Three points of view are presented by the contending defendants. They are:
(a) That the seventy-five per cent, of the estate should be immediately paid over to the widow and the guardian of the daughter absolutely and in equal shares. This is the view of the solicitor for the widow and daughter.
(b) That the seventy-five per cent, should be held as a trust fund upon the following trusts: to pay over the net income to the widow and daughter during the widowhood of the widow and the lifetime of the daughter; in case the widow never remarries, to pay the net income to her and the daughter and to the survivor of them for life; in case of the remarriage of the widow and the death of the child, or upon the death of the survivor .(the widow never having remarried) then the corpus of the fund to be paid over to the father, mother, brother and sisters, or to their heirs, in the proportions specified in the will. This is the view of the solicitors for the father, mother, brother and sisters.
(c) That the seventy-five per cent, should be held in trust, the income to be euqally divided between the widow and daughter during the lifetime or until the remarriage of the wife and upon the death or remarriage of the widow the corpus to be paid to the daughter if living. If the widow should survive the daughter, then the corpus should upon her death or remarriage, go to the father, *88mother, brother and sisters in the proportions specified by the will. This view is suggested by the solicitor for the complainants.
While the complainant in its capacity of complainant as distinguished from itself in its capacity of guardian, is a neutral party as between the contending defendants, yet it has conceived that its duty to what it conceives to be a trust under the will justifies it in venturing to suggest a construction which none of the defendants has seen fit to support.
All the questions involved in the conflicting contentions are involved in the argument by which the solicitor for the widow and child seeks to support the view designated above as “(a).” The argument to this end proceeds upon these points: that the language “in case of death” wherever used is referable to death in the lifetime of the testator; that neither the widow nor child having predeceased the testator, the limitation over in case of the death of either or both is therefore to be no longer regarded;., that the interest which the widow and child take is thus-an absolute and unconditional one unless the provision concerning the widow’s remarriage imposes upon her share a condition subsequent, upon the happening of which her share would go to the child; but that such a condition being in general restraint of marriage is condemned by the law as void, and the widow’s share is therefore as free and unhampered as is the child’s; that both take as tenants in common in equal shares; and though the gift to them is in the form of a gift of income from a trust fund, yet inasmuch as there is no limit fixed to the trust in point of time, and no beneficaries other than the widow and child are in any way entitled to either the income or fund, it follows that the fund now goes to them absolutely in equal shares.
The first question which this argument presents for consideration is whether the phrase, “in case of the death of," when used with respect to the wife and daughter refers to their death in the lifetime of the testator, or to their death at any time. Mr. Jarman in the second volume of his Treatise on Wills (5th Ed.) at *752, states the question which language of this character presents, as follows:
“Where a bequest is made to a person, with a gift over in case of his death, a question arises whether the testator uses the words ‘in case of’, in the sense *89of at or from, and thereby as restrictive of the prior bequest to a life interest; i. e., as introducing a gift to take effect at the decease of the prior legatee under all circumstances, or with a view to create a bequest in defeasance of or in substitution for the prior one, in the event of the death of the legatee in some contingency. The difficulty in such cases arises from the testator having applied terms of contingency to an event of all others the most certain and inevitable, and to satisfy which terms it is necessary to connect with death some circumstance in association with which it is contingent; that circumstance naturally is the time of its happening; and such time where the bequest is immediate (i. e., in possession), necessarily is the death of the testator, there being no other period to which the words can be referred.”
“Hence,” he proceeds, “it has become an established rule that, where the bequest is simply to A. and in case of his death, or if he die, to B., A. surviving the testator takes absolutely.” This is undoubtedly the general rule and has been recognized in this State. Rickards v. Gray, 6 Houst. 232; Jones v. Webb. 5 Del. Ch. 132; Marvel, Adm'r., v. Wilmington Trust Co., et al., 10 Del. Ch. 163, 87 A. 1014. That the words, “in case of the death” of the first legatee, occurring in such bequests as this general rule refers to, are doubtful and ambiguous in meaning is recognized by the Master of the Rolls in Cambridge v. Rous, 8 Ves. Jr. 12. They do not necessarily mean death in the lifetime of the testator. They may mean also death generally, in which case the effect would be to reduce the interest of the first legatee to a life interest with remainder over. The considerations mentioned by Mr. Jarman, however, have been accepted as sufficient by most courts to treat the words when uncolored by anything else in the will as referring to death in the lifetime of the testator. That the general rule to this effect has not been accepted without hesitancy is apparent, however, from the opinion of Lord Chancellor Broughman in Home v. Pillans, 2 Myl. & K. 15, where he says that questions of this kind rest more upon precedent than upon principle and that the construction which the general rule adopts has been accepted with reluctance and termed unnatural by one Lord Chancellor. It is founded, says he, “upon a supposition which, if not violent, is yet somewhat strong, inasmuch as the maker of a will does not naturally provide for the event of his surviving his legatees, the selected objects of his posthumous arrangements.” In this country, one jurisdiction at least appears to have refused to accord its rec*90ognition to the general rule. Neal v. Hamilton Co., et al., 70 W. Va. 250, 73 S. E. 971.
Of course it is the intent of the testator, after all, which all construction seeks to ascertain. And so, as observed by the Chancellor in Marvel v. Wilmington Trust Co., supra:
“Where the contingency of the death of the legatee can be referred to another period than during the lifetime of the testator, the court will not hesitate to so construe it.”
And in searching the entire will for a manifestation of intent that the testator meant his expression to refer to the death of the legatee at any time, either before or after his own, it was observed by the Vice-Chancellor in Fischer v. Fischer, 75 N. J. Eq. 74, 71 A. 488, that the court will, if the manifestation of such intent be “but slight,” nevertheless recognize it as controlling. Cases decided by Lord Thurlow show, in the opinion of Mr. Jarman, that that great equity judge thought that “very slight circumstances suffice to make the words under consideration refer to death at any period.” 2 Jarman on Wills, (5th Ed.) *754. Spealdng of this same general rule in another one of its phases, the court in Benson, et al., v. Corbin, et al., 145 N. Y. 351, 40 N. E. 11, remarked that:
“It is said to maintain its hold somewhat weakly and with a doubtful grasp, and to yield easily to any fact or circumstance indicating a different intention.”
In the earlier New York case of Fowler v. Ingersoll, et al., 127 N. Y. 472, 28 N. E. 471, the court said:
“The rule is an arbitrary one and has often been said to rest more upon precedent than upon reason, and in Vanderzee v. Slingerland, 103 N. Y. 47 [8 N. E. 247, 57 Am. Rep. 701], Judge Andrews said that ‘the tendency is to lay hold of slight circumstances in the will to vary the construction and give effect to the language according to its natural import.’”
In Rickards and Wife v. Gray, 6 Houst. 232, the Chancellor, in delivering the opinion of the Court of Errors and Appeals, at page 245 appears to note a distinction in the authorites between those cases where the gift to the first legatee is followed by words of limitation or benefit, as to A., his heirs and assigns, or to A. forever, or to A. for his own use and benefit and then over in case of his death, and those cases where the gift to the first legatee is *91merely in general words without any expressed indication that it is intended to be absolute and then over in case of his death. In the former case, he states the general rule to be applicable, whereas in the latter he indicates it to be inapplicable; that is to say, in the former case the contingency of death is referable to the period of distribution, while in the latter it is not. Without pausing to examine the authorities with the view of ascertaining whether a distinction founded on this consideration is in fact supported by them, this nevertheless seems to me to be deducible from the language referred to, viz. that a construction which cuts an absolute interest down to one for life will be more readily made where the language granting the absolute interest is general in character than it will be where the language granting the absolute interest is express and positive. If this is fairly deducible from the Chancellor’s language in that case, it is illustrative of and in harmony with the spirit of judicial opinion elsewhere, just referred to, by which the construction which the general rule imposes on the words “in case of death of” in such bequests as we are here considering, is regarded as resting on a slight foundation and easily overturned.
With these principles of law in mind, it is now in order to examine the language which the will in the instant case presents for consideration. The language disposing of the seventy-five per cent, provides at the outset that it is “to go to my wife and child.” If the testator had stopped here, it would of course be plain that the bequest would be of an absolute interest in the wife and child in equal shares. But the rest of the will is equally plain in disclosing an intent that the wife and child should not receive possession of the fund. It was to be held in trust. Whether or not the trust could be effectuated as a matter of law, because of indefiniteness of duration, is of no moment insofar at least as the intent of the testator was concerned. He wanted a trust and desired that only the income should be used. I take it that even though the wish of a testator, or the manner of its accomplishment, will not be complied with because of legal inhibition, nevertheless the language he uses, though vain to accomplish-what he wants, may be legitimately considered in the attempt to discover the intent which the whole will seeks to express. This testator wanted a trust, the income only *92to be used. When he proceeds to express his first limitation after haying said that the seventy-five per cent, should go to his wife and child, he says:
“In case my wife remarries the trust fund shall pass to my daughter.”
From this it would appear on first thought that upon the contingency named the trust was to end. But that can hardly be, because he had before stated that only the income was to be used. This, if standing alone, would not negative the idea of a termination of the trust upon the remarriage of the wife. But when connected with the later and final limitation over, it seems plain that the trust if the child succeeded to the wife’s share was nevertheless still to continue. Because in the further contingencies of the death of the child after the wife’s remarriage and of the death of the wife and child, the formula of language by which the father, etc., are to take is the same, viz. “this trust fund shall pass to my father,” etc. Now if the wife remarried the trust fund was to pass to the child. But if the child died after the trust fund thus passed to her, the will provides that the same trust fund should pass to the father, etc. Here is an inconsistency, therefore, which is irreconcilable except perhaps on the theory that the event of the death of the child was thought of by the testator as occurring in his own lifetime, and the gift to the father, etc., was merely substitutionary. Thus the construction contended for by the solicitor for the widow and child seems to be supported. But the death of the child to which the testator was referring could not have been meant by him as occurring in his own lifetime, because his widow was to have become remarried before its occurrence. The language is:
“In case of the death of my daughter, and my wife having remarried, * * * this trust fund shall pass to my father,” etc.
See Randfield v. Randfield, 2 De. G. & J. 57.
But it seems to be clear that “trust fund” whenever it is spoken of as passing to any one does not mean what it literally indicates. It means rather that the right to income is to pass. In terms to be sure the trust fund itself is made to pass to the child upon the widow’s remarriage, and no immediate reference to *93income is made in that contingency. But where the same formula of language passes the trust fund to the father, etc., there is an express direction that the income shall be apportioned between the father, etc. Connecting this last provision with the before expressed general wish that only the income should be used, it seems clear that the testator’s language, wherever he referred to the trust fund, must be taken to have referred to the income only. If this be so, no argument can be based on the idea that a termination of the trust upon the widow’s remarriage contemplates the daughter’s death in the testator’s lifetime. It is hardly likely that the testator would have desired his six year old daughter to receive an absolute gift of the fund, but that if his mature parents, brother and sisters became actual beneficiaries he would want a trust for their protection.
The coupling, however, of the death of the daughter with the prior remarriage of the widow is a circumstance so clearly referring the event of death to a time contemplated as posterior to the testator’s own, that it would be greatly understating its significance to describe it as slight, which according to respectable authorities is enough to overturn the general rule before referred to. It is more than slight. It is compelling. Certainly, therefore, in one event at least the testator was making provision against the daughter’s death subsequent to his own, viz., her death after the rémarriage of his widow. When he again mentions the possibility of his daughter’s death it is in connection with that of his wife’s also — the language then is “or in case both my wife and daughter die.” In this clause it is true there is no immediate coupling of death with such a posthumous event as the remarriage of his wife. But it is impossible to believe that the time of death here provided for must have been intended, unlike the other, to refer to death occurring in his own lifetime. The death of both the widow and child must have been thought of by the testator as referable to any time in general, after as well as before his own. Unless this be so we must suppose that the testator meant his wife’s interest, in case she survived him, to be absolute if she never remarried, but that his daughter’s interest should be similarly absolute only in case she survived her father and her death antedated her mother’s possible remarriage; for if she died after her mother’s re*94marriage, then by the plain language of the will the rights of the father, etc., immediately accrued. To refer the event of death when it is provided for in one case to an indefinite time generally and when it is provided for in the other case to the period of the testator’s death, would seem to me to be unwarranted. It would so far as the daughter is concerned make the existence or non-existence of an absolute interest in her depend upon the entirely irrelevant circumstance of whether her death occurred before or after her mother’s remarriage.' In another aspect of the matter, if the mother remarried, the daughter’s interest would be cut down to life, notwithstanding that that very circumstance of the mother’s remarriage, if it had any connection at all in the testator’s mind with his provision for his daughter, would in all probability have supplied a reason for wanting the daughter’s interest to be absolute so as to render her free from the possible adversities of new and untried home conditions. It seems to me that this result which attends the referring of the event of death in the various contingencies to different periods, is so unreasonable that we must conclude that the death of both the wife and the child was meant by the testator to be referred to the same period at all times, and he having"unmistakably indicated that his father, etc., were to take in one case after the death of his child, whether before or after his own, they were likewise to take in all cases after the death of both wife and child whenever occurring.
It follows, therefore, that the interests of both the wife and child were meant not to be entire and absolute in them, but that these interests are cut down by the subsequent language to life interests, with remainder over to the father, etc., in the proportions designated.
I am quite aware that such a construction results in this— that the testator has preferred that his father, mother, brother and sisters and their heirs shall enjoy this fund after his wife’s and daughter’s lifetime, and that he has put it beyond the power of his wife and daughter to 'dispose of any part of it for their necessities, and of his daughter to provide for her possible children, his own lineal issue, out of any of it. Such however seems to have been his intent. I can see no other logical construction to give to his language. If this is not what he meant, it is unfortunate that he *95used the language he did. Courts, however, cannot found a construction upon conjecture. They must take the language as they find it. If a casus omissus is suggested and no warrant is found for extending the construction to embrace it, the only alternative is for the court to ignore it or to remake the will, an undertaking which of course no one would advocate.
There being then a life interest in the income to the wife and daughter, it is in equal shares, with the provision that in case of the wife’s remarriage the daughter receives it all for life. (Whether this condition terminating the wife’s life interest is valid will be later considered.) Upon the expiration of their interests, there is a remainder over to the father, etc. Inasmuch as the remainder over to the father, etc., is sought to be kept in the form of a trust of unlimited duration with all the income payable to the beneficiaries, the law considers that the fund itself is given to them. Lorton v. Woodward, 5 Del. Ch. 505; Bishop v. McClelland’s Ex’rs., 44 N. J. Eq. 450, 16 A. 1, 1 L. R. A. 551; In re Goldmark, 186 App. Div. 447, 174 N. Y. S. 595; 28 R. C. L. 214; 2 Schouler on Wills, Executors and Administrators, § 1132. The trust, however, continues to exist while the interjests of the wife and dáughter are alive, it being a necessary instrument created by the testator for the convenience and protection of the estate in remainder.
The foregoing answers all the questions presented by the case except one; and this is whether the condition imposed upon the widow’s interest is valid. Where estates are affected by a condition against marriage, a material distinction is to be observed between those cases where the interest is given in such manner as to mark by way of limitation the duration of the estate by the continuance of the unmarried status, and those cases were a gift is made and its duration sought to be terminated by providing that subsequent marriage shall operate as a condition subsequent in defeasance of the gift. Morley v. Rennoldson, 2 Hare, 570; Arthur v. Cole, 56 Md. 100, 40 Am. Rep. 409. An illustration of the former, where the provision against marriage took the .form of a limitation and was held good, is found in King v. Phillips, 1 Houst. 349, where the bequest of a share was to the testator’s wife “during her widowhood.” It was held that the window’s interest in the share ceased upon her remarriage, that event marking the boundary of the interest given. *96The bequest in the instant case is not, however, in this form. Here there is first a gift as construed of an interest for life and a provision for its earlier termination in case of remarriage. It is similar in principle to the language found in the will which appears in Wilmington Trust Co. v. Jacobs, 9 Del. Ch. 77, 77 A. 78. But the validity of the condition was not passed upon in that case. So that the question of whether the law permits a husband to provide that his bequest to his widow may be defeated by a condition subsequent against remarriage, is an undecided one in this State. That a condition in general restraint of marriage is contrary to public policy and therefore void, seems to admit of no doubt. But there is an exception to the rule which, by overwhelming weight of authority both American and English, is as well established as the rule itself. This is that such a condition when imposed upon a widow is valid. Scott v. Tyler, 2 Bro. C. C. 431; Marples, et al., v. Bainbridge, et al., 1 Madd. 590: Grace v. Webb, 15 Sim. 384; Lloyd v. Lloyd, 2 Sim. N. S. 255; Gaven v. Allen, 100 Mo. 293, 13 S. W. 501; Knight v. Mahoney, 152 Mass. 523, 25 N. E. 971, 9 L. R. A. 573; Chapin v. Cooke, et al., 73 Conn. 72, 46 A. 282, 84 Am. St. Rep. 139; In re Biles' Will, 88 Misc. Rep. 452, 151 N. Y. S. 1097; Bryan v. Harper, et al., 177 N. C. 308, 98 S. E. 822; 40 Cyc. 1699; 28 R. C. L. 323; 2 Alexander on Wills, § 1065. In the cases just cited it will be observed that the exception was recognized where personalty was involved, and in some of them the absence of children was not recognized as altering the situation. The validity of such conditions when applied to widows has sometimes been placed on the ground that it was thought that the interests of children justified the condition. But this is not the only ground on which it has been put. Vice-Chancellor, Sir R. T. Kindersley, in Lloyd v. Lloyd, supra, placed it on the ground that the law recognized in the husband such an interest in his wife’s widowhood as to make it lawful for him to restrain her from making a second marriage by imposing a condition that on such marriage any provision he may have made for her shall cease. Similar in purport, though not so plainly stated, is the language of the Massachusetts court in Knight v. Mahoney, supra, which disapproves of the earlier Massachusetts case of Parsons v. Winslow, 6 Mass. 169, 4 Am. Dec. 107, cited by the solicitor for the widow. Chief Justice Gibson in *97Com. v. Stauffer, 10 Pa. 350, 51 Am. Dec. 489, though dealing with a cáse of real property, expressed the thought which these courts I think had in mind, in his characteristically vigorous manner, as follows: •
“It would be extremely difficult to say, why a husband should not be at liberty to "leave a homestead to his wife, without being compelled to let her share it with a successor to his bed, and to use it as a nest to hatch a brood of strangers to his blood.”
■ Not only was the validity of the condition when applied to the widow expressly recognized' by the Chancery Division in Allen v. Jackson, decided in 1876, 1 Ch. Div. 399, but for the 'first time in England a similar condition imposed upon a widower was likewise decided to be good. There áre "other authorities also which with like impartiality apply the same rule to husbands as to wives, and I can see no reason why it should not be so.
The solicitor for the widow assails "the condition also as being sirtiply in terrorem, there being as he contends no remainder over in case of breach except to the child who would take without testamentary provision as only heir at law — and such being so, there is no remainder over which the law will recognize" as rendering the doctrine of in terrorem inapplicable.' To this it is first to be said that there is authority to the effect that as to a widow "án exception to the doctrine is held to prevail. Chapin v. Cooke, et al., supra. But it is not necessary to examine the correctness of the rule laid down in that case, because as I construe the will there is a remainder over in case of the widow’s remarriage not only to the child, but as well to the father, etc. If the child should die, and the widow should then remarry, the fund goes to the ultimate remaindermen, for, though this contingency is not specially provided for, yet I have no doubt such is the meaning of the will. The case of the child’s dying after the remarriage of the wife is explicitly provided for by passing the fund over to the ultimate remaindermen. So that there is in fact a remainder over after the breach of condition by the wife and therefore the sort of case which calls for a discussion of the doctrine of in terrorem is not presented by the facts.
The solicitor for the wife argues that the validity or invalidity of a condition of this sort ought to be made to turn on whether the welfare of the child would be better served *98by holding the condition good under the facts of a given case,, or not good. In this particular case it is argued that inasmuch as the child is assured in any event of one-half of the income, it must be supposed that her father considered he had amply provided for her welfare regardless of whether his wife should ever remarry or not, and so under these facts there is no reason to hold that the condition against remarriage is void. The difficulty with this argument is that it finds no countenance in any of the authorities, and it assumes that consideration for children supplies the only rationale for the rule which subjects the widow to the burden of the. condition, an assumption which I have above endeavored to show is not warranted. Whatever the reasons may have been which occasioned the adoption of the rule to the effect; that a condition against remarriage is valid as to widows, it appears to be firmly fixed in the law uninfluenced by any such consideration as this last contention suggests. If the contention were accepted, courts would be drawn into the forbidden field of determining in a practical way the extent that a testator’s provision for his child ought in reason to have gone, for an inquiry of this nature would first have to be made in each case as a condition precedent to the determination of the final question of whether public policy would regard the condition in a given case as valid, or otherwise.
Let a decree be prepared in accordance with the foregoing.