delivered the opinion of the Court:
The leading question in the case arises on the several assignments of error relating to the admission of evidence tending to show the relations of testator, after the date of the will, with his next of kin and the family of one of the legatees,- his sentiments and feelings towards them, and his intentions in respect to provisions for them in the disposition of his estate.
As we have seen, evidence introduced without objection tended to show the great and constant affection of the testator'for Miss Hynes,,whom he had chiefly supported for years. Miss 'Throckmorton was the testator’s god-daughter, her mother was his second wife’s cousin, and her father had in youth been his protégé.
The will in controversy divided his entire estate between Miss Hynes and Miss Throckmorton; the share of the latter to be held in trust by her father until she attained the age of twenty-one years.
*569There was also testimony tending to show that the testator had become alienated from some of his next of kin because they had espoused the cause of the South during the civil war. Letters were also introduced from the testator to two of his brothers and several nieces, written between 1867 and the date of the will, tending to show the existence of affection for and interest in them.
The testimony to which the exceptions apply may be summarized as follows: (1) Correspondence between testator and certain of his next of kin during the years that followed the date of the will, tending to show the existence of an affectionate interest in them. (2) Visits made by testator to Washington Holt at the old family home in Kentucky; his expressions of affection for his nephew and family. (3) Gifts of land adjoining the home to Mrs. Washington Holt; the erection of costly improvements on the old home place, and the further gift to her of $40,000 in Missouri State bonds. (4) Declarations of testator in 1881 showing anger with the Throckmorton family. On one occasion, when asked of them, he said angrily, “I do not know the Throckmortons.” (5) Instructions by testator to his servants not to admit the Throckmortons to his house ; and his conduct at a levee of President Arthur in refusing to speak to Major Throckmorton and turning liis back upon him. (6) Declarations of testator between 1884 and 1893 that ho had made W ashington Holt his executor; that all he had would be his, and that he would arrange testator’s burial; instructions to the servants to turn over his keys to Washington Holt, as he would have charge of his affairs, etc. (7) Some other declarations of similar purport, among them one made in 1893 to one Barclay to the effect that he had made a will and therein given some pictures to Barclay’s mother.
All of this evidence was offered in support of the issue of forgery as well as that of revocation, and was objected to on both grounds.
*570The twelfth and thirteenth prayers of ca'veatees relating to the bearing of a part of the same on the issue of revocation were excluded :
“12. The jury are instructed that they can not consider alleged oral declarations of Joseph Holt to the effect that he had made a will creating Washington Holt his executor, or alleged declarations to the effect that Joseph Holt had made a will in Washington Holt’s favor as showing or tending to show revocation of any will which the jury may conclude from the testimony was duly executed by the said Joseph Holt prior to such declarations.
“13. The jury are instructed that they can not consider any letters or writings offered in evidence, not themselves testamentary in character or not being in effect a last will and testament, for the purpose of showing or tending to show revocation of any will which the jury may conclude was duly executed.by the said Joseph Holt prior to the date of such letters and writing.”
The court granted the following prayer of the caveators applying to the same issue:
“4. As to the question of revocation the jury have the right to consider such burning and mutilation of the paper-writing in controversy as they may be satisfied from the evidence existed when it reached the office of the register of wills, together with all the evidence in the case tending to show any change occurring after February 7th, 1873, in .the relations of the testator either to the persons named as beneficiaries in the alleged will or any of them or to those who would be entitled to his estate under the law if he had died intestate, and if the jury are satisfied from all this evidence taken together that such burning and mutilation was done by Joseph Holt or by his directions with the intention of revoking the' same, then the jury are instructed that they should answer the fourth issue, ‘Yes.’”
Further instructions, prayed by the caveatees, relating to the issue of revocation, by mutilation, and so forth, the *571burden of proof thereon, and the effect of a subsequent will or paper-writing thereon, were given to the jury.
1. No special instruction relating, to the particular bearing of the evidence of declarations and conduct of the testator, upon the question of forgery exclusively, seems to have been prayed by either party. Because of the omission of the caveatees in this respect, the caveators contend,that its effect can not now be complained of. They contend that, being clearly admissible as bearing on the revocation of the will by mutilation, it was not the duty of the court to restrict its consideration to that issue, unless called upon by the interested party to do so.
Ordinarily, this would be the correct practice. But it is not necessary to consider whether the circumstances of this case would constitute an exception; because we think it sufficiently appears, from a paragraph of the general charge, taken in connection with the express purposes for which the evidence was offered, that it was intended to bear distinctly also upon the issue of execution.
That paragraph, to which exception was duly taken by the caveatees, reads as follows:
“You are also to consider the circumstances touching the relations of Joseph Holt to all the parties to this controversy, the next of kin, the people of his own blood, and the devisees under this will, up to the time of the date of this will, bearing directly upon this question whether it would be likely that he would have made such a will at that time and also the subsequent conduct of the testator towards his next of kin or heirs-at-law, and as well towards the beneficiaries under this will, as bearing upon the probability, the likelihood, of his having made such disposition of his estate as is made by this paper in controversy.
“ If, upon the consideration of all this evidence, you reach the conclusion that this paper is not established to your satisfaction as the genuine paper of Joseph Holt, your answ'er to this first issue should be ‘No.’ If the evidence *572positively satisfies you that he did not execute it, your answer should also be 'No;’ for the question is, Was it executed by him?”
It is proper to add also, as bearing upon the weight of this evidence with the jury, the following paragraphs from the general charge:
“It is your province, gentlemen of the jury, to pass upon the evidence, and it is your duty to base your conclusions absolutely and solely upon the evidence. It is not your province to make a will for Joseph Holt, if in your judgment you should be satisfied that he ought to have made a different will from the will in controversy. It is not your province to unmake a will, unless the evidence is of such character as to satisfy you either that Judge Holt did not execute this paper, or, if he did execute it, that he subsequently revoked it.
“In passing upon these questions you should be influenced by no other motive than the desire to reach the exact truth. Let justice be done, no matter who is injured by your verdict. You are not responsible for the outcome of this case. It matters not to you, and it should not make any difference as to your conclusions upon the evidence in this case, whether the result will indicate prevarication or perjury, on the part of any witness who has been examined in your hearing, whether that witness be high or low socially. It matters not to you, and you should not permit these considerations to influence your judgment to the slightest degree, what may be the subsequent outcome of this case, who will be injured by your verdict, either pecuniarily or who ultimately may be subjected to criminal proscution. It is not your province to determine, if this will is not genuine, by whom it was forged. Your province is to determine,, simply, is it the genuine instrument of Joseph Holt? If it is the genuine instrument of Joseph Holt, then, as I have indicated to you, it is not your busi*573ness that the devise to these beneficiaries may cut out the next of kin and the blood relations of Joseph Holt.
“With the consequences you have nothing to do. Do not permit considerations of expediency or of sympathy for any one to enter into your deliberations. Render your verdict in accordance with the conclusions to which the evidence compels you, and let the consequences of such verdict adjust themselves and rest where they may.”
2. It seems to us quite clear, both on principle and authority, that this evidence, including the declarations of the testator tending to show, after the date of the will, renewed and increased affection for his next of kin, as well as estrangement from the family of one of the legatees, and an independent provision for the remaining legatee, for whom his affection remained unabated, was competent as tending to maintain the issue of revocation.
If the will had been found among the papers of the testator unmutilated, the evidence would clearly have been incompetent to prove revocation as an independent fact. If it had been there found, in its present condition, the presumption of revocation would at once attach. Coming to light at the time and in the manner in which this paper did, we are not sure but that the same presumption of revocation ought to be indulged.
Certainly it would be more just and more consistent with the reasons for the ordinary presumptions of law, to presume that the paper had remained under the control of the testator and had been spoliated by him, as he had the lawful right to do, rather than by a stranger, who, in suppressing or defacing the will of another, commits a heinous offense. The charge of the court, without expressly declaring the existence of a presumption under the circumstances of this case, nevertheless imposed the burden of proof upon the caveators to show that the acts of spoliation had been committed by the testator with the intention to revoke. *574The evidence to maintain this burden was necessarily circumstantial.
In consideration, then, of the condition of the will and the manner in which its appearance was made, we can not escape the conviction that the state of mind of the testator, during the more than twenty years in which the will was subject to his unrestrained control, as indicated by his changes of feeling towards the parties at interest, was a material circumstance to be considered' by the jury in determining whether the mutilation was the result of an intent to revoke. Lawyer v. Smith, 8 Mich. 411, 423; Patterson v. Hickey, 32 Ga. 156, 164; Burge v. Hamilton, 72 Ga. 568, 625; Callagan v. Burns, 57 Me. 449; Collyer v. Collyer, 110 N. Y. 481, 484; McDonald v. McDonald, 142 Ind. 55, 81; Miller v. Phillips, 9 R. I. 141, 144; In re Valentine's Will, 93 Wis. 45, 55; Pickens v. Davis, 134 Mass. 252; Sugden v. Lord St. Leonards, L. R. 1 Prob. Div. 154, 176, 219, 225, 232, 241; Gould v. Lakes, L. R. 6 Prob. Div. 1, 5.
3. This brings us to the consideration of the important question that has been chiefly argued by counsel, namely, whether the foregoing evidence was competent for the consideration of the jury in determining the issue joined on the genuineness of the will.
In that consideration, it must be remembered that the subscribing witnesses had long been dead; that the genuineness of their signatures had been supported and attacked by competent and- respectable evidence in the nature of opinions; and that whilst many witnesses had expressed opinions in favor of the genuineness of the will and the testator’s signature thereto, many others of equal credibility, apparently, and opportunities to know the handwriting of the deceased testator, had pronounced against both as forgeries. In addition, we have the mysterious, unexplained deposit of the instrument in one of the public mailing boxes. The jury were required to find if Joseph Holt actually wrote this will and affixed his signature thereto.
*575The contention of the appellants is supported by the citation of many cases denying the competency of such evidence, especially the subsequent declarations of the testator, on that issue.
The rationale of decision in a considerable number of these might, probably, be reconciled with the decision of competency under the exceptional circumstances of this case; but it would serve no useful purpose to occupy time with their review and the discussion of their respective weight. As the question is one of first impression in this jurisdiction, we are free to adopt the conclusion that seems to us to be the most reasonable, and that accords with the smaller number of cases relied on by the appellees. In one of these, whose analogy is striking, the Court of Appeals of Maryland, after much consideration of the principles involved, as well as the authorities, declared the evidence competent. Hoppe v. Byers, 60 Md. 381, 390. See, also, the charge of Mr. Justice Grier to the jury in Turner v. Hand, 3 Wall. Jr. 88, 92; Stephens’ Dig. Evidence, Art. 29.
We do not mean to say that such evidence would be competent to establish the execution and existence of an undiscovered will, any more than to prove the revocation of one actually produced without sign of spoliation; for that would be, in either case, to override the provisions of the. statute law prescribing the formalities of execution and revocation. Nor would the same evidence be of sufficient weight to warrant the overthrow of a formal will produced and with execution proved by living witnesses of unimpeached credibility.
In fact, it was said in Hoppe v. Byers, supra, that under circumstances like those just mentioned, the evidence ought to be excluded. But whether, in such event, the objection would lie to its competency or its credibility, the evidence would have so little weight that, if admitted, the court might properly charge the j ury that it could not afford support for a verdict against the will. The difficulty in the accurate *576reproduction of oral declarations, as well as the danger of the jury being biased or prejudiced thereby, and therefore misled, seem to constitute the chief grounds of objection to the competency of this kind of evidence; but the same dangers lurk, though in less degeee, in all oral evidence of declarations, admissions and confessions of parties against their own interests. In respect of this, we concur in the views expressed by Mr. Justice Lumpkin, in Patterson v. Hickey, 32 Ga. 164, as follows: “It is said that the reason for not hearing parol proof is, that there is not the ordinary security that it is true. This goes to the weight of the evidence. It is true, there are many cases in which it would be entitled to but little weight; nay, but a few in which it would be entitled to any. Yet, if there be others, in which it would subserve the cause of truth and justice, it must be heard, leaving its effect to those whose province it is to weigh it. There is little danger in this, when the courts can aid the jury in pointing out its legitimate tendency.”
Because, too, one who has executed and delivered a deed or a bond can not, by subsequent declarations, impeach its execution and obligation, it has been said, in many of the cases, that the same principle applies in the case of a will. In this we can not entirety concur. In the case of the delivered deed and bond the interest of the signer becomes wholly changed, and he can not be heard to disparage the title of the other party by declarations in his own interest. But the will is a radically different instrument; it is not made effective by delivery; it is ambulatory, and can take effect only upon the death of the testator; until then only is it beyond his power of recall by destruction, express revocation, or complete substitution by one of later date. By virtue of the statutory requirements in these respects, declarations before, at the time of, or subsequent to, the transaction inquired of, can not be received as sufficient to establish the same; nor should they be permitted to override positive, unimpeached proof of execution or revocation in accordance *577with the required formalities. But, in a case where such proof is lacking, or where, by competent evidence, the genuineness of the signature has been assailed to such an extent as to require the submission of the question of its forgery to the jury, we think the feelings of the testator towards all the parties, and his relations with them and apparent intentions as disclosed by his conduct and declarations, are competent for their consideration. We agree with the Court of Appeals of Maryland, that “they are not to be taken as direct proof to establish the paper (or to destroy it), but merely ás corroborative of such direct proof, or as a circumstance in a case of this character, where such direct evidence has been first given, proper for the consideration of the jury.” 60 Md. 393.
Substantially the same views were expressed by Mr. Justice Grier, who presided in the circuit court on the trial involving the same will that was in controversy in the case greatly relied on by the appellants —Boylan v. Meeker, 28 N. J. Law, 274. He said, in charging the jury (Turner v. Hand, 3 Wall. Jr. 88):
“While it is undoubtedly true that parol declarations of a testator made before or after executing his will ought not to be received as a ground for altering or annulling it, yet cases may arise where such declarations, in connection with other circumstances, may be taken into consideration, as, for example, where there is strong evidence of conspiracy and of fraud practiced on the testator, or that the instrument is forged and false.
“In order to elucidate this principle let us suppose a case. A will is produced in court, regularly proved according to law, yet, notwithstanding the legal proof it may possibly never have been seen by the testator, never have been signed and sealed by him, and consequently does not contain his will as to the disposition of his property. Suppose it to have boon made (as has sometimes been the case in Ireland and other places) by some person personating the *578testator, and simple and perhaps honest people have thus been prevailed upon to attest it.. In such a case the signatures may be so palpably a forgery as at once to detect the fraud to any judge of handwriting. Again, suppose the will disinherits a child, a grandchild, or other relative, who has been the favored and beloved companion of the testator’s life, whom he had' uniformly pointed out and always and invariably, through his whole life, declared his intention of making his heir, and in whose favor a prior will was duly executed. Suppose the devisee in this supposed will was some worthless fellow, unknown to the testator, or, if known, despised or abhorred by him. Suppose the witnesses to be of the same character, low and degraded, with whom the testator never associated. Would not such facts, if clearly proved, condemn such instrument in the mind of every rational man? Would not the moral impossibility that the testator could ever have made such a disposition of his property be sufficient to outweigh the positive testimony of such witnesses?
“It is easy to forge the handwriting of almost any man so that it may be almost impossible for the best judges to discriminate between the false and the true, and it is too true that persons may be found willing, for a sufficient consideration, to swear to any statement of facts. Fraud can be generally proved only by circumstantial evidence. A number of distinct facts, clearly proved, may be so utterly inconsistent with the truth of the instrument as most satisfactorily to establish the fraud. The fact that the testator had uniformly, through his whole life, declared that he intended a certain relative to be his heir, that he made his will in his favor, may be an important link in the chain of circumstances from which fraud, perjury, conspiracy, and forgery may be clearly proved.”
See, also, remarks of Jessol, M. R., on the competency of declarations in general, in Sugden v. Lord St. Leonards, L. R. 1 P. D., at page 241.
*579Again, we are unable to perceive a distinction in principle between the competency of declarations in a case like this, where the fact of execution is in question, and in one where the will is assailed as the product of fraud and undue influence. In the latter case, the authorities generally concur in holding that the declarations of the testator and his relations with the parties are competent circumstances tending to throw light upon his state of mind and disposition. The only limitation is, that it shall not be received as proof of the independent fact or the truth of the things declared, but as suppletory only to direct proof of the alleged fraud and undue influence. Some of these cases show the admission of such declarations upon very slight foundation of the direct proof first required. Stephenson v. Stephenson, 62 Iowa, 163; Lane v. Moore, 151 Mass. 87 (gift by endorsement and delivery of note): See, also, Parsons v. Parsons, 66 Iowa, 754, 758; Woodbury v. Woodbury, 141 Mass. 329, 334.
The question at issue in each class of cases is practically the same: Did the testator execute the paper of his own accord? Did he execute it at all? And the danger that the jury answez’ing either question may give undue weight to the declarations and conduct of the testator is equal. All that the court can do, imder the circumstances, is to caution the jury against such conduct, as was done in this case, and to set aside the verdict when convinced that the caution was uziheeded.
In the peculiar conditions of this case, it seems to us that any evidence from which it might be inferred that the testator would, or would not, have made such a will, would be a material aid in resolving the doubt created by the other evidence whether he did or did not write and sign it. If the proponents had produced a witness of undoubted veracity and disinterestedness, who could testify to declarations of the testator referring to and certainly identifying this particular will as having been written, signed and witnessed, as it appears to be, would not its reception and con*580sideration have been in accord with the dictates of reason and justice?
In order to prevent injustice, courts have even permitted the declarations of the testator to be given as suppletory proof of legal execution where a subscribing witness has either forgotten the facts stated in the attestation, or testified in denial of the same (Will of Cottrell, 95 N. Y. 329, 337; Beadles v. Alexander, 9 Baxter, 604); and also where it appeared doubtful if the formalities of the law had been complied with in the execution of a codicil. Adams v. Norris, 23 How. 353, 368.
Why, then, for the same reason, should not testimony be received from which the jury might, in case of doubt, infer that the testator had not written or executed the instrument at all? Would it not be important to know the character and state of mind of the testator in such a case? If a man of education and character, was he in a condition, mental or otherwise, to induce him to dissemble his affections and feelings for the parties or any one of them? Was there any cause or reason why he should falsely express his affections or dislikes, or should undertake to create hopes and expectations in the minds of his next of kin that he knew would be cruelly disappointed by the production of this will made years before? In short, what was the real state of the testator’s mind and feelings before, at and after the alleged date of the execution, in relation to his property and the conflicting objects of his bounty ?
The materiality of this inquiry, under the circumstances surrounding this will, seems plain to us.
The length of time of the declarations after the date borne by the instrument is of no materiality; for, if the. instrument be a forgery, the person procuring it would naturally give it a date when the disposition would seem more reasonable or more in accord with the testator’s sentiments thán at a later period; besides it would be a reasonable precaution to antedate a well-known change therein.
*5814. It follows from we have heretofore said,thatthe testator’s declarations tending to show the making of a will in Washington Holt’s favor, or appointing him his executor, were not competent to prove the execution of such a will either as a disposition of the estate or as a distinct revocation of the will offered for probate. They were admissible, and seem to have been admitted, only as one among other circumstances tending to show the state of mind of the testator and shed light upon the issue of revocation, by mutilation, of the will offered for probate; and as such we have held them admissible on the issue in respect of forgery also. The court gave a number of instructions hslced by the caveatees respecting the effect of the proof of a supposed later will, and none seem to have been prayed by the caveators. Had the twelfth and thirteenth prayers of the caveatees been modified so as to thus limit the weight or effect of this evidence of a different will, it would have been the duty of the court to grant them. If, however, palpable error had beemcommitted in this respect, the judgment would not be reversed, because it stands upon the verdict finding the will to be a forgery. Having so found, the jury could only answer the remaining question as they did — not revoked because never executed.
5. The next question arises on exceptions taken to the action of the court in permitting witnesses for the caveators to express opinions, founded upon their knowledge of the legal attainments and literary culture and style of the testator, that the will had not been written by him; and also in excluding like evidence when offered in rebuttal by caveatees.
Many genuine letters and documents of the deceased were in evidence for the consideration of the jury. Evidence was also given showing peculiarities of the testator in the matter of punctuation.
Among the witnesses who expressed opinions against the genuineness of the will and signature of the testator were *582several who said, in addition, that they had the same opinion because of the style of the composition. One testified that this was the chief ground of her opinion.
Two of them, officers who had served with the testator in the office of the Judge Advocate and were familiar with his handwriting, testified to the depth of his legal learning and the excellence of his English composition. One of these said of him: “He had the clearest power of expression and was the finést rhetorician witness had ever met.” Both criticised the word “inherit” as used inartificially in the will, and one of them said he did not think the testator would so use the expression.
Attention has been directed, on the argument, to the words following the recital of the name of the executor, Luke Devlin: “Whose character I believe to be of the highest standard; ” but we do not find that any witness referred to them.
In the course of their rebuttal, the caveatees proved by John A. Bingham, long and intimate acquaintance with the testator and familiarity with his handwriting, style of composition and general attainments. But the court refused to permit him to express an opinion in favor of the genuineness of the will and signature, formed upon either or both grounds.
It will be remembered that when the caveatees first closed their case they were admonished by the court that the rule in respect of rebuttal evidence would be rigidly enforced; and that they thereupon withdrew the announcement and examined a number of witnesses whose opinions were that the will and signature of testator were genuine. It does not appear that there was any reason why this witness could not have been called on the matter of the handwriting, simply, in the opening.
(1) Questions in respect of the order of the introduction of evidence are matters of such broad discretion with the trial courts that its exercise will not be disturbed except *583when abused. Lansburgh v. Wimsatt, 7 App. D. C. 271, 276; Olmstead v. Webb, 5 App. D. C. 38, 57; First Unitarian Church Soc. v. Faulkner, 91 U. S. 415, 418.
(2) If there was error in refusing to permit evidence by this witness in rebuttal of the opinions founded on the testator’s style of composition, it was cured by the subsequent action of the court.
For the same reason, it is not necessary to determine whether the opinions of witnesses concerning the genuineness of a writing are admissible when founded wholly or in part upon their knowledge of the technical learning, culture and literary style of the alleged writer.
It appears that when the case was closed, the court instructed the jury to disregard all opinions as to the falsity of the papers, founded “ either wholly or in part upon the character of the composition, or upon any other circumstance than the handwriting alone.”
We are not prepared to admit as clear, that evidence of opinion in respect of the authorship of a paper-writing by a competent judge familiar with the learning, culture and literary style of the alleged writer, is not equally competent and credible with that based solely upon knowledge of handwriting. But, conceding its incompetency, the error was corrected by the charge withdrawing it from the consideration of the jury. This rule of practice is well settled in the courts of the United States, if not elsewhere. Pennsylvania Co. v. Roy, 102 U. S. 451, 458.
It is only in a clearly exceptional case that error will be considered when assigned on the admission of incompetent evidence that has subsequently been withdrawn from consideration by the charge of the court, Hopt v. Utah, 120 U. S. 430, 438; Waldron v. Waldron, 156 U. S. 361, 383. Those exceptional conditions are not found here.
6. The last exception, upon which error has been assigned, also relates to the exclusion of evidence offered by the caveatees in rebuttal.
*584In opening their case, they offered P. Tecumseh Sherman •as a witness, and he expressed the opinion that the signature of his mother, Mrs. Ellen B. E. Sherman, was'genuine. He was not called upon to give an opinion as to the genuineness of the signature of his father, General W. T. Sherman. That was proved by Senator John Sherman.
Caveators introduced a witness, John B. Randolph, who testified that he had been employed for more than thirty years in the office of the Secretary of War, during which time U. S. Grant and W. T. Sherman had each acted as Secretary of War, and the latter had been General of the Army. During this service he had become familiar with their signatures, and had recently re-examined them. In his opinion neither signature attached to the will was genuine. On cross-examination, among other specifications of peculiarities in the 'genuine signature of General Sherman, he said that, “in the long quirl to the capital T the upper and lower lines met,” and that he had “ never seen one in which they had not met.” Subsequently he said that -“in four out of five” genuine signatures observed by him, the upper and lower lines of this “quirl” met. (They do not meet in the signature attached to this will.) He also thought the loop in the S was larger than usual.
Caveatees, in rebuttal, offered to prove by the witness, P. Tecumseh Sherman, that this failure of the lines to meet was “by no means an unusual feature in the signature of General Sherman, and that it is frequently, if not habitually, found therein;” and, also, that loops at the bottom of the S as large as that in the signature to the will were also usually found. The court excluded this evidence, on the objection of the caveators that it was not competent rebuttal.
Referring again to the express and timely notice heretofore mentioned as given by the court as regards the order to be observed in the introduction of the evidence of handwriting, which was acquiesced in by the caveatees, and *585considering the fact that this witness, though examined in the opening, had not been asked a question concerning this signature, we think it was within the discretion of the court to exclude his testimony.
We have had some doubt whether that part of it referring to the manner of making the “ quirl” was not admissible as tending to rebut that part of the witness Randolph’s testimony. Had Mr. Randolph testified that the peculiarity mentioned by him invariably appeared in the genuine signature of General Sherman; or had the witness offered in rebuttal examined the signatures of General Sherman in the office of the Secretary of War and been prepared to state that they did not show the peculiarity mentioned by Randolph, the evidence would have been competent in rebuttal.
But Randolph did not say that this was invariable. He gave the proportion of its occurrence as about “four out of five.” This might have been true as regards the signatures examined by him, and not true of those most familiar to the son of General Sherman.
What the latter proposed to state was that this continuous opening between the lines was not an unusual feature of the signatures familiar to him. What the former had said was, that, in the proportion of four to one, the signatures with which he was familiar did not show this continuous opening, and it does not appear that it was proposed to contradict this statement directly by the evidence offered in rebuttal.
Whilst inclined to the opinion that the discretion of the court might well have been liberally exercised in favor of the admission of this evidence, we do not think that its exclusion, considering, too, the indirect and slight contradiction between it and the evidence of Randolph, was an error for which the judgment should be reversed.
We are of opinion that the judgment should be affirmed, with costs; and it is so ordered. Affirmed.