delivered the opinion of the Court.

The appellant claims his freedom under the will of John Bates, deceased. His claim is based on the following clause in the will, viz:

“ I also give to my son-in-law and daughter, in addition to the eight negroes formerly given, the following, viz: James, Wesley, John, Berry, Emily, Bill, Mariah, and her child; also, Stephen, Margaret, and Green, to serve them eight years, and then to be free.”

The appellant is one of the three slaves last named in the foregoing clause of the will; and the eight years mentioned, having expired after the death of the testator, he instituted this suit to obtain a decree for his freedom.

In the case of Allison, a man of color vs Bates, (6 B. Monroe, 78,) this Court has given a construction to the will, by which all the slaves above named are entitled to their freedom, making no discrimination between those last named and the others. The language used in the will is only susceptible of two interpretations. Either all the slaves mentioned are to be free after eight years service, or the latter three only. The former interpretation was adopted by this Court in the case of Allison. Stephen’s right to freedom is guarantied under either construction. But it is now urged that Margaret -and Green only are to be free, and that Stephen is included by the word “also,” in the previous class. • If it *601■'were admitted that the slaves named in the clause, nre, ' by a proper exposition of the will, divided into, two '-classes,'having different destinies, one continued in sla- ' very, and the other emancipated, still Stephen’s right to freedom would be unquestionable. The language used places Stephen, Margaret and Green together’. They • cannot be separated without manifest violence to the ' plain import of words. The sentence will not bear any ' other grammatical or sensible construction. But 'if ■'■any doubt existed as to the 'true_cohstruction of the will, that one should prevdiTwh'ich would be most fa- ’ vorable to human liberty.

Extraneous testimony has been introduced with a •view to establish the following facts : That the testator in his lifetime, had expressed a decided opposition to 'the policy of emancipation; that the slaves named in the foregoing clause in the will, with the exception of the two last were- so young that their services for eight years would amount to little more than a reasonable •compensation for their support during the time; and that the testator did not intend to emancipate Stephen 'by his will, but if-such is its effect, it results from the 'mistake or inadvertance of the draftsman, the testator '•having intended fo devise Stephen as a slave for life.

The admissibility of this evidence, and Its effect on the construction of the will, have to be considered. ■ In the case of Bates vs Allison, (manuscript opinion June 'term, 1847,) this Court said, that “extrinsic evidence of the condition and circumstances of the property devised, "and of the testator and devisees, is no doubt admissible, to aid in ascertaining the intention of the testator. And where, by the use of such testimony, an ambiguity not apparent upon the face of the wi]l,is shown tchexist, the 'same kind of testimony may be resorted to for its explanation. ' But proof that the testator had contempla- . ted a different disposition of his property, or that he was opposed to emancipation, or considered it impolitic, or the understanding of the writer of the will as to ‘its meaning, or what the testator intended, is not legal or admissible.”

The testimony adduced does not show the existence of any latent ambiguity in the will. The language used is not applicable to more than one object. There is no doubt raised by the testimony, as to the thing given, or the person of the beneficiary. The devise is freedom, the appellant is the person designated for its enjoyment. There is no fact proved, creating any doubt as to the meaning of the will, or the object contemplated by the testator regarding the language of the instrument, as the true medium through which a knowledge on these subjects is to be derived. The evidence tends rather to assail the instrument itself, by exhibiting circumstances and declarations of the testator, inconsistent with its provisions in reference to the appellant. Is such evidence admissible!

The following doctrines in relation to the construction of wills, are well established:

Where the instrument is free from ambiguity, and there is no imperfection or inaccuracy in its language, the testator’s intention is to be collected from the words used by him, and parol evidence is not allowable for the purpose of adding to, explaining, or subtracting •from it, or to raise an argument in favor of any particular construction : (1 Phil. Ev. 545; 8 Bingham, 244; Wigram on Ec. Evidence, 65.)

Extrinsic evidence of intention, is inadmissible for the purpose of supplying a devise, or any other material provision, omitted by mistake, or to superadd any qualification to the terms used, or to evince a mistake in writing the instrument: Webb vs Webb, (7 Monroe, 626; 11 Johnson, 201; 7 Serg. & Rawle, 111; 5 Madd, 364; Worthington on Wills, side page 77; 60 Law Library.)

As a general rule, parol evidence is only allowable where a latent ambiguity, or a resulting trust is manifested by the same kind of testimony, or where fraud has been practised on the testator; and in the latter case the evidence is admitted, not for the purpose of explaining or altering the writing, but to show that it is void: (1 Johnson’s Chan. Rep. 231; 1 Yeates, 432; 3 Serg. & Rawle, 604.)

It sometimes becomes necessary to inquire what estate, or quantity of interest in the subject matter devised the testator had at the time of making his will. If , a person devise an estate tor lite generally, without saying whether for the life of another, or the life of the devisee, evidence is admissible to show what interest the testator had in the premises ; for if he owned the premises in fee, the devisee would have an estate for his own life, but if he was tenant for life of a third person only, then the devisee would take that interest. Or, if a testator bequeath such a sum in a particular stock, it will be a specific legacy if he has the stock; but otherwise if he has it not: (Phil. Evidence, 544.)

The attempt in this case is to show by evidence aliunde, a different intention on the part of the testator in reference to the disposition of Stephen, from that manifested by the language of the will. The person by whom it was written, does not prove that it is written otherwise than the testator directed it to be done; on the contrary, he proves that the name of Stephen is inserted exactly in that part of the will where the testator instructed him to place it. But he also proves that it was not the intention of the testator, as he understood it, to emancipate Stephen.

Now it is obvious from this evidence, that its object is, not to prove a mistake in writing the will, but an intention on the part of the testator, in conflict with that which- is collected from the words of the will itself. ' It is clearly inadmissible for this purpose, either according to the case of Bates vs Allison before mentioned, or the doctrines and principles on this subject, adverted to in this opinion.

Indeed the authorities go much further, and show that parol evidence is inadmissible for the correcting of a mistake in a will devising real estate, even when produced by the unskilfulness or inadvertence of the draftsman, and by which the avowed intention oí the testator has been defeated. The case of Webb's heirs vs Webb, (7 Monroe, 626,) fully asserts this principle. In the case of Iddingsvs Iddings, (7 Sergt. & Rawle, 111,) the same doctrine is maintained; and although the scrivenor, by *604inserting in the-will, a word, the meaning of which- he did not understand, thereby defeated t.he intention of the testator, yet parol evidence was rejected,-when offered for its correction. The modern English cases referred to in Worthington on Wills, side page 77, published in the 60th vol- of the Law Library, recognize the same doctrine. To allow parol evidence to be introduced to vary the meaning of wills, by showing a different intention on the part of the testator, from that, expressed in the will, would be in violation of a well, established rule of evidence, and in direct opposition to . the statute requiring wills devising real estate and slaves to be in writing. And although evidence of the situation and circumstances of the testator and devisee, and of the character and nature of the subject matter of the devise may be adduced, to enable the Court, by placing itself in the position of the testator,, to arrive at a correct knowledge of his meaning and intention, and thereby place a proper construction on , the will, yet such evidence is admissible only for the purpose of aiding in the construction of the language used, provided there, is sufficient indication of intention appearing on the face of the will, to justify its application. The enquiry must be confined to the meaning of the words used;, and hence, all extrinsic evidence tending to prove, not What the testator has expressed, but what he intended' to expresses inadmissible: (Wigram on Extr. Ev. 65.)

We are of opinion, therefore, that according to any allowable construction of the will of John Bates, deceased, Stephen, the plaintiff in error, is entitled to his freedom.

Wherefore, the decree of the Court below dismissing the bill of complainant, is reversed, and cause remanded for a decree in conformity with this opinion.