John A. Murphy was never married. He died February 2, 1908, seised of a quarter section of land. The defendants, Sion and Polly Ann Murphy are his parents, and but for the claim of plaintiff entitled to in [*257] herit the estate. The latter is the son of Anna Gehrt, nee Schnoor, who had married one Weasel, and been deserted by him, and was born out of wedlock. His claim is 'that deceased was his father and in his lifetime recognized him as his child generally and notoriously and also in writing. Whether such relationship existed need not be determined, for, as we think, the trial court rightly concluded that the evidence was insufficient to establish recognition such as exacted by statute.

1. Bastardy acknowledgment of paternity: evidence. It may be that there was a rumor or common report iñ the neighborhood that deceased was claimant’s father, but evidence thereof was not admissible to establish general and notorious recognition. Watson v. Richardson, 110 Iowa, 673. Nor was the other evidence sufficient for that purpose. ^ A physician who attended claimant’s mother in confinement testified that, at her instance, he informed deceased that she said he would pay for his services; that deceased in a joking way replied that he would see to that later and afterwards, upon being reminded that he had not been paid, advised him not to worry, and inquired how the mother and child were getting on. It will be noted that in neither instance did he acknowledge paternity; but, according to the witness, when riding with him, after the lapse of many years, decedent said, in response to an inquiry as to how he got out of the Schnoor deal, “Oh, I wrote a letter to the old man and fixed the matter up,” and admitted that the child was his. The credibility of the witness was assaulted by the testimony of a brother of decedent, and a second cousin, who met him in Crestón, testifying that he had asserted on Thanksgiving day previous to the trial with profane emphasis that the child was no more John Murphy’s than his. The only other evidence which might be relied on was of a joking response of decedent that he wished he had another boy, when several were present, to a remark of an associate that he be [*258] lieved he had as good or a bigger boy than Murphy, and of the circumstance that at a gathering he laughed, but did not reply when referred to in a joking way as papa or dad to Anna Wessel’s child. Manifestly his conduct toward his bantering associates ought not to be construed into a recognition of the child as his offspring. In the first of these conversations the “boy” talked of was not identified as plaintiff, and in the last his silence was no more than the discreet evasion of recognition. That it was his purpose to avoid recognition plainly appears from plaintiff’s testimony that decedent, though residing but three and onéfourth miles from his home until he died, never spoke to him, and in passing did so apparently without noticing him. Even if it might be found that he admitted his relationship to the physician, the evidence failed to show that he had done so to others, and therefore the record is wanting in proof that such recognition, as there might have been, was general or notorious, both of -which in the absence of writing is exacted by statute as a condition to inheritance.

2. Evidence: nonexpert testimony: comparison of handwriting. II. The main contention of plaintiff is that after his birth decedent wrote a letter to his mother’s father, with whom she resided, in which he acknowledged his paternity of her illegitimate son. The mother can neither read nor write, but testified that her sister Mary then read a letter to her purporting to be addressed to her lather, Jacob Schnoor, though she did not undertake to recall its contents. A brother, Henry Schnoor, then twenty years old, testified that shortly after plaintiff was born his father received a letter addressed to him and postmarked Arbor Hill (the local post office), purporting to be signed by John A. Murphy, in which the writer stated that he knew Anna’s child was his; that he was unable to support a wife and the old folks too; and that he would like to settle in some satisfactory way. He also thought, upon [*259] having his memory refreshed, that something was said about the young and old not being able to get along well together. An elder brother and two sisters gave like testimony to the receipt of such a letter and its contents. The loss of the letter was sufficiently established, diligent search having been made where Jacob Schnoor, who had been dead several years, kept his papers, and no one pretending to have seen it for more than fifteen years. The brothers and one sister were shown three proven signatures of deceased, and, over objection, testified that in their opinion the lost letter was in the same handwriting. Another sister, Mary, then thirteen years of age, had lost her eyesight some eight years prior to the trial. She testified to having read the letter to plaintiff’s mother, and that the signature thereto was in the same handwriting as the name of ’deceased she had seen him write in an autograph album about a year before seeing the letter. She had seen him write on no other occasion, and was not otherwise familiar with his handwriting. While the witnesses were able to read and write, their education was limited as appears from their examination. None testified to having any experience in the comparison of hand-writings.

Was the testimony of these witnesses admissible? Section 4620 of the Code declared that: “Evidence respecting handwriting may be given by experts, by comparison or by comparison by the jury with writings of the same person which are proved to be genuine.” The wording of this section is slightly different from the corresponding section of the Code of 1873. Section 3655 thereof: “Evidence respecting handwriting may be given by comparison made by experts, or by the jury, with writings of the same person which are proved to be genuine.” The meaning, however, is the same, the change being made that the statute might accurately express what was intended, and there is no warrant for the contention that the transposition or the in [*260] sertion of the words “by comparison” was designed to render competent opinions of nonexperts, not familiar with the controverted handwriting. Under the section, as it now stands, the evidence of experts may be given. How? By comparison. Such comparison may also be made by the jury, and farther the Legislature has not spoken. The purpose of the statute was to establish the rule with respect to this class of evidence, as the decisions in other jurisdictions were in conflict. See Moore v. United States, 91 U. S. 270 (23 L. Ed. 346); State v. Thompson, 80 Me. 194 (13 Atl. 892, 6 Am. St. Rep. 172); Wigmore on Ev. section 1991 et seq. Aside from such comparisons, evidence of handwriting may be (1) by witnesses who saw the person write or t© whom admissions of having written have been made; and (2) by witnesses familiar with the handwriting of the person alleged to have been the writer, and who are able to testify from such familiarity to belief respecting the writing in question. In a sense the last mentioned testimony is by comparison, for the witness testifies from the mental impression or picture of writing shown to be genuine to his judgment concerning that in dispute. In any event, the testimony is based on knowledge of the handwriting of the person. Of course, much depends on the intelligence, memory, and observation of the witness. The circumstance. that he'has seen the person write but once and then his name only, and this long previous to the trial, is sufficient in law to establish his competency if the impression of the character of the handwriting because of acuteness of observation and accuracy of recollection has been gained and retained. Grreenleaf Evidence, 577; Wigmore on Ev. sections 694, 695, 696; Wilson v. Van Leer, 127 Pa. 371 (17 Atl. 1097, 14 Am. St. Rep. 854); In re Diggins, 68 Vt. 198 (34 Atl. 697).

But the books are agreed that a witness not shown to be familiar with the handwriting, and not qualified to speak as an expert, may not testify from comparison made [*261] of genuine writings with that in controversy. This was formerly put on the ground that an opinion by a nonexpert would Te of no assistance to court or jury, either being quite as competent to determine the “similitude of hands.” At the common law, however, the jury was not permitted to compare the writing in dispute with other documents, even though identified as genuine unless otherwise in the case. Were the writing in dispute lost or destroyed, the reason of the rule would seem to have failed only to the extent that the nonexpert might be permitted to testify to a comparison the jury or court might have made had it been adduced at the trial. But no such documents were before the court, so that as the law stood, independent of statute, the testimony of three of the witnesses was not admissible. Nor does the statute in authorizing comparison with any writing proven genuine obviate this conclusion, for by declaring what evidence by comparison may be considered the section of the Code quoted impliedly excludes all other of that kind. Mixer v. Bennett, 70 Iowa, 329. In that case the exclusion of the testimony by comparison of a nonexpert was based on the statute, and not on the ground that the jury or court was as competent to decide. Moreover, the authorities are in sharp conflict as to whether an expert witness may testify by comparison where the writing in controversy is lost. 17 Cyc. 173. This court has held such evidence admissible (Hammond v. Wolf, 78 Iowa, 227), but from that it does not follow that we will go a step farther and approve of such testimony by nonexperts, notwithstanding the ordinary rule as well as the statute making them incompetent. The theory on which expert witnesses are permitted to testify is that' the handwriting is always in some degree the reflex of the nervous organization of the writer, which, independently of the will, and unconsciously, causes him to stamp his individuality in his writing. In re Gordon’s Will, 50 N. J. Eq. 397 (26 Atl. 268). The value of such [*262] evidence depends largely on the identification and number of similar characteristics or lack thereof between the disputed writing and the standards. The appearance or lack of one characteristic may be accounted as a coincidence or accident, but, as the number increases, the probability of being a mere coincidence or accident disappears, and conviction as in cases of circumstantial evidence may become irresistible. The court or jury ordinarily will derive aid by a comparison in the light of testimony by experts, but the nonexpert is not able to point out differences or similarities such as mentioned, and their testimony by comparison would be of little or no value if received. Especially is this true where comparison is attempted to be made with an instrument not seen for many years.

3. Same: expert evidence. Testimony of experts by comparison repeatedly has been declared of the lowest order. Patton v. Lund, 114 Iowa, 205; Borland v. Walrath, 33 Iowa, 130; Whitaker v. Parker, 42 Iowa, 85; Jackson v. Adams, 100 Iowa, 163. It has been declared weak and unsatisfactory. Browning v. Gosnell, 91 Iowa, 448. But its weight is for the jury. Ball v. Skinner, 134 Iowa, 298. Because of its unsatisfactory character, the Legislature in its wisdom declined to authorize and impliedly declared incompetent evidence by comparison by persons not qualified to mark the distinctions or similarities on which the value of such evidence dejDends. The district court rightly excluded the testimony by comparison of the three witnesses.

4. Same: nonexpert evidence. The other witness was but thirteen years old when she claims to have read the letter. Her familiarity with deceased’s handwriting was based on having seen him write his name in an autograph album about nineteen years previous to the trial. At that time she could read, but, as said, had been blind for a long time prior to the trial. The cir [*263] cumstanee of having seen him write but once and that so long ago, while not authorizing the rejection of her testimony arbitrarily, greatly impairs its value. Doubtless it is possible that the single instance created such an impression on the mind that it has continued all the years intervening with such distinctness as to enable her to recall its resemblance with a signature to a letter which also is but a memory of long ago and say both are in the same handwriting, but it is so out of the ordinary and improbable that the court rightly declined to base its judgment thereon.

Our conclusion is that recognition in writing was not proven. — Affirmed.