DUNKLIN, J.
The only controverted issue involved upon this appeal is the proper construction of the will of R. L. Hagood. The conclusion of the majority is, substantially, that as the language of the will is of itself clear and free from ambiguity, ail agreed facts shown in the statement of facts should be excluded from consideration in arriving at the testator’s intentions, because such evidence would tend to vary the express terms of the will.
Some of the quotations shown in the opinion of the majority to the effect that a will cannot be reformed to correct a mistake made by the testator in its execution are inapplicable here and tend to confuse rather than to elucidate the issue, since appellants did not present that issue, but elected to stand upon the will as written, interpreted in the light of facts dehors the will shown in the agreed statement of facts.
All the authorities agree upon the general rule announced in the quotations in the opinion of the majority that parol evidence will not be -permitted to contradict, add to, or explain the terms of a will which are free of ambiguities or obscurities. But there is another rule of construction equally as well established, that the terms of a will or contract, which, standing alone, are definite and certain, may be shown b3r parol evidence to be ambiguous and the ambiguity may be removed by evidence of like character. 3 Jones on Evidence, § 472.
“Ambiguity is defined as duplicity; indistinctness; an uncertainty of meaning or expression used in a written instrument.” 1 Words and Phrases, p. 199.
“A ‘latent’ ambiguity, as defined by Lord Bacon, is ‘that which seems certain and without ambiguity for anything that appeareth upon the deed or instrument, but there is some collateral matter, outside of the deed, that breedeth the ambiguity.’ ” 3 Words and Phrases, p. 31.
Prof. Wigmore, in volume 4, § 2462, of his able and exhaustive work on Evidence, traces the history of the general rule forbidding the introduction of parol evidence to vary the terms of a written instrument. He notes that until the middle of the fifteenth century land could not be alienated by will at all, and that for a long time after that period “through the lack of a liberal and sympathetic search for the testator’s meanings, the spirit of rational interpretation was hindered.” 1-Ie notes further that a like strictness of interpretation of all legal documents continued until about the beginning of the eighteenth century, when there set in “a growing spirit of liberality.” And in section 2470 he notes the stages of progress of liberalism in the construction of wills, both in England and the United States, from Lord Coke’s time down to the present, giving the following quotations of authorities:
“1831, Sir James Wigram, Y. G., Extrinsic Evidence in Aid of the Interpretation of Wills, Proposition V: ‘For the purpose of determining the object of a testator’s bounty, or the subject of disposition, or the quantity of interest intended to be given by his will, a court may inquire into every material fact relating to the person who claims to be interested under the will, and to the property which is claimed as the subject of disposition, and to the circumstances of the testator and of his family and affairs, for the purpose of enabling the court to identify the person or thing intended by the testator, or to determine the quantity of interest he has given by his will. The same (it is conceived) is true of every other disputed point, respecting which it can be shown that a knowledge of extrinsic facts can, in any way, be made ancillary to the right interpretation of a testator’s words.’
“In 1833, Parke, J., in Doe v. Martin, 4 B. & Ad. 770, 785: ‘It may be laid down as a general rule that all facts relating to the subject-matter and object of the devise * * * are admissible to aid in ascertaining what is meant by the words used in the will.’
“In 1842, Sugden, L. C. in Attorney General v. Drummond, 1 Dr. & W. 356 (interpreting a deed containing the words ‘Christian’ and ‘Protestant Dissenter’): ‘The court is at liberty to inquire into all the surrounding circumstances which may have acted upon the minds of the persons by whom the deed or will (it matters not whether it was one or the other) was executed. * * * The court therefore has not merely a right, but it is the duty to inquire into the surrounding circumstances, before it. can approach the construction of the instrument itself.’ _ .....
“1886, Blackburn, J., in Allgood v. Blake, L. R. 8 ICxeh. 160: ‘The general rule is that in construing a will the court is entitled to put itself in the position of the testator and to consider all material facts and circumstances known to the testator with reference to which he is to be taken to have used the words in the will, and then to declare what is the intention (i. e. sense) evidenced by the words used, with reference to those facts and circumstances which were (or ought to have been) in the mind of the testator when he used those words. As said in Wigram on Extrinsic Evidence: ‘The question in expounding a will is, not what the testator meant —as distinguished from what his words express —but simply what, is the meaning of Ms words.’ But we think that the meaning of words varies according to the circumstances of and concerning which they are used.’ ”
And in 3 Jones on Evidence, § 477, the author, after referring to the former rule of strict construction of wills, says:
“The rule now is unquestioned that extrinsic evidence in aid of the interpretation of wills is admissible for the purpose of showing the object of the testator’s bounty, the property devised, and the quantity of interest intended to be given. Evidence may be received as to every material fact relating to the person who claims under the will, and to the property devised, as to the circumstances of the testator and his family and affairs, so as to lead to a correct decision of the quantity of interest the claimant is entitled to by the will. This is true as to every disputed point respecting which it can be shown that a knowledge of extrinsic facts can aid in the right interpretation of the will.”
In 4 Wigmore on Evidence, §§ 2471 to 2477, inclusive, the author announces the general rule, which he says has never been questioned, and exceptions thereto, that while extrinsic facts and circumstances are admissible to discover the intent of the testator, his declarations of such intention are rejected because the tendency of such proof would be 'to overthrow the written instrument, or else establish by parol evidence a transaction which is required to be shown by an instrument in writing. See, also, 1 Greenleaf on Evidence, §§ 290, 291; 2 Underhill on Wills, § 90S.
The most frequent application of the rule allowing evidence of facts dehors the will occurs when the language, employed to designate the property devised or the objects of the testator’s bounty, is applicable alike to more than one piece of property or more than one person, and such evidence is introduced to identify the property or person intended.
And in section 289, 1 Greenleaf on Evidence, the author states his individual belief to be that in all other cases than those parol evidence to show the testator’s intention should be excluded under the general rule; but he cites no decision supporting a doctrine so broad in scope as that.
The authorities abound with statements that the court should place itself in the position of the testator and construe his will in the light of his surrounding circumstances, to the end that his intention may be discovered and carried out, if that purpose can be accomplished without violating some rule of law. 1 Greenleaf on Evidence, § 289; 1 Jar-man on Wills, § 422; 2 Underhill on Wills, § 909. In Finley v. King’s Lessee, 3 Pet. (28 U. S.) 377, 7 L. Ed. 712, Chief Justice Marshall said:
“The intent of the testator is the cardinal rule in the construction of wills; and if that intent can be clearly perceived, and is not contrary to some positive rule of law, it must prevail; although in giving effect to it, some words should be rejected, or so strained in their application, as materially to change the literal meaning of the particular sentence.”
The statement occurs frequently in the text-books that wills are construed in accordance with the same rules that apply to deeds and contracts. But in Lambert v. Paine, 3 Cranch, 97-138, 2 L. Ed. 377, 391, the following was said:
“Wills are expounded more favorably, to carry the intent of the testator- into effect, than conveyances at common law, which take effect in the lifetime of the parties; wills being frequently made by people enfeebled by age or indisposition, and without the aid of counsel learned in the law. Therefore, words no.t so technical for the purpose, have in a great variety of cases, for above 100 years, been construed by .the judges, to carry a fee, which would not do so in a deed.”
Other cases of like import are cited in Hancock v. Butler, 21 Tex. at pages 813, 814 (orig. edition), bot. page 726 (later edition).
The authorities are uniform upon the proposition that when a will is made, the presumption obtains that the testator intended, thereby to dispose of all his property, and no presumption will be allowed that he intended to die intestate as to a part of his estate, if the words used in the will may fairly carry the whole. 1 Underhill on Wills, §. 464; Hardenbergh v. Kay, 151 U. S. 112-120, 14 Sup. Ct. 305, 38 L. Ed. 93.
By the express terms of the will in controversy in this suit, R. L. I-Iagood devised his entire estate. He left the will unchanged at the time of his death, although he survived J. O. Hagood, one of the beneficiaries originally intended, by more than 3 years. One of the provisions of the will limits the action of the county court with respect to his estate to the probate of the will, and the return of an inventory and appraisement of the property belonging to his estate — a provision wholly inconsistent with intestacy of one-half of the estate; and the devise of the entire estate which testator might own at the date of his death to the beneficiaries named, to the exclusion of all other kindred, indicates an intention that such other kindred should be excluded from a participation in said estate, especially as the will was never changed after.the death of J. O. Hagood. The facts so enumerated create an ambiguity and uncertainty as to whether or not the testator intended that the will should cast the whole estate upon the survivor -of those beneficiaries in the event one should die before the death of the testator. And independent of testator’s instructions to his attorneys with respect to how he desired his will to be drawn, and independent even of the
agreement as to tlie testator’s intention in fact, tlie facts enumerated above, considered in connection with the further facts and circumstances shown in the record, including the joint acquisition by the three brothers of practically all the property owned by them, the fact that R. L. and J. O. Hagood never married, that they lived in the same home with R. B. Hagood and his family for several years prior to the execution of the two wills and up to their respective deaths, and were cared for by R. B. Hagood and his family during their long illness from consumption, with which both died, in connection with an absence of proof of any facts such as financial needs or special affection, tending to show why testator might probably have desired his other kindred to share in the estate in the event of the predecease of one of the beneficiaries, all lead irresistibly to the conclusion, in the opinion of the writer, that in making the will in question the testator intended that, in the event of the death of One of said beneficiaries prior to the death of the testator, the survivor of the two should take the whole estate, if he should survive the testator also. If the will can be so construed as to give effect to that intention without doing violence to the language therein employed, interpreted in accordance with the law controlling in the construction of wills, then that interpretation should be adopted and the will enforced accordingly; otherwise it should be held that the testator died testate as to one half his estate and intestate as to the other half. And that is the next question to be considered.
Authorities, practically without number, might be cited announcing the general rule, that in construing wills the terms used must be interpreted in their plain and ordinary sense. But lexicographers often give different meanings of the same words according as they are einployed in common parlance, or in theology, law, etc., and it cannot be doubted that words used in deeds or wills which have a well-fixed meaning in law should be given that construction by courts rather than some other meaning not popularly understood by those not versed in the law. Eor illustration: In early times a conveyance “to A. and his heirs’’ and one “to A. for life, remainder over to his heirs,” were substantially the same. In either ease A. acquired only a life estate in the property, and after his death his heirs took the remainder under the conveyance. In other words, the word “heirs” was a word of purchase according to its plain meaning to the lajmlan. But later, by judicial construction, it was interpreted as a word of limitation indicating the quantity of estate conveyed, which would be a title in fee simple, or fee tail, as the case might be, and that meaning was finally fixed and established by the decision in Shelley’s Case. See 2 Reeves on Real Property, § 892.
If the beneficiaries named had both survived the testator, then under the common law the will vests in them a joint tenancy, and under the common law, in case of the death of one before the decease of the testator, the will vests the whole estate in the survivor, “because,” as said in 1 Jarman on Wills, §§ 340, 341, “as joint tenants take per my et tout, or, as it hás been expressed, ‘each is to take of the whole, but not wholly, and solely,’ any one of them existing when the will takes effect will be entitled to the entire property.” If one of the intended joint tenants dies before the date the devise becomes effective, then no- joint tenancy is created, and it is perfectly clear that the underlying and controlling reason for so construing the will as to cast the whole estate upon such surviving devisee is the purpose to give effect to the intention of the testator that such should be the result. And the rule applying in case of a devise which would create a joint tenancy if both devisees survive the testator, furnishes another instance of giving effect to the testator’s intention at the sacrifice of the plain meaning of words in that a devise to two or more is construed to mean a devise to one only, viz., the survivor.
By article 2471 of the statutes of descent and distribution of Texas, it is provided that where two or more persons own an estate jointly and one dies before a severance, his interest in said estate shall not survive to the remaining joint owner or joint owners, but shall descend to the heirs of the decedent in the same manner as if his interest had been severed before his death. That is a statute of inheritance after title has vested in two or more joint tenants, and does not necessarily control in the present suit because J. O. and R. B. Hagood never became joint tenants. But several states have enacted statutes, which further provide that upon the death of an intended joint tenant by will before the death of the testator, the interest which he would have acquired under the will had he survived the testator shall pass to his heirs.
If, by reason of our statute, it can be said that the will itself does not show an intention on the part of the testator that the beneficiary surviving at the death of the testator should take the whole estate, yet that intention clearly appears from extrinsic facts and circumstances noted already. And if the intention of the testator is to be the controlling factor, regardless of the specific terms of the will, in the one instance, why not in the other?
Under the common law, a devise or conveyance to two or more with a provision denoting the share each is to take, would, upon the face of the instrument, constitute the takers tenants in common. And Mr. Jarman in his work on Wills, § 341, following his statement of the rule obtaining if the devise is to- joint tenants, says:
“While it is equally clear that if the dev-isees or legatees in any of these cases had been
made tenants in common, the failure of the gift as to one object would not have entitled the other to the whole by the mere effect of survivorship. Where, however, the devise or bequest embraces a fluctuating class of persons, who, by the rules of construction, are to be ascertained at the death of the testator; or at a subsequent period, the decease of any of such persons during the testator’s life will occasion no lapse or hiatus in the disposition, even though the devisees or legatees are made tenants in common, since members of the class antecedently dying are not actual objects of gift.”
In passing, it is proper to note that to do this, the designation of several beneficiaries by a collective term is construed as meaning one of that class only. It is quite evident that the discussion by the author, of the rules in the two instances referred only to wills showing plainly by their own language the creation of the character of estates devised and without reference to cases in which the intention of the testator, in case of ambiguities in the will, might be shown by extrinsic evidence, a .subject discussed in later chapters of the work.
Many authorities are cited in the opinion of the majority defining a class, substantially, as a body of persons which can be designated by the same general term, such as “children,” “grandchildren,” “nephews,” “brothers,” or “sisters.” In some of the definitions found in the authorities, it is expressly stated that the designation of the beneficiaries by name, or a statement of the share each is to take, will destroy the class. It is also a rule that a class may be created by the inclusion of a part only of those standing in the same relation to the testator. The definitions of “class” referred to are given in numerous cases where the beneficiaries named did constitute a class within the meaning of those definitions. If those decisions could be construed as holding that persons not coming within those definitions cannot constitute a class, then to that extent they would be mere dicta and not necessarily authoritative. Except as announcing general principles, little aid can be procured from adjudged cases with their ever differing facts. In Smith v. Bell, 6 Pet. 68, 79, 8 L. Ed. 322, Chief Justice Marshall said:
“The construction put upon words in one will has been supposed to furnish a rule for construing the same words in other wills, and thereby to furnish some settled and fixed rules of construction which ought to be respected. We cannot say that this principle ought to be totally disregarded; but it should never be carried so far as to defeat the plain intent, if that intent may be carried into execution, without violating the rules of law. It has been said truly (Gulliver v. Poyntz, 3 Wils. 141) ‘that cases on wills, may guide us to general rules of construction; but, unless a case cited be in every respect directly in point, and agree in every circumstance, it will have little or no weight with the court, who always look upon the intention of the testator as the polar star to direct them in the construction of wills.’ ”
That statement has been quoted with approval in Colton v. Colton, 127 U. S. 300-309, 8 Sup. Ct. 1164, 32
L.
Ed. 138, and in Adams v. Cowen, 177 U. S. 471, 475, 20 Sup. Ct. 668, 44 L. Ed. 851.
If the two beneficiaries named in the will in controversy had been designated as those brothers with whom the testator made, his home, without naming them, then they would have constituted a class within the meaning of the definitions of that term usually given, and would have given the one surviving when the will took effect the whole estate. That the brothers named in the will were the ones with whom he made his home is admitted, but instead of identifying them by collective and descriptive terms, and in order to establish their identity beyond cavil, he named them, and, by so doing, according to the usual definitions of a “class,” he metamorphosed what would have resulted in a devise of the whole estate to R. B. Hagood into a devise of
only
one-half of it, and that too contrary to the testator’s intention in fact. Viewing such a construction in the light of fundamenal principles, the inquiry would not be inappropriate: When this magic in mere names of persons?
If the question, whether or not the will in the present suit, was to the testator’s two brothers as a class should be determined by precedents, then only those cases should be considered which involved devises or legacies to persons identified in such a manner as not to come within the usual definitions of “class.” Applying that test, the decided weight of such precedents sustains the contention of the appellant in the present suit, assuming that the cases which are considered more nearly in point are those specially noted in the opinion of the majority, and without attempting an examination of the numerous decisions otherwise referred to. In the case of Herzog v. Title Guarantee & Trust Co., 177 N. V. 86, 69 N. E. 283, 67 L. R. A. 146, by a residuary clause of the will, the remainder of the estate after satisfying other bequests was given to testator’s four children, who were all named, share and share alike; and it was held that the gift was not to a class.
In re Hittell’s Estate, 141 Cal. 432, 75 Pac. 53, 54, the gift was to two sisters, as tenants in common, who were not related to testator, but with whom he lived for 10 years prior to his death. It was held that the gift was not to a class, and that the death of one of the beneficiaries prior to the death of the testator caused a lapse of one-half of the estate. The further question whether or not the devise was intended to be a devise to a class was discussed and the conclusion reached that the circumstances relied on to show such intention were not sufficient to accomplish that result, thus leaving the inference to be drawn that if such intention was so shown, the decision might have been different.
In Melton v. Sellars (Ky.) 181 S. W. 346, by the Court of Appeals of Kentucky, the beneficiaries were identified by terms of gen-
aral description, and it was held that the gift was to a class, tested by the usual definitions of what is necessary to constitute a class.
In Ritch v. Talbot, 74 Conn. 137, 50 Atl. 42, the gift was to testator’s two brothers, Charles and Christopher Talbot, share and share alike. Charles Talbot, one of the brothers, died prior to the death of the tes tator. The court said:
“The bequest to the two brothers, Charles and Christopher Talbot, was one to them individually and not to them as a class. By operation of section 541 of the General Statutes, the legacy to Charles Talbot was in legal effect a gift to him, and to his issue in case he should not survive the testator.”
Hollowing are cases cited in appellant’s brief to support his contention that the gift was to him and J. O. Ilagood as a class:
In Bolles v. Smith, 39 Conn. 217, the gift was to “Frederic M. Smith, Valentine W. Smith, and Timothy W. Smith, sons of my late brother, Roswell Smith, and to their heirs and assigns forever, equally, the residue of my estate * * * ” upon the payment by them of a legacy to testator’s wife. The gift was held to be to a class because such appeared to be the intention of the testator as gathered from the whole will in the light of surrounding circumstances. And in the opinion in that case the court said:
“It is true that by the law of Connecticut the survivors, in ordinary cases of joint tenancy, do not take the whole estate. But the argument does not depend upon the jus acerescendi. If they are joint tenants, prima facie they take as a class. If they take as a class the lapsed legacy goes to the survivors.”
And tli© latter holding is supported by Lockhart v. Vandyke, 97 Va. 356, 33 S. E. 613.
In Warner’s Appeal, 39 Conn. 253, the remainder of the estate left after other devises was divided into four equal parts, one of which was given “to the sons of my two sisters, deceased, Henry S. and Charles K. Warner.” It appeared that the mother of each of said nephews had successively been the wife of the father of those nephews. One of those nephews having died prior to the death of the testator, it was held that the gift was to the three as a class, and the two survivors took the entire fourth interest that had been bequeathed to the three, because such appeared to be the intention of the testator, viewing the will as a whole.
In Jackson v. Roberts, 80 Mass. 546 (14 Gray), a certain residue of the estate was bequeathed to Elizabeth, Ellen, Edward, Grace, and Mary, five children of an adopted daughter of testatrix, share and share alike, the issue of any of said beneficiaries to take the share of their parents by representation. It was further provided in the will that in event any one of three of the children, Edward, Grace, and Mary, should die without issue, then the share of that one should go to the survivors of Ellen, Edward, Grace, and Mary, thus omitting Elizabeth, the issue of any one of those four to take by representation the share of their parent. It was held that the devise was to a class, and that the shares of Grace and Ellen, who died without issue and before the death of the testatrix, passed to the survivors. It is noted in the opinion that the fact that testatrix, by codicil to the will, expressly confirmed it after the death of Grace clearly indicates her original intention that the bequest was to the children as a class. It was further said that the mention by name of individuals who compose a class is not conclusive that a class was not intended, and in that connection occurs the following:
“And it is not to be doubted, that when the intention of survivorship is in any other way plainly shown by the will itself, or by the will and such evidence of extrinsic facts as is legally admissible for the purpose of showing it, such intention must prevail.”
(Numerous English authorities are cited in support of that announcement.
In Re lye’s Estate, 182 Mich. 699, 148 N. W. 727, the gift of the property in controversy was “to my sister Hattie Butterfield, and my brothers, Wesley and Dwight Skinner, to each an undivided one-third.” The testatrix was survived by the two brothers named, but her sister Hattie Butterfield died prior to her death. In holding that the devise was to the sister and two brothers as a class, the court said:
“The presumption is that a testator intended to dispose of his entire estate, and not to die intestate either as to the whole or any part thereof, and the will should be so construed, unless the presumption is clearly rebutted by the provisions of the will or by the evidence to the contrary. 40 Cyc. 1409; Toms v. Williams, 41 Mich. 552-565, 2 N. W. 814; Des Grand Champ v. Duflo, supra [169 Mich. 104, 135 N. W. 98]. Iu the instant case we have not only the presumption to aid us, but we have also, the express language of the testatrix, that she intended to give the rest, residue, and remainder of her estate to her sister and brothers. * * * We recognize the general rule contended for by appellee that, where property is given to several persons by name, to be equally divided between them, they take as tenants in common, and not as joint tenants, or as a class. But the authorities hold, we think, that this rule yields to a different construction when it plainly appears from the will that it was the intention of the testator that the survivors should take the whole. We think that this is such a case, and that for the reasons stated the trial court erred in its conclusion in reversing the action of the probate court.”
Numerous authorities are cited in the opinion in support of that construction.
In Re Dangdon’s Estate, 129 Cal. 451, 62 Pac. 73, the gift in controversy was “unto my nephews Callaghan Byrne, James W. Byrne, and Ered Byrne, in equal proportions.” The bequest was held to be to those nephews as a class, and that Fred Byrne having died prior to the death of testatrix, leaving no descendant, that part of the estate intended for him passed to the other two nephews who survived the testatrix. In that case a section of the Civil Code of California was quoted, which reads:
“In case of uncertainty arising upon the face of a will as to the application of any of its provisions, the testator’s intention is to be ascertained from the words of the will, taking into view the circumstances under which it was made, exclusive of his oral declarations.”
It was held that there was such uncertainty upon the face of the will, and hence the surrounding circumstances of the testatrix at the time she made the will could he shown by parol evidence aside from her oral declarations.
The familiar common-law rule of evidence for explaining a latent ambiguity in wills and other written instruments was thus by statute, made applicable to wills showing patent ambiguities also. According to the old common-law rule, ambiguities apparent upon the face of a written instrument could not' be explained by parol evidence of extrinsic facts, and, presumably, this statute was passea for the purpose of changing that rule. Other decisions practically to the same effect as those discussed above are Page v. Gilbert, 32 Hun (N. Y.) 301; Swallow v. Swallow, 166 Mass. 241, 44 N. E. 132; Hoppock v. Tucker, 59 N. Y. 202; Schaffer v. Kettel, 96 Mass. (14 Allen) 52S; Stedman v. Priest, 103 Mass. 293; Mann v. Hyde, 71 Mich. 278, 39 N. W. 78; Roosevelt v. Porter, 36 Mise. Rep. 441, 73 N. Y. Supp. 800; Meserve v. Haak, 191 Mass. 220, 77 N. E. 377; Smith v. Haynes, 202 Mass. 531, S9 N. E. 158; Hazard v. Stevens, 36 R. I. 90, 88 Atl. 980; Chase v. Peckham, 17 R. I. 385, 22 Atl. 285; Security Trust Co. v. Lovett, 78 N. J. Eg. 445, 79 Atl. 616; Lockhart v. Vandyke, 97 Va. 356, 33 S. E. 13; McCoy v. Houck, 180 Ind. 634, 99 N. E. 97.
If, in a deed or will disposing of real estate, the word “heirs” can be construed as a word of limitation and not as a word of purchase, as it plainly indicates, if a devise to two or more as joint tenants can be read as a devise to one only in the event he alone survives the testator, if a devise to several as tenants in common can be read as a devise to one only in the event the devise is to a class and that one only survives the testator, if the intent of the testator is the cardinal rule in the construction of wills, and if, as said in Jackson v. Roberts, 14 Gray (80 Mass.) 550, supra, and supported, in effect, by practically all the other authorities last cited, “No rule of law gives an inflexible sense and effect to a bequest made to children of a family, by their several names, nor to a bequest to them ‘equally’ or ‘in equal shares,’ ” then, under the facts and circumstances shown in the statement of facts in the present suit, excluding the declarations of the testator and the opinion of his attorneys therein appearing, no sound reason is perceived why the will of R. L. Hagood, which, upon its face, purports to be a devise to his two brothers as joint tenants and not as tenants in common, should not be con-
strued as a devise to a class, according to the undoubted intention of the testator; and for the reasons noted, the writer is of the opinion that the judgment of the trial court should be reversed, and judgment here rendered in favor of the appellant for the property in controversy.