Vinje, J.
Appellants contend the evidence shows (1) that Martha was born in wedlock; (2) if not, that she is the illegitimate child of Thomas Richmond; and (3) that he sufficiently acknowledged her in writing to be his child to entitle her to- inherit.
It appears from the evidence that Thomas Richmond was born a slave in Randolph county, Missouri, near College *636Mound, on the farm of Samuel Richmond, who owned his parents, John and Queen Richmond. They lived together in a separate cabin and continued to be the slaves of Samuel Richmond until slavery was abolished in Missouri about January 1, 1865. Thomas was upon the Richmond place as a slave from the time of his birth until he left, about January 30, 1864. Margaret, a daughter of Samuel Richmond, married one Caswell Smith, and they lived upon a farm adjoining the Richmond place. One of the slaves owned by Caswell Smith was Sarah, who was known by the name of Sarah Smith. Caswell Smith inherited her from his father, who lived in Horth Carolina, and she was brought to Missouri some ten years before the war. She was then grown and had a child who afterwards died. After coming to Caswell Smith’s place she had a girl child, a mulatto, who died when three or four months old. Her next' child was a boy, colored, named William, who died about 1870; her next child was a girl, colored, named Adeline, who died about 1890; and her next child, was a girl, Martha,, colored, who is one of the appellants in this case. The date of her birth is in dispute, and is variously testified to as having been as early as 1863 and as late as 1865. About January 30, 1864, Thomas Richmond ran away from the Samuel Richmond place and went' to Macon, about fwenty miles distant, and on January 31st enlisted for three years in the Union army. He was mustered into service at Benton Barracks, St. Louis, on February 8, 1864, and was discharged June 14, 1865. According to his discharge he was then twenty-five years of age. Thomas Richmond frequently visited Sarah Smith and had intercourse with her, certainly as late as January 30, 1864. Sarah, however, never assumed the name of Richmond, but up to the time of her death at Macon, Missouri, about 1886, was known only by the name of Sarah Smith. Martha, now called Richmond, up to the time she came to Wisconsin in 1891 was known by the name of Martha Smith. After Thomas Richmond’s discharge from the Union army he worked in St. *637Louis and in Illinois, and came to Grant county, Wisconsin, in 1866 or 1867, and continued to reside there until bis death in 1909. Most of the time he lived on a small farm in the town of Beetown, near Lancaster, but some of the time he lived in Lancaster. A few years after Thomas settled in Grant county, and about 1871 or 1873, he went back to Missouri to bring his father and mother to Wisconsin, and did so. From statements he made before he went it seems he contemplated bringing with him Sarah Smith also. When he went into the army she had two children living, William and Adeline, of whom Thomas spoke as his own; he also spoke of Sarah as his wife. He found, however, when he returned to Missouri that, besides Martha, there had been born to Sarah two other children, one known as Sam Oliver and the other called Bob. He did not bring Sarah with him to Wisconsin. From the time the parents of Thomás came to Wisconsin they lived with him on his farm near Lancaster. In 1891, his mother becoming feeble and desiring to have assistance for her in the house, he, through others, as he could neither read nor write, sent several letters to Martha requesting her to come to Wisconsin and live with them. Only one of these letters, which was written by Mr. John Oarthew, has been preserved. It reads as follows:
“LANCASTER, Wis., Sept. 19 — 91.
"My Dear Daughter: Your letter of the 2nd of this month duly came to hand. I have been so very busy could not answer before. I am first rate but mother is not so very well. You speak of coming on railroad. You had better come on to Oassville, Grant Co., Wis., on the O. B. and Northern and I will meet you there if you will send me a letter a week ahead what day you will start or when you will arrive at Oass-ville. Hoping you are all well and will come all safe soon, I am, Your affectionate father,
“Thomas Richmond.”
The envelope in which this letter was sent has not been preserved, and Mr. Oarthew cannot swear positively to what ap*638peared thereon. Mr. John Mankel and Mr. N. H. Suttle also wrote letters to Martha, for Thomas about the same time. Each testifies that he cannot remember the contents of the letters except that they concerned Martha's coming to Wisconsin. The recollection of the witnesses is that said letters began “Dear Daughter” or “Dear Daughter Martha” and that they were signed “Tour father, Thomas Richmond.”
" Martha came to Grant county in 1891 and lived upon the Richmond place with Thomas for eight or nine years, when a difference between them arose and she went to Lancaster and worked out for a while, during which time she occasionally visited him on his farm. She then went back to the farm for a short time, but afterwards went to South Dakota, and from there to Phoenix, Arizona. Some weeks before Thomas died, and while he was sick, he sent money to her to pay for her return to Wisconsin, and she came back and was with him during the last two weeks of his life. He died June 22, 1909, leaving an estate inventoried at about $5,000. He left surviving him two brothers, Henry Richmond, the respondent, and Peter Richmond.
Thomas Richmond was a pensioner of the United States. Under date of January 15,1898, the pension bureau at Washington sent to him certain' questions to be answered and to be returned with his next quarterly voucher to the pension agents at Milwaukee. He signed the paper by mark under date of July 5, 1898, and answered some of the questions. To the question, “Are you married ? If so, please state your wife’s full name and her maiden name,” the answer given was, “No. Wife is dead. Maiden name was Sairih Smith. Died in Macon City, Missouri, in the year 1868, as near as I can remember.” • To the question, “When, where, and by whom were you married ?” no answer was given. To the question, “Have you any children living? If so, please state their names and the dates of their birth,” the reply was “Yes, Martha Richmond. I think she was born in the year 1865.” *639These statements were made out for Richmond by one W. E. Webb, in whose presence they were signed. The date of Sarah Smith’s death should have been given as 1886 instead of 1868. How the mistake occurred does not appear.
The trial court found that Martha Richmond is not the child of Thomas Richmond, and that any acknowledgment of her as his daughter made by Thomas was made through his mistake as to her paternity. The trial judge evidently based his findings upon the truth of the testimony contained in the depositions .of Margaret Smith and Margaret E. Terry, both of whom testified that Martha was bom in December, 1864. It is a verity in the case that Thomas Richmond left the Samuel Richmond place about the last of January, 1864; that he joined the service at Macon, Missouri, January 31st, and was mustered in February 8th at Benton Barracks, St. Louis. If it be a fact that Martha was born in December, 1864, and that, as found by the trial court, Thomas did not have access to or intercourse with Sarah Smith later than January 30, 1864, then it follows that Martha is not the child, legitimate or illegitimate, of Thomas Richmond. If, however, Martha was born before Thomas Richmond went into the army, as testified to by Milly Smith, or if Thomas Richmond had intercourse with Sarah after he joined the army, or if Martha was born on or before November 1, 1864, then the basis upon which the trial judge made his findings falls. Macon, at which place Richmond joined the army, was only twenty miles from the Samuel Richmond place. While the testimony is to the effect that he was not seen in or about the Richmond neighborhood after he left it to join the army, still it cannot be said to be an absolute verity that he did not return during his term of service. It is also' apparent from the testimony of Margaret Smith and Margaret E. Terry that their recollection of the birth of Martha as occurring in December, 1864, was vague and unsatisfactory. It is true that Margaret Smith, who was seventy-five *640years old at tbe time ber deposition was taken, bases ber evidence, that sbe knew tbe time of Martha’s birtb, npon tbe recollection that sbe was born about a month later than ber own. son, and be was born November 7, 1864. Sbe testifies, however, that sbe bad not refreshed ber recollection from any records or documents at tbe time sbe gave ber testimony, and •it is apparent on reading it that sbe was much -confused as to dates, which of course would be natural for a person of ber age, especially considering tbe remoteness in time of tbe facts to which sbe testified. As an instance of bow tbe witness confused dates tbe following will suffice: To tbe question, “Tell us as near as you can about when be left to go to tbe army ?” sbe replied, “He left during tbe winter, along after Christmas in 1864, as well as I remember.” Her counsel then suggests, “You mean in 1863 after Christmas, don’t you?” and sbe answered, “What I meant was 1863.” “Q. You mean after Christmas, 1863?” “A. It is 186^ after Christmas, I guess. Just not knowing these things beforehand I just can’t remember. I am getting old.” Tbe testimony of Margaret E; Terry, who was seventy years old at tbe time ber deposition was taken and who lived on a farm adjoining that of Samuel Richmond, is still more vague' and unsatisfactory. Sbe does, however, testify that Martha was born in December, 1864, but sbe says sbe was not present at ber birtb, adding: “I am white and sbe was a low-down, black slave. We didn’t visit negroes.” Tbe addendum • suggests at least tbe unimportance of tbe event to ber. Sbe claims sbe remembers tbe date by a remark ber mother made at tbe time Martha was bom, but,what that remark was is not disclosed by tbe testimony.
Were tbe question of tbe paternity of Martha dependent solely upon tbe date of ber birtb, as shown by tbe testimony, then this court would perhaps not disturb tbe finding of tbe trial court that Martha was not tbe child of Thomas Richmond. But in our view of tbe case tbe question of tbe pa-*641temity of tbe child must be determined not alone upon tbe testimony as to wbat particular month and year Martha was bom, but upon tbe whole evidence in tbe case. It is obvious that two women of tbe ages of seventy and seventy-five, respectively, called upon to testify as to tbe month and year in which a negro child was bom forty-six years previously, without any opportunity for refreshing their recollections from family records or otherwise, would be apt to make a mistake of at least a few months, if not years, as to the date of such birth. Such a mistake, even as to the date of the birth of one of their own children, would not be strange or unlikely, much less that of a child of a negro slave. We deem the fact that Thomas Richmond up to the time that he had the misunderstanding with Martha in 1901, and even thereafter, shortly before he died, always spoke of and acknowledged her as his child, more persuasive and controlling than the testimony of the two above mentioned witnesses. No one was in a better position to know the truth of the matter than were he and Sarah Smith, and it does not appear that either of them claimed Martha had any other father than Thomas Richmond. The latter, in 1898, when answering the pension department, said Martha was his daughter, and that he believed she was horn in 1865. He fixed the date of her birth later than did Margaret Smith and Margaret E. Terry. It seems, therefore, that Thomas Richmond did not regard the fact that Martha was born as late as 1865 as any reason why he was not her father. As before stated, he knew best whether or not it was, and his conclusion on that subject, adhered to for practically over forty years, ought not to be lightly brushed aside or nullified by the dim, uncertain memory of two old women as to the exact date of such an unimportant event as the birth of a slave child over forty-seven years ago.
The trial judge speaks of Thomas having acknowledged Martha to be his daughter under a mistake as to her paternity. *642The record is barren of any fact that came to the knowledge of Thomas Eichmond as to the paternity of Martha after 1872. In 1871 or 1872 he visited Macon, Missouri, where Sarah Smith then lived, and where she continued to live till the time of her death in 1886. He then knew that she had several children of which he was not the father, yet as late as 1891 he writes to Mrn'tha and addresses her as “Dear daughter” and signs himself “Your father” or “Your affectionate father.” ' She came to live with him and kept house for him for eight or nine years, during which time he held her out as his daughter, spoke of her as such, and she was known as such in the neighborhood in which he lived. That the children William and Adeline were his is not disputed, although they were known by the name of Smith in Macon, where they resided until their death. Practically the only doubt that is cast upon the paternity of Martha is the testimony of Margaret Smith and Margaret E. Terry to the effect that she was born in December, 1864. If there be a mistake as to that, then there is no greater uncertainty as to the paternity of Martha than of that of Adeline and William. We have therefore reached the conclusion that the county judge was correct in finding that Martha was the child of Thomas Eichmond and that the trial court erred in finding that she was not the child, either legitimate or illegitimate, of Thomas Eichmond.
It is strenuously urged by the appellants that Thomas Eichmond and Sarah Smith were married according to the custom of slaves and that Martha was bom in wedlock. There is considerable testimony to sustain the claim of such marriage. Of course if there was a lawful marriage Martha needed no legitimation. Assuming, however, that Martha was illegitimate, the question arises, Did Thomas Eichmond in writing acknowledge himself to be her father so as to entitle her to inherit his estate? Sec. 2274, Stats., provides-that “Every illegitimate child shall be considered as heir of *643tbe person who, shall, in writing signed in the presence of a competent witness, have acknowledged himself to be the father of such child, and shall in all cases be considered as heir of his mother, and shall inherit his or her estate, in whole or in part, as the case may be, in the same manner as if he had been bom in lawful wedlock.” At least two facts must co-exist to constitute legitimation under the statute, namely, an illegitimate child and a written acknowledgment of the paternity thereof by some person according to the statutory provisions. The statute applies only to illegitimate children. Grant v. Mitchell, 83 Me. 23, 21 Atl. 178. But it is not essential that the written acknowledgment shall be made for the express purpose of establishing heirship or of a compliance with the statute. It may be made in entire ignorance of the statutory requirements and for purposes other than that of creating heirship, provided there is otherwise a compliance with the statute. In re Rohrer, 22 Wash. 151, 60 Pac. 122; Blythe v. Ayres, 96 Cal. 532, 31 Pac. 915. Its requirements are an acknowledgment by a person in writing, signed in the presence of a competent witness, that he is the father of the illegitimate child. When such a writing is produced the statute is satisfied. It does not require that the witness should attest or subscribe the writing or acknowledgment — only that the person making it shall sign in the presence of the witness. In re Pederson's Estate, 97 Minn. 491, 106 N. W. 958; Blythe v. Ayres, supra. Neither does it require that the acknowledgment of paternity shall be in precise formal language. It is sufficient if it appear with reasonable clearness and certainty from the written words that the paternity of the child is acknowledged. Crane v. Crane, 31 Iowa, 296. So, also, the law of the domicile of the person making the written acknowledgment, and not that of the domicile of the child or the mother, governs the question of legitimation. Blythe v. Ayres, supra; Moen v. Moen, 16 S. Dak. 210, 92 N. W. 13.
*644In tbe present case, in reply to inquiries made by tbe government relative to a matter of great importance to tbe deceased, be declared that Martha, was bis daughter. Tbe acknowledgment of paternity was made in unmistakable terms and under such circumstances .as to exclude inadvertence or mistake. Tbe instrument was signed by bim in tbe presence of Mr. Webb, a competent witness. It must be held tbat sucb writing fully satisfies tbe requirements of tbe statute. Tbe pension statement is supplemented by letters written to Martha addressing ber as daughter, or dear daughter, in which tbe deceased declared himself to be ber father. At least one of these letters was signed by bim, through Mr. Car-tbew, and in bis presence as a competent witness. A number of cases, in addition to those already mentioned, in which various similar declarations have been held a sufficient acknowledgment of paternity under substantially similar statutes, is added for reference. Brown v. Iowa Legion of Honor, 107 Iowa, 439, 78 N. W. 73; Britt v. Hall, 116 Iowa, 564, 90 N. W. 340; In re De Laveaga’s Estate, 142 Cal. 158, 75 Pac. 790; Thomas v. Estate of Thomas, 64 Neb. 581, 90 N. W. 630. In the following cases tbe evidence was held insufficient to show acknowledgment under tbe statute: Watson v. Richardson, 110 Iowa, 673, 80 N. W. 407; Lind v. Burke, 56 Neb. 785, 77 N. W. 444; Moore v. Flack, 77 Neb. 52, 108 N. W. 143.
Tbe question is raised by counsel for appellants tba.t since tbe deceased, in compliance with tbe statute, acknowledged Martha to be bis daughter, she is entitled to inherit even though in fact she was not bis child. Having found tbat be was ber father, it becomes unnecessary to decide tbe question, and' it is reserved for future determination.
By the Oowrt. — Judgment reversed, and cause remanded with directions to affirm tbe judgment of tbe county court.