Catherine Donnellan died testate in the city and county of San Francisco, and her will was there probated. The tenth clause of her will is as follows:

Upon petition for distribution Annie Sheridan, appellant herein, and Mary Smith, respondent herein, each claimed distribution as the person meant in the paragraph above quoted. The court’s decree favored Mary Smith. As has been said, Annie Sheridan appeals, and so also do certain heirs at law, who insist that the latent ambiguity disclosed to exist in paragraph 10 has not been removed by extrinsic evidence; wherefore the bequest fails and the property descends to them as heirs at law.

The evidence showed that Catherine Donnellan, born Catherine Riley, the testatrix, came to San Francisco from Ireland about fifty-five years ago, and there died, after the death of her husband without issue. At the time she left Ireland she had a sister Mary who remained in Ireland. This sister, by her marriage to Michael Cook, had two daughters, Annie and Mary. After the death of Cook she married John Donohoe. The daughter Mary married a man by the name of Smith, and lives in Ireland with the testatrix’s brother Thomas Riley. Mary has never been in the United States. The other daughter, Annie, married a man by the name of Sheridan, came to the United States about twenty-five years ago, and at the date of the will lived and still lives in Brooklyn, state of New York. The testatrix left Ireland before either of the daughters of her sister Mary was born, never saw either of the daughters, and, being unable to write, never personally wrote to either of them. This statement discloses the latent ambiguity and uncertainty as to the person whom the testatrix intended to indicate by the tenth paragraph of her will. Both Mary and Annie were her nieces, both were the daughters of her sister Mary, and both were the stepdaughters of John Donohoe, the second husband of testatrix’s sister. If the testatrix intended the niece Mary, then there is a grave error in description, for the niece Mary was never a “resident of New York.” If the testatrix meant to designate the niece who was a resident of New York, then that niece was Annie and not Mary.

Extrinsic evidence was taken, as was proper, to explain this latent ambiguity, and that evidence was this: In May, 1909, testatrix gave instructions to her attorney to draw her will. This he did, and was a witness to it. The testatrix told her attorney that she had a sister named Mary who had a daughter named Mary; that she did not know where the daughter was; she thought she was in New York;-she did not know what her name was. “In fact,” concludes Mr. Quinlan in his testimony, “everything that she told me about Mary I included in the description as she gave it to me and put it in the will as closely as I could adhere to her general statements to me.” Still further the testatrix explained to her attorney that she would not make provision so that another branch of her family “the Boston Rileys” would take the share allotted to the niece in the event that Mary could not be found, because “she had made inquiries from the Riley family in Boston as to the whereabouts of this party, and that they had refused to give her information, and she thought that if they were interested in this portion of the estate that they would not make any endeavor to find or locate this person.” The further testimony was that Annie was the elder sister, Mary the younger. Thomas Riley, a nephew of the deceased, living in Boston, Massachusetts, testified that he received a létter from the testatrix in her lifetime, which letter was appended to the deposition, and in which the following inquiry was made, “Let me know about my sister Mary’s daughter”; that he had written to Catherine Donnellan in answer to this letter, and in so writing wrote about the niece Annie Sheridan, who was the only niece of Catherine Donnellan who ever lived in New York. Annie Sheridan deposes that her father died when she was about six years of age, and at that time the testatrix, living in San Francisco, wanted her, Annie, to come to San Francisco to live with testatrix; that she, Annie, had been called Mary when she first came to America, and had been so called by Margaret Riley, an aunt by marriage at Boston, where she, Annie Sheridan, first lived.

Over objection and exception the court admitted in evidence from the deposition of Mary Smith the following: “I remember my mother getting a letter from the decedent over twenty years ago, in which she asked my mother to send me out to her home in California, as the decedent stated that all her family were dead.” And also the following from the deposition of Thadeus Doyle: “Mary Riley, the deceased (mother of Annie Sheridan and Mary Smith), told me several times that her sister Catherine (the testatrix) was married to a man named Donnellan, and that she received letters from America, and that Catherine Donnellan wanted Mary Riley’s daughter, the present Mary Smith, to be sent to her in America.” This is in substance all of the extrinsic evidence given to relieve from the latent ambiguity and to aid in the construction of the will. It will be noted that it presents absolutely no conflict.

It is a fundamental and indisputable proposition that wherever doubt arises as to the meaning of a will, such doubt is resolved by construction and that construction is one of law,—it is an application of legal rules governing construction either to the will alone or to properly admitted facts to explain what the testator meant by the doubtful language. In those cases where extrinsic evidence is permissible there may be a conflict in the extrinsic evidence itself, in which case the determination of that conflict results in a finding of pure fact. But when the facts are thus found, those facts do not solve the difficulty. They still are to be applied to the written directions of the will for the latter’s construction, and that construction still remains a construction at law. In such cases where the evidence of the facts is in conflict, it is permissible for the court or for the jury, to find the facts and those findings, under firmly established principles, will not here be disturbed. But the application to the will itself of the facts found, admitted or established, presents a qu'estion of legal construction, which is as purely a question of law as is a construction of the will without resort to extrinsic evidence. Therefore, if the facts have been found by the court upon conflicting evidence, this court, accepting the findings, will still review the construction of the court in probate and determine whether or no a wrong construction at law has been reached. If the facts are admitted, or established without conflict, the justness of the application which the court made of those facts in its construction will equally, as a legal proposition, be the subject of review. Again, it is fundamental that in all cases where extrinsic evidence is admissible to aid in expounding the will, the evidence is limited to this single purpose. It is considered, for the purpose of explaining and interpreting the language of the will, and is never permitted to show a different intent or a different object from that disclosed (though perhaps obscurely) by the language of the will itself. (6 Wigmore on Evidence pp. 2472-4.) So it is declared by section 1340 of the Civil Code: “When, applying a will, it is found that there is an imperfect description, or that no person or property exactly answers the description, mistakes and omissions must be corrected, if the error appears from the context of the will or from extrinsic evidence; but evidence of the declarations of the testator as to his intentions cannot be received.”

Broadly speaking, there are two classes of wills presenting latent ambiguities, for the removal of which ambiguities resort to extrinsic evidence is permissible. The one class is where there are two or more persons or things exactly measuring up to the description and conditions of the will, as “.to my niece, daughter of my sister Jane, residing in Alameda County,” when there are two or more nieces, daughters of his sister Jane so residing; or a devise “of my ranch in Alameda County” when the testator owns two or more ranches in that county. The other class is where no person or thing exactly answers the declarations and descriptions of the will, but where two or more persons or things in part though imperfectly do so answer. To this latter class belongs the case under consideration. The very apparent distinction between the two classes is this, that in the first class resort is had to extrinsic evidence to determine the person or thing upon whom or which the language of the will is to operate. That person or thing once being determined, as matter of law, the construction follows so much as matter of course that the legal aspect of it is sometimes regarded as negligible. And thus in some cases it is said that under circumstances such as this, the court’s or the jury’s determination of the intent of the testator is one of fact. Logically this is not true, for always must the intent of the testator be determined by the will itself and that determination is a conclusion of law. But the person or thing which is the matter in doubt being decided upon as matter of fact, the application of the intent of the will that that person should receive the property, or that the property should go to such a person, presents absolutely no difficulty. For that reason, as has been said, the fact that there still remains to the court the duty of construing the will in this particular is sometimes lost sight of. Thus a will gives a legacy “to my niece, daughter of my sister Jane, in San Francisco.” It is disclosed that the testator has a sister Jane who lives in San Francisco, and who had two nieces. Extrinsic evidence is admitted to show that at the time of the making of the will the sister Jane had but one daughter, that the testator knew this niece, that before the birth of the second niece he departed from the state, was never thereafter in communication with his sister, and did not know of the existence of the second niece. Here the finding of fact would clearly be that the testator had in mind the elder niece. This finding at once removes the only difficulty in the way of. construction; but, nevertheless, legally and logically, no matter how simple may be the construction which follows, there is the necessity of construction in the declaration by the court that the testator by his will did bequeath a legacy to the elder niece.

Much more apparent is this legal construction in the second class of cases where errors in description or designation are found, so that, however the will may be construed, the construction results in rejecting some part of such designation or description. Here, manifestly, the construction is one of law, because it necessarily results in striking from a written instrument some words or phrases contained therein as being errorneous and not expressing the true meaning and intent of the testator. To these propositions, which we think to be very plain, it is unnecessary to do more than cite Taylor v. McCowen, 154 Cal. 798 , [ 99 Pac. 351 ] ; Estate of Dominici, 151 Cal. 181 , [ 90 Pac. 448 ] ; McKeough’s Estate v. McKeough, 69 Vt. 41 , [ 37 Atl. 275 ] ; In re Welch’s Will, 78 Vt. 16 , [ 61 Atl. 145 ] ; 2 Underhill’s Law of Wills, sec. 909 et seq. ; 2 Woerner’s American Law of Administration, p. 893 ; Wigram on Wills (O’Hara’s ed.) 142 ; Charter v. Charter, 7 H. L. 364 ; Hiscocks v. Hiscocks, 5 M. &. W. 332.

In the case before us the court is called upon to do this precise thing. It becomes its duty to declare either that the use of “Mary” as the name of the devisee was error, or that the use of the phrase “a resident of New York” is erroneous description, and should therefore be rejected; or, finally, to declare that the doubt raised by the latent ambiguity cannot be resolved and the devise or bequest must lapse. If the name is rejected as error, then the devise or bequest is a perfectly clear and intelligible one—“to my niece, a resident of New York, said niece being the daughter of my deceased sister Mary.” If the description be rejected, the devise or bequest becomes equally clear and intelligible and is “to my niece Mary, said Mary being the daughter of my deceased sister Mary.” As has been said, the evidence was without conflict, and the court by its construction adopted the theory that the name controlled and that the description was erroneous and should therefore be rejected. But if, as respondent implies, the court adopted the maxim “neritas nominis tollit errorem demonstrationist as controlling, and the declaration of Redfleld that “where persons are correctly named, any amount of false description will be rejected as surplusage, and even a great degree of certainty will not induce a court to give the devise of a person named to one described but not named,” (Redfleld, 3d ed., 404), it unduly extended the application of the maxim and the meaning of the text-writer. There is no rule in the construction of wills which prefers a name to a description. The matter is clearly expounded in Theobald on Wills, p. 268. He says:

What, then, fairly appears by the uneontradicted evidence in this case? First, that the testatrix knew that she had one niece, the daughter of her sister Mary. Nothing in the evidence from the beginning to the end points to knowledge upon her part of the existence of more than one such niece. Second, she thought this niece’s name was Mary, but she knew that the niece whom she had in mind had come to America and was living in New York. She even wrote to her Boston relatives to learn the exact whereabouts of this niece whom she knew had come to America. All of the evidence of her inquiries concerning the whereabouts of this niece, her refusal to will over to the Boston Rileys the share bequeathed to the niece in case the niece should not be found, points irresistibly to the conclusion that she had in mind the niece who was in America. All of this evidence is absolutely unexplainable upon the theory that the niece she meant was the one who was living with her mother in Ireland and who had never left her home. Still less possible is it to explain the theory that Mary was meant, if credit be given to the hearsay evidence admitted by the court; for whatever else that evidence shows, it establishes knowledge upon the part of the testatrix that Mary was not and had not been in America, because she desired Mary to come to America. Again it appears affirmatively that the testatrix did know that her niece was in America and (bearing in mind that she knew of the existence of but one niece) it further appears that she wrote to the Boston Rileys to learn the exact whereabouts of the niece in America. Balancing then the probabilities of error, and considering that Mary was the mother’s name, and the name which it would not be unlikely her daughter would bear, it appears to us that the inference that the niece in America, the one “resident in New York” was intended is overwhelming, and that the error was an error in name and not in description. Rejecting from the consideration the evidence above quoted, which without question was improperly admitted, and the inference becomes well nigh irresistible. Treating for a moment the evidence improperly admitted. The first is the declaration of the niece Mary as follows: “I remember my mother getting a letter from the decedent over twenty years ago, in which she asked my mother to send me out to her home in California, as the decedent stated that all her family were dead.” The witness purported to give the contents of a written instrument —a letter—without proof either of the destruction of the letter or that the deponent had ever herself read the letter. The second is the declaration of Thadeus Doyle, also as to the contents of letters, whose destruction was not proved, which Doyle never asserts that he read, and the contents of which he was testifying to upon statements of those contents made to him by another. Here was hearsay upon hearsay, and it needs no citation of authority to support the statement that such evidence was inadmissible.

Finally, it should be added, as to the appeal of those cbntending that the devise lapses because the ambiguity is not removed by parol evidence, that what this court has already said is a declaration that the extrinsic evidence is sufficiently explicit to solve the ambiguity and points to the niece Annie as the one meant by the testatrix.

Therefore the appeal is sustained, and in the indicated particulars the decree is reversed with directions to the trial court that if there be no other or further evidence presented upon the rehearing of the matter, section 10 of the will shall -be construed to apply to and to mean the niece Annie.

Lorigan, J., and Melvin, J., concurred.