The appellant William B. Kennedy and the appellee Yada Kennedy Couillarcl were divorced in 1955. They had owned a 41.80-acre tract of land as tenants by the entirety. During the negotiations in the divorce case the couple purportedly executed a warranty deed conveying the land to Kennedy’s parents, who later reconveyecl it to him. This is a suit by Mrs. Couillard to cancel both conveyances, on the ground that her signature upon the first deed was a forgery. The only question we need consider is whether the chancellor was right in holding that the deed was in fact forged.
*354There were two hearings in the court below. At the first one Mrs. Couillard denied having signed the deed in question. Dr. Orlando Stephenson, testifying as a handwriting expert, explained his reasons for thinking the deed to be a forgery. Mrs. Couillard’s testimony was contradicted by Kennedy and his mother, who both said that they saw Mrs. Couillard sign the deed. After the hearing the chancellor gave a written opinion in which he discounted the conflicting testimony of the interested parties and concluded that he should accept the disinterested opinion of Dr. Stephenson, which was then the only expert testimony in the record.
Before a decree was entered the case was reopened for a second hearing. The notary who took the acknowledgments, appearing as a witness for Kennedy, identified her own signature as genuine, but she was unable to remember after seven years whether Mrs. Couillard had signed the deed in her presence. Kennedy also introduced the testimony of another handwriting expert, Earl Davenport, who was of the opinion that the questioned signature was genuine. After the second hearing the chancellor announced that the additional testimony had not affected his earlier conclusions. We think the court was in error, for the supplemental proof destroyed ■whatever value Dr. Stephenson’s opinion may have had and tipped the scales decidedly in favor of the appellants.
Dr. Stephenson’s testimony, standing alone, was undeniably convincing. He based his opinion solely upon a comparison of the disputed signature with 19 specimen signatures that the former Mrs. Kennedy (now Mrs. Couillard) had written upon a single sheet of paper. The witness pointed out that in the challenged signature there was a break in the name Yada, where the pen had been lifted from the paper, but there was no such break in any of the 19 specimens. Moreover, in all of the 19 specimens there was an upward flourish of the pen at the end of each name (Yada and Kennedy), but there was no similar flourish in those names as they appeared in the deed. Upon this reasoning Dr. Stephenson concluded that the signature in dispute was a forgery.
*355Now it happened that on May 17, 1955, the day on which the deed in controversy was apparently executed, Yada Kennedy signed three other documents: A waiver of service, an agreement to take depositions, and an application for the purchase of traveler’s checks. These three signatures are contemporary and admittedly genuine. Davenport, the appellants’ handwriting expert, demonstrated with much force the great similarity between the three authentic signatures and the one upon the questioned deed. Among other things, all four contain the break in the name Vada that is absent in the 19 specimens. Again, all four lack the flourishes that characterize every one of the 19 specimens. In fact, Davenport pointed out certain inconsistencies in the various flourishes that led him to believe that they did not represent the writer’s normal penmanship and had been deliberately inserted in the 19 specimens prepared for eomparison purposes.
It has frequently been held, for obvious reasons, that specimen signatures written solely for use at the trial are not an admissible basis for comparison. Jones on Evidence (2d Ed.), § 1294. But even if these 19 specimens should be considered admissible they have no persuasive force. Their dissimilarity to the three genuine signatures of May 17, 1955, is too marked to be overlooked. Yet Dr. Stephenson’s conclusions were based upon no other comparison. If he is right in saying that the signature upon the deed under attack is a forgery, then we should have to conclude upon exactly the same reasoning that the three authentic contemporary signatures were also forgeries.
Dr. Stephenson’s testimony is now shown to prove nothing except the fact that the 19 specimen signatures contain characteristics that distinguish them from the signature in issue. With his testimony eliminated the appellee is left, with only her denial of having signed the deed—a denial that is somewhat weakened by the fact that she apparently thought it best to disguise her true signature before submitting it for comparison. The appellee’s testimony is contradicted by her former hus*356band, by the latter’s mother, by the expert Davenport, and to some extent by the notary public. We are compelled to conclude that the appellee did not sustain the burden of proving her case by a preponderance of the evidence.
Reversed.