the construction of what is said. — See Manigault v. Bailey, 1 Bailey’s Eq. 298. It is not the province of a court to incorporate into a will provisions which it may be supposed the testator would have adopted if they had occurred to him. Nor is it the province of the court to provide for a contingency, neglected in the will, because there is room for conjecture that the testator would have done so, had he anticipated it. — 1 Jarman on Wills, 744-50; Parsons v. Parsons, 5 Vesey, 578; Holmes v. Craddock, 3 Vesey, 317 ; Roper on Legacies, 322, 326, ch. 21, § 9.
If, then, it be granted that the grand-son took only a life-estate, it would not open the door for the three eldest sons *546to come in. The will gives them an interest in a portion of the property, in a certain contingency, which never happened ; and if every intendment is made for the appellants, we can attain nothing more than conjecture that they were to take in any other contingency. Therefore, in deciding that the grand-son took the entire or absolute interest, to be defeated only in the specified contingency, we avoid a partial intestacy, which it is evident the testator did not contemplate. The case of Read v. Snell, 2 Atkyns, 642, is not like the case in hand. There, the devise was to R. or the heirs of her body ; and if she died, leaving no heirs of her body, then over. “Heirs of the body” was construed to mean children, and “or” was construed to mean ‘and ;” so that the devise, thus understood, was to R. and her children living at her death, and it was therefore decided that R. took only a life-estate.
Geo. Goldthwaite, for appellants.
1. If the words, “leaving a wife,” where they occur in the bequest, were used by the testator as words of condition, he must have intended to make the substitution of his sons to any portion of the prior gift dependent on the accident of the grand-son leaving a widow. It is incredible that any sane man should have made such a will. What possible motive or reason can be assigned for it ? To make provision for the grand-son’s widow out of the gift to the grand-son, in case he died without children, and to bestow all the balance of the property, after deducting the provision for such widow, on certain of the testator’s own children, is intelligible to every one ; but, for the testator, after making this disposition, to add, “in case my grand-son leaves no widow to provide for, my children shall have nothing,” is intelligible to no human being. It is nob merely irrational and absurd, but unnatural and startling. It makes the testator not only perpetrate the absurdity of making the gift over to the sons depend upon a matter with which it-can have no possible connection, but reverse the principles of human action, by making a contingency operate against the sons, which should operate in their favor. It makes provision for the three so^s, although the widow may survive, and takes it from them when there is no widow to provide for. A construction which leads to results so absurd and irrational, will not be adopted if it can be avoided. — Pearsall v. Simpson, 15 Yesey, 29-33 ; Key v. Keij, 19 Eng. Law & Eq. 617-24; Hart v. Fulh, ib. 438-43 ; McKinnon v. Seioall, 6 ib. 327.
*546As Chief-Justice Rice, being related to one of the administrators of Benjamin Sherrod’s estate, sits in the case only by consentof the parties, we deem it proper to observe that the foregoing opinion has the entire approval of every member of the court.
The decree of the chancellor is affirmed.
Note by Reporter. — The foregoing opinion was delivered at the January term, 1S58. A rehearing having been granted, on the application of the appellants’ counsel, the cause was held under’ advisement until the January term, 1863, when the following opinion was delivered by Mr. Justice Stone. The annexed brief is an abstract of the argument submitted by the appellants’ counsel on the rehearing.
Can any other construction be given ? Did the testator intend that the gift to the grand-sou should not be divested, unless he left a widow, and died without descendants? • His purpose was, in case the grand-son died without descendants, to make provision for his widow, if any, and also for the three eldest sons. This is the precise case for which he has provided, and which, if it had happened, would have divested the original bequest. That he both said this, and meant this, is certain; but, did he also mean that, if the grand-son died without descendants, no provision should be made for the sons ? The bare statement of the proposition is almost as monstrous and startling, as the will itself becomes under its adoption. If the testator, after using the words of absolute gift to the grand-son, had said, “but, should my grand-son die, leaving a widow but no descendants, in that case I give to my three oldest sons,” &., entirely omitting the gift over to the widow, it may be conceded that leaving a widow would be an event which must happen, to divest the original gift; but why ? because, the gift to the widow being omitted, the two contingencies designated could only be referred to the gift to the sons. However unnatural and absurd the disposition would have been, there would be no escape from it, for no other meaning could possibly have been given to the words. But here there is no such necessity. The gift to the widow follows the words “leaving a wife,” not immediately, but directly alter the only condition, so far as she was concerned, which was imposed ; and it is perfectly certain that, as to her, they were not used, in any proper sense, as a condition on which she should take, but were merely intended to designate a person for whom, if living at the death of the grand-son, provision was to be made out of the prior gift, in case he died without descendants. The testator thought of and expressed the case of there being a widow, with reference to the provision he intended to make for her, and not as a Condition, or contingency, on which the gift over to the sons was to depend. — Murray v. Jones, 2 Vesey & B. 313 ; Morton v. Wliitalcer,' 1 Durn. & East, 346.
This construction is exemplified by a devise to A in fee, and at his death, without issue then living, one-half to his • widow, if any, and the other half to B. Could there be any doubt that, in such case, the executory devise to B would depend alone upon the contingency of the death of A without issue living? The present case is an absolute bequest to the grand-son, and, in case of his death without descendants, to his widow, if any, and the residue to the three sons. Omit the gift .to the widow, and it is like a devise to A, and at his death, leaving a widow, but without descendants then living, to B; in which case, the devise over would depend on the two contingencies.
The case of Fear sail v. Simpson, supra, tends strongly to confirm the view last insisted on. There, the legacy was in trust to pay the interest to the separate use of A. for life, and after her decease, as to the capital, for her children ; if no' children, to pay the interest to her husband during his life ; “and from and after his decease, in case he shall become entitled to the interest,” then to pay the principal to the first cousins of the testatrix. A died, leaving no issue, nor any husband, he having died before her; and it was contended, that the legacy over to the cousins did not take effect, because the husband of A. never became entitled to the interest, — the event upon which alone the legacy was given over. Sir William Grant said : “ The only question is, whether Richard Stafford taking for life, was a condition to the cousins of the testatrix taking the capital. That would be a most absurd condition, undoubtedly; for there is no sense or reason in making the right of her first cousins depend upon a fact totally disconnected with any intention as to them. What was it to them, whether Richard Stafford took or not ? Such a construction is not to be unless absolutely necessary.” “ It was doubtful whether Richard Stafford would live to become entitled to the interest. The testatrix, giving the capital over after his death, recollects that he may not live to take the interest ; but, if he does, she makes the death the period at which her first cousins are to take. It is not a condition precedent, but fixes the period at which the legatees over shall take, if he ever takes. The words will bear that construction, and the reason of the thing seems to require it.”
Murray v. Jones, supra, in its facts, as well as the reasoning of the' learned judge who delivered the opinion, also bears a striking analogy to the case at bar. In that case, the question was as to the construction of a residuary clause, after a bequest to'the testatrix’ younger children ; “but, in case I shall have but one child living at the time of my decease,” or all but one die under twenty-one, or unmarried, to another family ; which was construed not to be a condition. The master of the rolls (Sir Wm. Grant) uses this language : “In my opinion, the first case put by Lady Bath — namely, that of her having but one child living at her death — does not contain a condition that she shall have one child living at that time. At first sight, the proposition relative to the having but one child may seem to include in it, and to imply, the having one. That is true, if the proposition be affirmative ; but by no means necessarily so, if the proposition be hypothetical or conditional. The proposition that A has but one child, is as much an assertion that he has one, as that he has no more than one ; but, when the having but one is made the condition, on which some particular consequence is made to depend, the existence of one is not required for the fulfillment of the condition ; unless the consequence be relative to that one supposed child. As if I say, that in case I have but one child, it shall have a certain portion, it is in the nature of the thing necessary that the child should exist, to be entitled to the portion ; but, if I say, that in case I have but one child of my own, I will make provision for the children of my brother, it is quite clear, that my having but one child is no part of the condition, on which the supposed consequence is to depend. My having one child of my own, would rather be an obstacle, than an inducement to the making of a provision for the children of another person. The case I guard against, is the having a plurality of children ; and it is only' the existence of two or more that can constitute a failure of the condition, on which the intended provision for my brother’s children was to depend. The plain sense of the proposition is, that, unless I have more than one, the provision shall be made.”
'If the testator in the present case had said, “but, if my grand-son shall die, leaving only a widow to provide for, 1 will make a certain provision for her, and also for the three sons, out of the prior gift,” holding the words “only a widow” to exclude all except descendants of the grand-son, the case would be identical with the case of .Murray v. Jones. And this is precisely what the testator did mean, when he used the words “leaving a wife but no. descendants.” The gift over was in part to the widow, the balance to the three sons, in case the grand-son died without descendants. What the testator was guardiug against, as to the ulterior gift, was not the existence of the widow, but the grand-son leaving descendants.
The construction insisted on by the appellants, is also established by the fact, that the gift over to the sons is residuary in its character. The intention of the testator was, as has been said, in case the grand-son died without descendants, but leaving a widow, to substitute her and the three sons in the place of the grand-son. He gives the widow the portion which she would be entitled to, as such, by the laws of Alabama, if the grand-son was the absolute owner, and the “balance” (residue) to his three oldest sons, &c. The testator, in the gift over, makes the widow the special legatee, and the three sons the general legatees. The presumption as to residuary gifts is, that the testator prefers the general legatee to all the world except the sjoecial legatee; that but for the sake of the particular or special legatee, he would have given all to the residuary or general legatee. It is upon this presumption that all lapsed legacies fall into the residuum. — Cambridge v.Bous, 8 Vesey, 15. •
If the testator intended the words “leaving a wife” to operate as a condition, he intended that the three sons should take no portion of the prior gift, if the grand-son left no widow. The intention of the condition, and the intention of the residuary clause, cannot exist together. They are irreconcilable. The testator could not have intended to say, “were it not that there may be a widow of my grand-son to provide for, I would, in case of his death without descendants, give over all I have bequeathed to him, to the three sons ;” and in the same breath to say, “but, should he leave no widow, the sons shall have nothing.” The whole clause would be an unmeaning, senseless, contradictory jargon. By the residuary clause, the testator shows, not that the existence of the widow is the inducement for giving to the sons any portion of the prior gift, but that she is the obstacle which prevents him from giving them all ; that the provision for the widow was the only limitation on his bounty to the sons.
But it may be urged, that in the case of the devise to A. in fee, and on his death without issue living, one half in fee to his widow, if he leaves a widow, and the other half to B., although it may be clear that the testator intended the limitation over to B. should vest upon the sole contingency of the death of A. without issue; still, that on a devise to A. in fee, and on his death, leaving a widow, but no issue living, then, in fee, one half to the widow, and the other half to B.; while the result in the two cases is the same, on the death of A. leaving a widow, but no issue, yet, from the collocation of the words in the last case, it is apparent that the testator intended to provide for the one case only which he specified; and that whatever he might have done, had it occurred to him that A. might die, leaving neither widow nor issue, the omission to provide for that case is fatal; that the testator having only declared certain events upon which the original devise was to be divested, it remains in force until these events. happen. To this position an answer is furnished by the principles of necessary implication, and the presumption of the residuary clause.
The principle of implication, as applied to wills, is illustrated by the cases of Jones v._ Westcomb, Ch. Pr. 316 ; Gulliver v. Wiclcett, 1 Wils. 105 ; Statham v. Bell, Cowp. 40; and Murray v. Jones, supra. In the first case, the devise is to A. (the wife of the testator) for life, and at her death, in fee, to the child of which she was enciente ; and if the child should die before twenty-one, in fee, one-third to A., and the remaining two-thirds to B. and C. The wife was not enciente; there was no child ; but the court held the devise over to the wife good. The event expressly declared and provided for, did not happen; the case of there being no child born was a casus omissus; it had not occurred to the testator, and was not provided for. But the court implied, that, although the case which the testator supposed did not happen, he would have made the same devise over to the wife in the case which did happen, had it occurred to him. There could have been no absolute certainty that he would have done so, for many testators might give their wives more, if about to become mothers, than if not in that situation. Gulliver v. Wickett, supra, was upon the same .will; and Statham v. Bell identical in principle. These cases are referred to, principally, for the purpose of showing that necessary implication, as it is termed, is not always an implication which is irresistible; as where, upon a devise to the heir on the death of A., it is implied that A. takes for life; for, in that case, as the heir does not take until A. dies, if A. does not take, no one can.
Blackstone says, that necessary implication in wills ‘•maybe only strong probability,” (2 Black. Com. 381;) and Lord Eldon, in Wilkinson v. Adam, (1 Yesey & Bea. 421, 466,) defines the same words “not natural absolute certainty, but so strong a probability of intention that an intention contrary to that which is imputed to the testator cannot be supposed.” Mr. Jarman, in language which conveys the precise idea more definitely and distinctly, states the principle thus : “ Sometimes, however, an executory gift is made to take effect in defeasance of a prior gift; i. e., to arise on an event which determines the interest of the prior devisee or legatee ; and it happens that the prior gift fails ab initio, either by reason of its .object (if non-existing at the date of the will) never coming into existence, or by reason of such object (if a person in esse) dying in the testator’s life-time. It then becomes a question, whether the executory gift takes effect, the testator not having in terms provided for the event which has happened, although there cannot be a shadow of a doubt, that, if asked, whether, in case of the prior gift failing altogether for the want of an object, he meant the ulterior gift to take effect, he would have answered in the affirmative.” Jones v. Westconib, with the other cases last cited, as well as Murray v. Jones, are referred by that author to this principle.
The principle being settled, let it be applied to the clause in question ; and to present it more clearly, take the naked case of direct gift from the testator, without the intervention of a prior gift. This may appropriately be done, because the rule as to devises which take effect or are divested upon condition is the same. In one case the heir, and in the other the prior devisee, is defeated. In . the first, the estate does not arise, until the contingency on which it depends happens ; and in the last, the original devise remains in force, until the happening of the event which is to defeat or divest it. The construction of the same words in either case would be the same, unless varied by some other part of the will. Thus, if a testator was to say, “should I die leaving a widow, but no descendants, in that case, I give her one-half of my estate, and the other half to my nephew;” the question as to the nephew taking, in case the testator died leaving neither widow nor descendants, would be the same which would be presented had the testator used the same words in defeasance of a prior gift.
In the cases of Jones v. Westcombe, and Gulliver v. Wiclcett, supra, the supposed case, on which the devise over to the wife was made, was the birth of a child, and the child not attaining twenty-one years; the actual case was the having no child. The implication, based upon the devise in the supposed case, was, that the testator would have made the same devise in the real case — that, as he gave the wife^ a portion of his estate, if he had a child, and the child died before twenty-one, he would have given her the same, if he had known there would be no child. In Murray v. Jones, the supposed case was that of the prior legatee leaving but one child; the implication was, that, as the testatrix made provision for the nephews if there wTas but one child, she would do no less for them, in case there was no child to provide for. So, in the case given above, the question would assume this shape: If a testator, in case of his dying without descendants, but leaving a widow, makes a certain provision for his nephew as well as his widow, would he make at least the same provision for his nephew, if there was no widow to provide for ? Had the question been put to a testator, could there, to use the language of Mr, Jarman, “have been a shadow of a doubt, that his answer would have been in the affirmative;” or, in the words of Lord Eldon, is there not so “strong a probability” that he would have done so, that “ the contrary cannot be supposed?”
It is true, that the cases cited in illustration of the doctrine of implication have sometimes been referred to the principle of a substantial performance of the condition imposed : that what has happened is equivalent to that which the testator supposed would happen ; but they cannot be rested upon this principle, without assuming or implying that the testator would have regarded the real as equivalent to the supposed case ; which is not absolutely certain, so that it comes to implication at last. The testator thought of the one case only, and the devise was made tvith reference to that case alone ; the case which did happen never occurred to him, and he made no provision for .it; hut, hy what he did in the supposed case, the implication is made as to what he would have done in the real case, had it occurred to him.
The case of a direct devise, without the intervention .of a prior absolute gift, has been taken to illustrate the doctrine of implication, for the reason, that the error, if error there be, in the opinion of the court, is in the too strict adherence to the letter, rather than the spirit of the rule, as laid down in Harrison v. Foreman, (5 Yesey, 207,) in these words : “ It is perfectly clear that, where there are clear words of gift, giving a vested interest to parties, the court will never permit that absolute gift to be defeated, unless it is perfectly clear that the very case has happened, in which it is declared that the interest shall not arise.” In the application of this rule to the case at bar, the opinion on the former hearing holds, in effect, that the bequest to the grand-son being in words of absolute gift, it can only be defeated in the very case in which it is declared by the testator it shall be defeated; that the case thus declared being the leaving a widow and no descendants, and that case not having happened, as a necessary consequence, the prior gift to the grand-son is not defeated. The result is, that while it is conceded that the heir may be defeated by implication, the prior legatee or devisee never can be; for, if the very case happens which is declared by the testator, there is no implication; if it does not, implication cannot be resorted to. In Harrison v. Foreman, the fund was bequeathed to A. for life, and after her decease to P. and S. in equal moieties; and in case of the death of either of them ■in the life-time of A., then the whole to the survivor living at her death. P. and S. both died in the life-time of A.; and Sir R. P. Arden, the master of the rolls, held, that the gift to P. and S. was a vested interest, to be divested upon the contingencies declared in the will; and as these contingencies never happened, the original bequest remained in force. The question as to the defeasance of the vested interest by implication did not arise, for there was no ground to rest it upon. - The case supposed and provided for, it is true, did not happen: the case which did happen — that of P. and S. both dying in the life-time of A. — may not have occurred to the testator, and certainly was not provided for hy him ; but it was impossible to do more than conjecture how he would have given the fund, had the case which actually did happen, occurred to him. And this is the difference between that case and the present. That the prior devisee cannot be defeated by implication, although the heir may be, seems to be against all reason; and no adj udged cases have been cited, which directly sustain such a doctrine. Murray v. Jones, (supra,) and the others cited for the appellants, are directly against it; for, in none of them, did “ the very case” declared by the testator, upon which the gift over was to take effect, happen.
The rule is laid down by Mr. Jarman, thus: “That estates, once vested, will not be divested, unless all the events which are to precede the vesting of the substituted devise happen,” (1 Jarm. on Wills, 756,) which gives full scope for implication, as well as construction; and the same learned author gives a warning note, against extending the case of Harrison v. Foreman too far, referring'to the more recent decision of Joslin v. Hammond, ,3 Myl. & R. 110. Hobnes v. Craddoch, (3 Yes. 317,) another authority cited in the opinion of the chief-justice, may well be questioned. Mr. Jarman does question it, and holds it irreconcilable with Pearsall v. Simpson, (supra,) and that class of cases, and there is little doubt that Sir Wm. Grant would have decided it differently. — 1 Jarm. on Wills, 746, note i. Conceding, however, the correctness of each of these decisions, and of every other case cited in the opinion, there is not one of them which goes so far as to hold, that a prior absolute gift cannot be divested by implication ; and if the words used by the master of the rolls in Harrison v. Foreman, were intended, as probably they were, to assert no more than the general proposition, that an absolute bequest remains in force until the title of the substituted legatee becomes complete, by the happening of the events which are to divest the original bequest, he states the rule accurately; or, if spoken with reference to the case before him, are correct; but, if intended to be taken as an unbending and universal rule, subject tono exception, they go further than the case before him warranted, and are not sustained either by principle or authority.
If .the principles of implication have been correctly stated and applied, it follows necessarily, that, in case of a devise in fee to the testator’s grand-son, and on his death, leaving a widow, but no descendants, then, in fee, one half to the widow, and the other half to the son of the testator; on the death of the grandson, leaving neither widow nor descendants, the son would take by implication; and it would be a singular anomoly, if, in the substituted devise, the son occupied the position of residuary, and the widow that of special devisee, the son should take nothing. But, in truth, this position, instead of weakening the implication, makes it irresistible; for, as has been seen, the presumption which attaches to every residuary bequest is, that the testator prefers the general to all the world but the special legatee. This presumption is inevitable: the testator can mean nothing else; and applying it to the devise in question, it is the same as if the testator had expressly declared, that, were it not that there might be a widow of the grand-son to provide for out of the prior gift, he would have given all to the three sons. This presumption is, of itself, a conclusive answer to every argument on behalf of the appellees; for it shows with unerring certainty the intention of the testator, in the case which has happened, to give over the whole of the prior gift to the three sons; that with reference to that case, he did not use the words “leaving a wife,” as a condition; and at the same time, marks the distinction between Harrison v. Foreman and the present bequest, as just the difference between conjecture and irresistible implication, — doubt and absolute' certainty.
By giving effect to this presumption, the intention of the testator is sustained, without violence to the words; every absurd, startling and monstrous feature disappears ; every seeming inconsistency is reconciled; and the well settled principles of implication are sustained. The blood of the testator in the line of the original legatee has become extinct, and the estate returns to his own blood, subject alone to that provision for the widow of the grand-son which our own law deems most just. No words could be used which express this intention more clearly, and no disposition made which bears upon its face the impress of a more considerate forethought and comprehensive justice. By interpreting the words of the master of the rolls in Harrison v. Foreman, not with reference to the particular case in which they were used, but as an inflexible rule of universal application, the whole doctrine of implication as to -the defeasance of vested interests is repudiated, the adjudications of the ablest judges overruled, and the presumption of the residuary clause, not merely defeated, but reversed, — made to speak directly the opposite of what was intended by the testator. Holding the words “leaving a wife” to be a condition intended by the testator to operate against the divestiture of the original bequest on the grand-son dying without descendants, goes still further, and, in addition to the results just adverted to, assumes that the same words were used in a double sense; the one rational, the other absurd. In the one sense, the disposition is marked by consideration, reason, and justice; in the other, the self-same words serve only to give expression to an act of unspeakable folly, startling inconsistency, and unnatural caprice.
STONE, J.
It is not denied that Vm. S. Swoope took a vested estate under the will of his grand-father ; but it is contended that, under the obvious intention of the testator, as expressed in his will, and by necessary implication arising from its language, the gift over to the eldest three sons of the testator was not made to depend on the contingency of said Swoope’s dying, leaving a -wife surviving him ; that that was rather an obstacle in the way of the gift over ; that the real and only condition on which the gift over was made to depend, was that the grand-son, dying, should leave neither child, nor descendant of a child, surviving him.
The authorities conducing to support this proposition are the following : — Murray v. Jones, 2 Vesey & B. 313 ; Pearsall v. Simpson, 15 Vesey, 29; McKinnon v. Sewell, 5 Sim. 78 ; Hulton v. Simpson, 2 Vern. 722 ; Gulliver v. Wickett, 1 Wils. 105; Jones and Westcomb, 1 Eq. Cases Abr. 245; Andrews, dem. Jones and Fulham, 2 Eq. Cases Abr. 264 ; Earl of Newburgh v. Eyre, 4 Russ. 454; Robinson v. Robinson, 1 Burr. 38-50 ; Doe, ex dem. James v. Hallett, 1 M. & S. 124; Hill v. Smith, 1 Swanst. 195 ; Doyne v. Cartright, 1 Coll. 482 ; Wainwright v. Wainwright, 3 Vesey, 558 ; Key v. Key, 19 Eng. L. & Eq. 617.
Murray v. Jones is probably the strongest of these cases for appellants. In that case, Lady Bath, by will, gave the income of the bulk of her estate, first to her father, and then to her husband, successively, for life; and at the death of the survivor, gave the property to the children of Lady Bath ; but gave it over to Mrs. Markham, now Mrs. Fawcett, and her children, upon any one of the following events : 1st, in case Lady Bath should have but one child at the time of her decease, be the same a son or daughter; 2d, in case she should have two or *560more sons, and no daughter or daughters, living at the time of her decease, and all of them but one shall depart this life under the age of twenty-one; 3d, in case she should have two or more daughters, and no son or sons, living at the time of her decease, and all of them but one should die under twenty-one, and unmarried ; or, 4th, in case she should have both sons and daughters, and all but one, being a son, should die under twenty-one, or, being a daughter, should die under that age, and unmarried. Lady Bath died, never having had a child; and the property was decreed to Mrs. Fawcett, formerly Mrs. Markham, and her children ; Sir Wm. Grant, master of the rolls, remarking, that the true condition on which the gift over was to take effect, was not that Lady Bath should have one child, but that she should not have more than one at the time the gift over was to be enjoyed. The argument of Sir Wm. Grant •in this case is very vigorous ; and from his acknowledged ability, it is contended that it should have great weight.
Looking into the cases, a list of which is given above, it will be seen that the presumed intention of the testator, not clearly expressed, on which some of them are made to rest, or the necessary implication, mentioned in the books, brought to the aid of others, was in every instance indulged, not to defeat a vested estate, but, in most of them, to prevent an intestacy. The exceptions are the following :
Hill v. Smith, 1 Swanst. 195, in which it was attempted to divest the title of the testator’s son, a primary legatee, on some confusion and inaccuracy in the language of the will. The court, from the general tenor and context of the will, came to the conclusion, that the testator did not intend to cut down his son’s estate on the event which had happened.
Doe v. Hallett, 1 M. & S. 123, is very like the last in principle. The testator, by mistake, described a certain person as an only son, when there was another living. He then made provisión for such [supposed] only son, and for the other sons afterwards to be born. This family was the primary object of his bounty ; and the first son dying, the *561second son was declared entitled to the devise. This was pronounced to he the intention ol the testator, gathered from the language of his will.
In each of these cases, it will be perceived, that the effort was made to exclude the primary object of the testator’s bounty; in one of them, his heir-at-law. To accomplish such object, requires express language, or implication amounting to such “strong probability of an intention, as that the contrary can not be supposed.” — 1 Jarman on Wills, m. p. 465, and note.
Another case relied on for appellants is Hart v. Tulk, 19 Eng. L. & Eq. 438. In that case, the court became satisfied that there was an obvious mistake in the draft of the will, by employing the word “fourth,” where “fifth” was intended. This conclusion was attained, by considering the whole tenor of the will, which evidently contemplated equality of benefit to each and every of the testator’s children; and correcting this mistake, carried into effect the general intent of the will. There is much in the will which tended to show the mistake. ’ ■
We have stated the extreme cases on one side, which are presumed to make in favor of the appellants. There are many cases of a contrary tendency. In the case of Shuldham v. Smith, (6 Dow. 22,) testator had devised real estate in trust to pay the clear rents, issues, and profits, and in certain proportions, to certain persons in the will mentioned, for life; and then proceeded to devise as follows: “And from and after the death of the survivor of them, the said L. S.” &c., (naming the several persons to whom the above life-interests were given,) “then I give and devise, all and singular, the said manor, messuages, lands, &c., unto all and every the children of my late sister, E. C. by her three several husbands,” (naming them,) “that shall be then living, and their heirs and assigns forever, equally to be divided between them as tenants in common, and not as joint tenants; and if there should be but one such child, and no issue of any of the other children then li oing, then, and in that case, I give and devise all my said real *562estate in Ireland, unto such surviving child, his or her heirs and assigns, forever. At the death of the surviving annuitant, there was only one child of the sister E. C. then living; but there was issue of several of the other children then living. It was held by the house of lords, in concurrence with the unanimous opinion of the judges attending, that there was an intestacy, from the death of the surviving annuitant; the event which happened not having been provided for.
In Doe, ex dem. Raddiffe v. Bagshaw, (6 Term Rep. 512,) the devise was to Margaret, an only child, for life,, remainder to the first son of her body, if living at the time of her death, and the heirs male of such son, and, in default of such issue male, remainder to R. B. Margaret had an only son, who died during the life of his mother, leaving a son. It was held, that Margaret took only a life-estate; that neither her son nor grand-son, took anything under the will, but the devise over to R. B. took effect.
Of similar import are the following cases : Wingrove v. Palgrave, 1 Pr. Wm. 401; Holmes v. Craddock, 3 Vesey, 317; Doe, dem. Vessey v. Wilkinson, 2 Term Rep. 209; Doe v. Jessep, 12 East, 288 ; Doe v. Rawding, 2 B. & Ald. 441; 1 Jarman on Wills, m. pp. 744 et seq. See, also, Brown v. Clark, 3 Vesey, 166 ; Scott v. Chamberlayne, ib. 302.
In Harrison v. Foreman, (5 Vesey, 207,) the court said: “It is perfectly clear that, where there are clear words of gift, giving a vested interest to parties, the court will never permit that absolute gift to be defeated, unless it is perfectly clear that the very case has happened in which it is declared that interest shall not arise.” This sentence is somewhat obscure; but the context shows, that the writer meant to say, that an absolute gift shall not be defeated, unless the case has happened in which it is declared that the estate shall cease.
Under the will of Benjamin Sherrod, an absolute estate vested in Wm. S. Swoope, subject to be defeated on a contingency. Has the event happened, on which it was declared the estate should cease ? It has not literally hap*563pened; but we are asked to presume that the testator did not mean precisely what he said, because we can not perceive any reason why the fact that Wm. S. Swoope should have and leave a wife surviving him, should incline the testator to make larger provision for his own children, than he would have made if his grand-son had left neither wife nor child. We concede this. But then a testator may give no reasons, or very foolish reasons, for his acts; yet, if he have testable capacity, his will must be carried into effect. Or, he may omit to provide for many possible events, when it is morally certain that, if such contingencies had occurred to him, he would have made provision for them ; still, courts can not supply such omissions. It is our province to expound, not to make wills.
We adopt as our own the language of Chancellor Johnson in Manigault v. Deas, 1 Bailey’s Eq. 302 : — “If speculations on the subject of intention were admissible, I should probably arrive at the conclusion to which this argument leads. Ignorant of the feelings which might have operated on the testator, I am ready to confess that I can see no reason” [why the leaving a wife should be a condition of the gift over”]. “But we are forbidden by the rules of law to indrrlge in conjecture. The testator’s power of disposition over the property is unlimited. If he will, he may indulge his partialities, and his prejudices, and exercise wisdom or folly in the disposition of his estate.” .
If we enter upon the broad sea of speculation as to the probable intention of the testator, whither shall we be drifted, and where find a safe anchorage ? When the testator made his will, it was possible that Wm. S. Swoope would die unmarried. It was also possible that he might die leaving a wife and children, or a wife without children, or children without a wife surviving. It was also possible that he might leave children, who might all die unmarried, and leaving no descendants. The will makes express provision for only one of these contingencies. It is said there is an implication to meet one of the other contingencies— the one which has happened. How about the others ? *564Suppose the grand-son had died, leaving children and no wife; would there then have been an implied gift over ? No one will contend for such result. The absolute title was not intended to be divested by that event. Suppose he had died, leaving children, and they had all died under age, and without issue ; what then ? See note to Holmes v. Craddock, 3 Vesey, 321.
We have indulged in these reflections to show the great peril that must attend every step we take, when we travel out of the language of a will in pursuit of a conjectured intention of the testator. The will must be expounded by its own terms, and not by the after-accidents that may befall thé devisees, unless those accidents are provided for by the will. The rules of construction, as compiled by that accurate writer, Mr. Jarman, forbid the latitude now invoked. — 2 Jarman on Wills, m. p. 742. He says :
XI. “ That in general, implication is admissible only in the absence of, and not to control an express disposition.”
XII. “That an express and positive devise can not be controlled by the reason assigned, or by subsequent ambiguous words,” &c.
XIII. “ That the inconvenience, or absurdity of a devise, is no ground for varying the construction, when the terms of it are unambiguous; nor is the fact that the testator did not foresee all the consequences of his disposition, a reason for varying it,” &c.
XXI. “That the construction is not to be varied by events subsequent to the execution,” &c.
The rule as to implications is declared in the following authorities : Gardner v. Sheldon, Vaughn, 261; Bamfield v. Popham, 1 Pr. Wms. 56 ; Brown v. DeLaet, 4 Bro. C. C. 435, and note (a); 1 Jarman on Wills, 435, and note 2 ; 2 Lomax on Ex’rs, 19; Rathbone v. Dyckman, 3 Paige, 9. Necessary implication is defined to be, “such a strong probability that an intention to the contrary can not be supposed.” It is sustainable only on the principle of carrying into effect the intention of the testator.
The following authorities seem to us to be precisely in *565point, and to show that the chancellor did not err in the decree rendered in this cause: Devise and bequest for the benefit of a daughter ; “ and from and after the decease of my said daughter, in trust to convey and assign the said several last mentioned freehold and household estates, and the said £1000 stock, unto the heirs, executors, and assigns of my said daughter,, for and according to all my estate and right therein respectively. Nevertheless, in case my said daughter shall intermarry, and have no child or children, then the said estates and money in the funds shall belong to my son George Wm. Russen; or, in case of his decease before my said daughter, then to such child or children as he may happen to leave,” &c. After the death of testator, the said George Wm. died without issue, having made a will, disposing of the property; and subsequently Mary Ann Russen, the daughter of David Russen, the testator, intermarried with one Noble, and was his wife at the time of the trial; but she had no child. The court said: “I think that, upon the construction of this part of the will, independently of the contingent executory gift over, there is an equitable estate for life, with an equitable remainder to the heirs, executors,. administrators, and assigns; and that Mrs.Noble has an absolute estate, subject to be defeated by the executory gift over. And if this be so, the question is, whether the particular event in which the vested estate was to be divested, can now happen; and having regard to the intention of the testator, and the words in which the gift over is expressed, I am of opinion that the gift over was to take effect only in the event of Mrs. Noble’s marrying, and dying without issue, in the lifetime of her brother, or of such child or children as he might happen to leave ; and as he died in her life-time, and had no child, I think that the contingent executory gift cannot take effect.” — Jackson v. Noble, 2 Keene, 590.
A testator devised to Thomas Cooke, and added: “ But my will is upon this further condition, that in case the said Thomas Cooke shall die an infant, unmarried, and without issue, then J do hereby give and devise” [the premises] *566“unto the said Wm. Cooke and his three other children,” &c. Thomas Cooke entered upon the premises, lived to attain his majority, married, and died, never having had issue. The question was, whether Wm. Cooke and his three other children took under the will. It was ruled that they* did not. Lord Ellenborough, in delivering the opinion of the court, held the following language : “If the court should confine the estate’s going over to a dying without issue, they must reject the words ‘infant, unmarried.’ If they should retain them, and read them as the counsel for the defendant contended they should be read — viz., ‘If he die an infant, or die unmarried, or, being married, die without issue/ this would be, in effect, reading 'the will as if it had given the estate over on any one of these single events.” — Doe, ex dem. Everett v. Cooke, 7 East, 271.
So, in the case of Williams v. Chitty, (3 Vesey, 545,) where there was a devise to A. and her heirs; but, if she dies under twenty-one and unmarried, to B. and her heirs. A. died in the life-time of the testator, under twenty-one, and without issue, but had been married. It was conceded in the argument that B. did not take under this clause of the will.
See also, Humberstone v. Stanton, 1 Ves. & B. 385 ; and Joslin v. Hammond, 3 Myl. & K. 110.
The decree of the chancellor is affirmed.
R. W. Walker, J., having been of counsel, not sitting.