Those parts of the will of David Lawson that are so obscure as to require the advice of this court relate to the bequests to the Freedmen’s Association and to Fairfield to be used as he pleases.
1. Who can take the legacy payable to the proper officers of the “Freedmen’s Association”? We cannot advance a single step toward the solution of this question unless resort may be had to parol evidence, because the record shows that there was no such organization or corporation in existence as the Freedmen’s Association at the date of the execution of the will; and this expresses but a small part of the difficulty, for the further finding is that except a single item of parol evidence, the admissibility of which is one of the questions reserved, there was absolutely no evidence of any kind to identify the object of the testator’s bounty.
The evidence' in question consisted merely of the oral instructions given by the testator to the scrivener, Fair-field, “ that he wanted to give the income of the projJerty in question in trust for the education of the freedmen; that there was a Freedmen’s Association organized by the Methodist Church people located in Cincinnati, Ohio, and that he wanted it payable to the officers of that association.”
Now it is very common to admit parol evidence in cases for the construction of wills. The difficulty here is not owing merely to the fact that the evidence is oral, but to its relation to the written words of the will. The law is imperative that the entire will must be in writing, and herein are found the rules and limitations that must be applied to such evidence. The intent must in every case *509be drawn from the will, but never the will from the intent. The test therefore to be applied in all eases where evidence like that under consideration is tendered, is, whether there appears on the face of the will sufficient indication of intention to justify the application of the evidence. The words of the will are so controlling that if they apply with exactitude to one person, such person will take the legacy, although parol and extrinsic evidence might make it perfectly clear that another person less exactly described was the one intended.
This principle was applied by this court in the recent case of Dunham et al. v. Averill et al., 45 Conn., 61, where the legacy was to “ The American and Foreign Bible Society,” and it appeared that that society was one mainly supported by the Baptist denomination; but that there was another society supported by the Congregational and Presbyterian denominations, named the “ American Bible Society,” sometimes called “The American and Foreign Bible Society,” and that the testator’s sympathies and preferences were all with the latter; and evidence was offered that while the will was being drawn the testator said to the scrivener that he wished to give the money to the Bible Society sustained by the Congregationalists and Presbyterians; that he was not sure as to its corporate name, but believed it to be “ The American and Foreign Bible Society ”; but the evidence was held not admissible. So it has been uniformly held that parol evidence cannot be received to correct a mistake in the will. Avery v. Chappel, 6 Conn., 270; Comstock v. Hadlyme Ecc. Society, 8 id., 254; Tucker v. Seamen's Aid Society, 7 Met., 188; Jackson v. Sill, 11 Johns., 201.
The principle we are contending for is also applied in another class of cases, where parol and extrinsic evidence is admitted. I refer to the rule derived from the maxim, “Salsa demonstrate non noeet, cwm de corpore constat where the office of the parol evidence is to reject that part of the description which is false, but in such case it is indispensable that enough remains in the words of the will to show plainly the intent, but in no case can any words be added to the description.
*510Another prominent rule is, that when the question is one of construction the parol or extrinsic evidence must be ancillary to a right understanding of the language of the will; hence all direct evidence of intention as contra-distinguished from evidence to show the meaning of the written words in the will is inadmissible. This rule is well illustrated by the case of Goblet v. Beechey, given at length in the second American edition of Wigram on Extrinsic Evidence, p. 287, Appendix, and also briefly reported in 3 Simons, 24. Nollekins, the sculptor, by a codicil to his will desired that “ all the marble in the yard, tools in the shop, bankers, mod, tools for carving, &c., should be the property of the plaintiff. A lady who was an attesting witness was offered to prove that before she subscribed her name she read the codicil in the hearing of the testator, and when she came to the word “mod ” she asked him what he meant by it, and he replied “models.” Sir John Leach, Yice Chancellor, held the testimony inadmissible, but allowed an inquiry as to the meaning of the term itself from the testimony of sculptors. See also cases referred to in 2 Phillips’s Evidence, (Cowen & Hill’s notes) p. 754.
So far the rules referred to, if applied to the evidence in question, rigidly exclude it. Is there then any exception or additional rule under which it may be received ? The case shows that it was sought for the purpose of ascertaining the beneficiary, to prove the specific intention of the testator by his oral declarations to the scrivener who drew the will. There is only one rule that can be invoked as applicable to such a case. This is stated very clearly by Lord Abinger, Chief Baron, in Hiscocks v. Hiscocks, 5 Mees. & Wels., 363, whose opinion, Redfield says, in his Treatise on Wills, vol. 2, p. 566, is universally admitted to have settled the law that such evidence is only admissible in the one instance there stated, namely, “ where the meaning of the testator’s words is neither ambiguous nor obscure, and where the devise is, on the face of it, perfect and intelligible, but, from some of the circumstances admitted in proof, an ambiguity arises as to which of the two or more things, or which of the two or *511more persons,, each answering the words in the will, the testator intended to express. Thus, if a testator devise his manor of S. to A. B. and has two manors of North S. and South S., it being clear he means to devise one only, whereas both are equally denoted by the words he has used, in that case there is what Lord Bacon calls ‘ an equivocation,’ that is, the words equally apply to either manor, and evidence of previous intention may be received to solve this latent ambiguity ; for the intention shows what he meant to do; and when you know that, you immediately perceive that he has done it by the general words he has used, which, in their ordinary sense, may properly bear that construction. It appears to us that, in all other cases, parol evidence of what was the testator’s intention ought to be excluded, upon this plain ground, that his will ought to be made in writing, and if his intention cannot be made to appear by the writing, explained by circumstances, there is no will.”
Now it seems to us that under this rule the proposed evidence cannot apply, because the words of the will describing the beneficiary do not apply equally to two or more, “ each answering to the words of the will.” On the contrary the words used are not applicable to any known organization, either voluntary or incorporated. Such in substance is the finding. When therefore we learn from the parol evidence what the actual intent was, we do not “immediately perceive that the testator has effectuated his intent by the general words he lias used; ” on the contrary, the effect of the evidence in this case is rather to increase the mystery that hangs over the words in the will. The name “ Freedmen’s Association” in itself considered would naturally import an association composed of freedmen, as the names “ Lawyers’Association,” “Doctors’Association,” “Farmers’ Association,” would indicate the membership of each.
It is very strange, if the testator gave such instructions to the scrivener as the evidence indicates, that no one of the prominent features of his description should find its way into the will as written. The prominent things in his description were, the religious body that organized the asso- *512. ciation and its location at Cincinnati, Ohio, but of these things the words of the will are silent, and it does not appear how or why the words “Freedmen’s Association” alone were used; there was no discussion concerning the name; no suggestion that the name used would be sufficient, nor that the Cincinnati society had ever been so called. As the case stands upon the record the instructions given by the testator were not carried into effect by the scrivener, and the court has no power to correct the mistake, as it would upon like evidence correct a mistake in a contract. We should be virtually making a will as to the beneficiary from the actual intent proved only by parol.
We have not deemed it necessary to review the numerous eases bearing upon this question. While there is now substantial harmony among the courts concerning the abstract principles that apply, there is, it must be confessed, considerable diversity in their application. We have therefore preferred to test the somewhat extraordinary features, of this case by a pretty strict application of the principles of evidence and construction, and our conclusion is that the parol evidence cannot be received for the purpose of showing that the legacy in question is payable to the officers of “The Freedmen’s Aid Society of the Methodist Episcopal Church located in Cincinnati, Ohio; ” and it is pleasant to know that this society will not be disappointed by this result, for it appears that, although they well knew the terms of the will and the fact of the pending litigation, yet they have never claimed the legacy in question.
2. The next question is, whether the trustees named in the will, (or others to be appointed by the court for the purpose,) can rightfully use and appropriate the income for the education of the freedmen as constituting a definite class of persons?
It is contended that, as the purpose and object of the bequests under consideration are the education of the freedmen, who constitute a definite class of persons, the charity will not be suffered to fail for the want of a competent agent to administer it. If this were the only difficulty in *513the case it might easily be overcome, for there is no doubt that the court can supply the want of a trustee. There is in fact no such want here. Corbin and Fairfield are named as trustees. But in each of the clauses where bequests are rnade for the education of freedmen the trustees have no discretion given them in the will. On the contrary, all discretion is taken away by tbe express direction to pay the income over to the proper officers of the Freedmen’s Association. When they have performed this duty there is nothing left for them to do under the will, and the court cannot prescribe an additional duty without in effect making an addition to the will. But it is argued that the freedmen are the cestuis que trust, and that if the trustees only pay the money for their education it effectuates the intention of the testator as indicated in the will; that the certainty required is only to point out the class, no matter how indefinite may be the particular recipients of the benefit within that class. It is found that the term “ freedmen,” as used in the will, refers to that class of persons who were emancipated during the late civil war and their descendants. As matter of common knowledge we may be permitted to say that the numbers composing this class are now about six millions. Of all these, only a very few individuals could by any possibility receive any of the benefits contemplated by the will. It is not within the range of probability that different individuals or corporations, separately charged with the duty of disbursing the testator’s bounty, would so perform it as to benefit the same individuals of the class. A change therefore in disbursing agents, or a change in the mode of selecting beneficiaries, not provided for in the will, constitutes in effect a change of the bequest. Hence, in addition to a definite class it is indispensable that the will itself should prescribe some mode of selection, or give to some person a discretionary power to select; in short, a will must be executed in the way and manner which the testator provides, and if, owing to the indefiniteness of the object or the mode provided, this cannot be done, then the subject of the trust is not disposed of, but results to the benefit of those tq *514whom the law gives the property in the absence of a valid will.
The cogent reasoning of Btxskibk, J., in giving the opinion in Grimes's Exrs. v. Harmon et al., 35 Ind., 198, furnishes most ample support for the positions we have taken in this discussion. In that case the bequest was “ to the Orthodox Protestant Clergymen of Delphi, and their successors, to be expended in the education of colored children, both male and female, in such way and manner as they may deem best, of which a majority of them shall determine; my object being to promote the moral and religious improvement and well-being of the colored race.” The court, after a most exhaustive and able review of all the authorities, held that the legacy was void for vagueness and uncertainty; placing the decision upon the following propositions, among others,—that the testator intended that his beneficiaries should be selected from the children of the colored race residing within the United States; that the persons composing the class were very numerous and as each one had a beneficial interest in the fund it would be utterly impossible to execute the trust; that under the will as construed by the court the trustees had no power or discretion to select the beneficiaries from the class designated; that there is no difference in principle whether a devise be immediate to an indefinite object, or to a trustee for the use and benefit of an indefinite object; that if it beymmediate, to an indefinite object, it is void, and if it be a trust for an indefinite object, the property that is the subject of the trust is not disposed of, and the trust results to the benefit of those to whom the law gives the property in the absence of any other disposition of it; and that if a charity does not fix itself on a particular object, but is general and indefinite and no plan or scheme is prescribed and no discretion is given in the will to select the beneficiaries, it does not admit of judicial administration.
3. The remaining question relates to the construction of item second of the will, which reads as follows: “I do give, devise and bequeath unto my executor hereinafter *515named, all of my real estate in whatever place situated, the same to be sold by him after my decease, and the proceeds to be held by him in trust for the education of the freedmen, and the annual interest and income arising from the same to be paid by him to the proper officers of the Freedmen’s Association, or to be disposed of and used as he pleases.”
The construction of this clause of the will, we think, must depend on the question whether the idea of a trust adheres to the proceeds in the hands of the executor to the last, whether he pays it over to the officers of the Freedmen’s Association or exercises his own pleasure and discretion as to whom it may be paid to, or whether the trust drops out altogether at the commencement of the alternative clause, so that the bequest was either a trust or no trust at the will of Fairfield.
We think the first is the better construction. In the first place we think, if the testator had intended a personal gift to Fairfield, as he was a lawyer, and was employed to draft the will, and was made executor to administer it, that if he had so understood it the terms of the will would have been more explicit, for it would have occurred to both that such a gift, covering as it did the principal part of the estate, would necessarily excite the suspicions of the disinherited heirs, who would desire to defeat it. And in the next place, we think the language used indicates that the testator intended to stamp all this property with a permanent trust. The bequest is not to Fairfield by name, but to his executor, and the testator expressly says that the proceeds of the sale of all the real estate are to be held in trust by the executor, and then follows a statement of the purposes.
Upon this construction is the trust one that can be enforced ? Lord Lastglale, in Knight v. Knight, 3 Beavan, 148, 174, defining the certainty required to create a valid trust, says:—“ Any words by which it is expressed, or from which it may be implied, that the first taker may apply any part of the subject to his own use, are held to prevent the subject of the gift from being considered certain. And a *516vague description of the object, that is, a description by which the giver neither clearly defines the object himself, nor names a distinct class out of which the first taker is to select, or which leaves it doubtful what interest the object or class of objects is to take, will prevent the objects from being certain within the meaning of the rule.” In 1 Jar-man on Wills, (5th Am. ed.,) p. 680, it is said that “ if the gift be expressly in trust, though to be disposed of in such manner and for such purposes as the devisees think fit, they are trustees, and the beneficial interest results to the heir or next of kin, and a gift ‘ to be expended and appropriated in such manner as the donees or a majority of them shall in their discretion agree upon,’ would probably without the words ‘in trust’ produce the same result, for technical language of course is not necessary to create a trust. It is enough that the intention is apparent.” Citing Fowler v. Garlike, 1 Russell & Mylne, 232; Buckle v. Bristow, 10 Jur. (N. S.), 1095; Gibbs v. Rumsey, 2 Ves. & B., 292. See also Wheeler v. Smith et al., 9 How., 79. In Morice v. Bishop of Durham, 10 Ves. Jr., 526, Lord Eldon says:—“If a testator expressly says he gives upon trust, and says no-more, it has been long established that the next of kin will take. Then, if he proceeds to express the trust, but does not sufficiently express it, or expresses a trust that cannot be executed, it is exactly the same as if he had said that he-gave upon trust, and stopped there.”
We therefore advise the Superior Court that the oral declarations are inadmissible for the purpose claimed; that “ The Freedman’s Aid Society of the Methodist Episcopal Church located at Cincinnati, Ohio,” cannot take the legacy given to the Freedmen’s Association; that the bequests to the freedmen as a class are void for uncertainty; and that Fairfield takes nothing under the will.
In this opinion the other judges concurred.