SMITH, J.

Appellant was tried under an indictment charging him with the crimes of burglary and grand larceny. He was acquitted on the larceny charge, but was convicted on the burglary count of the indictment and given a sentence in the State Penitentiary.

A bank at Altus had been entered by burglars, who, after entering the vault where the safety deposit boxes were kept, broke into several of these boxes. When the sheriff arrived at the scene he observed finger prints on the boxes, and he caused photographs of these finger prints to be made. Appellant had been confined in the State Penitentiary, and while there his finger prints had been taken, and at his trial W. H. Bennett, who testified that he was an expert in the comparison and identification of finger prints, having devoted many years to that study, further testified that the finger prints on the boxes were identical with those of appellant made at the penitentiary.

The witness Bennett was subjected to a searching cross-examination by counsel for appellant, in the course of which he stated that the finger prints of no two persons were identical. After eliciting this opinion, counsel for appellant moved to exclude it, and assigned as error the refusal of the court so to do.

We think that no error was committed in this ruling, for the reason, first, that counsel's cross-examination brought out the answer. A second reason is that the witness stated that his answer was not an opinion, but was a fact, which had been demonstrated to be true after multiplied thousands of tests, and that he himself had made innumerable tests without finding any finger prints of different individuals to be identical.

The case of Moon v. State of Arizona, 198 P. 288, 22 Ariz. 418, annotated in 16 A.L.R. 362, gives an interesting review of the development of the study of finger prints, and in the annotator's note it is said that the courts have uniformly held that evidence as to the correspondence of finger prints is admissible to prove identity.

The cross-examination of the witness Bennett was devoted to an attempt to discredit the ability of the witness to distinguish between different finger prints, and, upon the redirect examination of the witness by the prosecuting attorney, the record contains the following recitals: "At this point the prosecuting attorney requested the court to permit the witness to make a demonstration to the jury of his ability to compare finger prints. The court granted the request over the objections and exceptions of the defendant. Whereupon the witness, in the presence of the jury and with the assistance of the sheriff, had each of the jurors to place his left thumb on a piece of paper and sign his name thereto. Witness left the room, and the sheriff had Mr. W. T. Blaylock, one of the jurors, to place his left thumb on a piece of paper without his name. Witness was called back and given these pieces of paper. After a few minutes, he said it was Mr. Blaylock's finger print."

Counsel for appellant asked the court to exclude from the consideration of the jury this demonstration, and to admonish the jury to disregard it, and an exception was saved to the refusal of the court to so instruct the jury.

In the case of Moon v. State of Arizona, supra, the exact demonstration permitted in this case was made, and that action was assigned as error; but the Supreme Court of Arizona, after saying that there was no pretense that any trick or device had been employed, held that the test was proper to determine the skill of the witness in identifying finger prints.

In holding that no error had been committed, the court quoted from the brief of counsel for the State as follows: "To a layman, unsophisticated and incredulous, the idea that a finger laid on a clean sheet of paper, leaving no visible trace, thereby leaves a signature upon that paper, absolutely and positively, is a fact startling enough, but to see that finger print development under the fingerprint powder is a demonstration impressive and convincing. It might well be that, until a juryman witnessed this demonstration, he would never believe that a plain porcelain slab would reveal the incriminating finger print, but, having seen their own finger prints developed from invisible impressions on sheets of paper, it was no longer a question of speculation; it was to the juryman a fact as commonplace as radium, or wireless, or flying in the air."

The court further said that, "for obvious reasons, the admission of experimental testimony must largely rest in the discretion of the trial judge, and the exercise of this discretion will not be controlled, unless it is manifestly abused."

See also the annotated case of State of Nevada v. Kuhl, 175 P. 190, 3 A.L.R. 1694, 42 N.W. 185, where a somewhat similar demonstration was made in the presence of the jury, and the testimony held admissible, after an extensive review of the authorities.

It has always been held competent to examine a witness who proposes to testify as an expert, touching his qualifications as such.

At 62 of the chapter on "Expert and Opinion Evidence," in 11 R.C.L. page 646, it is said that: "He (the expert) may also be subjected to tests to determine the value of his opinion. Thus a handwriting expert has often been required to pass on other signatures submitted to him on the witness stand, though they were not relevant to the case." There is a conflict in the authorities, however, as to whether this examination should be confined to writings already in evidence.

The presiding judge wrote into the bill of exceptions the following statement:

"While the jury were deliberating on the case of the State of Arkansas v. Willard Hopkins, and after the same had been submitted to then, and after they had been deliberating for a considerable period of time, without having reached a verdict, one of their number, Mr. Will Mullens, as I understand his name, came to me and asked me if they could find the defendant guilty and recommend in their verdict that his sentence be suspended during the time he remained within the jurisdiction of the court and supported his widowed mother. I told him they could do this, but that it was not binding on the court. This was in the absence of the defendant, his counsel and the other jurors. Later in that day another member of the jury on this case, Mr. Eli Turner, as I understood his name, came to me and asked me practically the same question that Mr. Mullens did, and I told him that they could. This was also in the absence of the defendant, his counsel, and the other members of the jury. I further certify that, shortly thereafter, the jury brought into open court a verdict finding the defendant guilty of burglary and not guilty of grand larceny, with the recommendation in accordance with what I told them they could do."

The Attorney General has virtually confessed that this proceeding was erroneous, and such we think it was.

In the case of Pearson v. State, 119 Ark. 152, 178 S.W. 914, a member of the jury addressed the following note to the presiding judge: "If the jury should find the defendant guilty as charged in the indictment with a recommendation for leniency, has your Honor the authority and will you assess his punishment at twenty-one years in the State Penitentiary, or for life?" The judge answered the above note by writing thereon the word "No." When this communication was had between the court and the jury, neither the defendant nor his attorneys were present, and they knew nothing about the incident. It was there said:

"Therefore, treating the exception as true, the court erred in communicating with the jury in the manner set up in this exception, in the absence of the defendant and his counsel. This inquiry on the part of the jury and the answer thereto by the court was tantamount to giving instructions to the jury in the absence of the defendant and his counsel. If the appellant or his counsel had been present, then they might have objected to the court's answering the inquiry in any manner at all, and they might have objected to the answer that the court gave. It is unnecessary to determine whether the answer was correct. In Kinnemer v. State, 66 Ark. 206, 49 S.W. 815, the court re-read the instructions to the jury in the absence of the defendant exactly as at first given before the jury retired to consider of their verdict. Of this procedure, we said: `The instructions could not be re-read in his absence, for, although they were read exactly as at first given, the defendant had the right to know and see that such was the case, and to be present for that purpose'" (citing authorities).

In the recent case of Wacaster v. State, 172 Ark. 983, 291 S.W. 85, it was said:

"Section 3192, Crawford Moses' Digest, provides how a jury, after it has retired for deliberation, shall acquire information on any point of law or about any part of the evidence, if there is a disagreement, that they must require the officer to conduct them into court, where the information required must be given in the presence of, or after notice to, the counsel of the parties. Its provisions are mandatory."

See also Wawak and Vaught v. State, 170 Ark. 329, 279 S.W. 997, and Aydelotte v. State, 170 Ark. 1192, 281 S.W. 369.

The court should not have charged the individual jurors as to the law of the case in the absence of the accused and his counsel, and, for this error, the judgment must be reversed, and the cause will be remanded for a new trial.