Bussell, J.
The defendant was charged with burglary. Omitting the proof as to the corpus delicti, it appears from the evidence that a lady heard a noise in a dwelling-house and saw a negro man make his escape by jumping from a window of the house to the ground. The only real question in the case was as to the identity of the culprit. Among other circumstances submitted to show that the accused was the burglar, the court permitted evidence to be introduced as to the conduct of a bloodhound which was put upon tracks on the ground near the window from which the alleged *849burglar made Ms escape. The admissibility and probative value of tMs testimony were properly brought in question at the trial, thus raising an issue of law which, so far as we are aware, has not been adjudicated by a court of last resort in this State. Counsel for the plaintiff in error objected to the admission of the testimony, and also to the instructions of the court as to its consideration by the jury. In substance, the testimony to which objection was made is as follows: The sheriff, in company with others, took the dog to the window from which they had been told the party was seen to jump, and the dog struck a trail by the side of this window and followed it to a place known as the Harris place. There the sheriff pulled the dog off, and he carried her back to the house from which they originally started, and made a circle of the yard, to see that they had the right track. No other trail was found in circling the yard, and the dog was taken back to the window, and she trailed the track to Indian Spring street, and at the end of the street she took the trail down the railroad-track and trailed it until they reached the office at Pepperton. The dog left the railroad-track here and turned to the right four or five steps, and followed the track 100 yards, and then got back on to the railroad-track and followed it to Daniel’s crossing. When the party reached the crossing the sheriff pulled the dog off the •trail and went to Mr. Melton’s. He then carried the dog back to the trail, going south down the railroad until they got to the first wagon road at the north end of Flovilla. The dog here trailed to the left of the railroad until they reached the intersection of the Flovilla and Jackson dirt road, and when she struck the Jackson 'and Flovilla road she went a little to the northwest, back toward Jackson. The trail went to Henry Byar’s house. The sheriff “pulled the dog off and tried to circle around the house.” He left the dog with Mr. Lewis and went for some other parties, and when he got back the dog took the trail again and went back to the Byars house, in which the defendant was found in bed. To this evidence the defendant objected upon the grounds: (1) that it was immaterial and irrelevant; (2) that the proper grounds, had not been laid for the admission of the evidence as to the trailing of the dog; (3) that it was hearsay, and the witness was testifying what he learned from the dog; (4, 5) that the constitution grants to every defendant the right to be confronted, face to face, with every witness who testifies against him; that this was an effort to introduce *850a dog as a witness, and the admission of snch testimony would amount to a denial of the defendant’s constitutional right to be confronted by the witness and to have the witness sworn; (6) that the defendant would thereby be denied the statutory right of cross-examination, and could not inquire how many tracks the dog had counted, whether they were fresh or cold, whether the dog traced them continuously, or lost the track and refound it, or whether the dog was certain about it beyond a reasonable doubt; and (7) that the proper foundation was not laid for the introduction of such evidence.
The court charged the jury that before they could consider evidence as to a track-dog, it must appear that the dog was able at the time to follow a scent of the track of a person, and was certain and reliable; that it would not be sufficient to show that the dog was of "pure-blooded stock characterized for these qualities,” but that it must also be shown that the dog was led to the point where the circumstances tended to’ show clearly that a person had been, and that the tracks trailed by the dog were the tracks which the circumstances showed were tracks made by him; that, "these things having been shown, then you may consider the evidence of the trailing of such tracks as a circumstance in the case, and only as a circumstance, as this evidence alone is not sufficient to convict the defendant. Such evidence must be corroborated by other evidence or other'facts and circumstances that would be sufficient to induce you to believe that the defendant is guilty beyond a reasonable doubt.” These instructions were excepted to on the ground that there was no evidence to connect the defendant with the making of the tracks, or to show that the dog recognized or identified the defendant as the person who made the tracks she had been trailing, and that the judge should not have charged upon the subject of “track-dogs” at all, but should have confined himself to charging the law of circumstantial evidence; for the reason that the charge was not the law of this State, and tended to confuse and mislead the jury and to cause them to believe that “this track-dog was gifted with supernatural powers and could unerringly point out the person who was guilty of the burglary.”
1. It is not to be doubted that there is a deep-seated impression in the popular mind which perhaps unduly magnifies the well-known trait of the bloodhound to distinguish individuals by means of his *851acute sense of smell. As pointed out by Chief Justice Sullivan in Brott v. State, 70 Neb. 395 (97 N. W. 593, 63 L. R. A. 789), “The bloodhound hás, of course, a great reputation for sagacity, and there is a prevalent belief that in the pursuit and discovery of fugitive criminals he is practically infallible. It is a commonly accepted notion that he will start from the place where a crime has been committed, follow for miles the track upon which he has been set, find the culprit, confront him, and, mirabile dictu, by accusing bay and mien declare, ‘Thou art the man.5 This, strange misbelief is with some people apparently incorrigible. It is a delusion which abundant actual experience has failed to dissipate. It lives on from generation to generation. It has still the attractiveness of, a fresh creation. ‘Time writes no wrinkles on its brow.’ ” This traditional confidence in the scent of a bloodhound is perhaps of mythological origin, and its antiquity perhaps inspires awe, but, as a matter of fact, it is unsafe to regard the result of the trailing as an infallible test of identity. Nevertheless, under favorable conditions the trailing of a bloodhound and its conduct may present such independent circumstances as may aid the jury, in the light of other testimony, to determine the question of identity. For this reason Mr. Wigmore, in his work on Evidence (vol. 1, § 177), says: “It seems to be conceded that evidentially the fact that a well-trained bloodhound of good breed, after smelling a shoe or other article belonging to the doer of a crime, has tracked to the accused, is admissible to show that the accused was the doer of the criminal act.” The behavior of the animal to which the testimony of the witness relates may be the result of an impression made on the peculiarly strong sense of smell which such a dog is well known to possess, by a casual event or human act incapable of being perceived by the human senses. For this reason we think that evidence as to the behavior of a dog in following a' track may be received in evidence. But such evidence should be received and considered with great caution; and proof as to the behavior of a dog is entirely valueless unless it be clearly shown that the dog in question is of a breed possessing that peculiarly strong sense of smell, and has had the training which will enable him to discriminate between the tracks of different human beings.
Due to the fact that the path of every human being, at every step from the cradle to the grave, is strewn with the putrescent ex*852cretions of his body, and that the particles of waste matter given off by particular individuals, coming in contact with the olfactory nerves of a dog, produce an impression which the dog is able to recognize as distinct and different from all other impressions, it is possible for a bloodhound sometime to trail a person unerringly for a long distance. But the value of this circumstance as evidence of identity may be diminished or even entirely destroyed by other circumstances which must be considered by the court and the jury. Its value may depend on the freshness of the track, or upon whether the trail lay through open country or followed a road or a street, or crossed such public thoroughfares. In attempting to separate one smell from ten, fifty, or a hundred others with which it was intermixed on a public thoroughfare, it is evident that' the fact that there is no means of inquiring into the reason why the dog selected the one smell rather than the other should be taken into consideration. In Brott v. State, supra, not only was the evidence as to the conduct of the bloodhound rejected as unsafe evidence, but it was held to be “condemned both by reason and instinct,” the Chief Justice saying that if the conviction were sustained, “ courts, in this golden age of enlightenment, would now and again be under the humiliating necessity of adjudging that some person be deprived of his property, his liberty, or his life, because, forsooth . . a certain dog indicated by his conduct that he believed the scent of some miscroscopic particles supposed to have been dropped by the perpetrator of the crime was identical with, or closely resembled, the scent of the person who had been accused and put upon trialand in Pedigo v. Commonwealth, 103 Ky. 41 (116 S. W. 344, 42 L. R. A. 432), and in State v. Adams, 85 Kansas, 435 (116 Pac. 608, 35 L. R. A. (N. S.) 870), in which the evidence as to the conduct of bloodhounds was held to be admissible, there were strong and vigorous dissenting opinions on the part of the minority of each of those courts. There is much to commend in the strong reasoning of Guffy, J., in his dissenting opinion in Pedigo v. Commonwealth, supra, but the principal reasons adduced by him as an argument for excluding the testimony, in our opinion, can profitably be employed so as to safeguard the admission of such testimony by imposing the proper conditions for its use.
Before testimony as to the way a certain dog acted can be of any rightful use, it should appear that the person testifying is reliable, *853that the dog whose actions are to be described was able to scent a track under the given circumstances, and that the dog did follow such scent or track to the location of the defendant. When all this is established it is not proof of guilt, but merely some evidence that the party tracked had been at the place where the crime was committed; the inference that the defendant was at the place in question, drawn from the conduct of the dog, being in the nature of a lesson in location, taught by the exercise of canine instinct. Of course, the weight to be attached to the testimony is for the jury, but, its competency being a question for the court, no evidence a,s to the conduct of the dog should be submitted by the court to the jury without a preliminary investigation on the part of the court '(and preferably in the absence of the jury) as to its admissibility. In Pedigo v. Commonwealth, supra, it was held that “In order to make such testimony competent, even when it is shown that the dog is of pure blood and of a stock characterized by acuteness of scent and power of discrimination, it must also be established that the dog is possessed of these qualities, and has been trained or tested in their exercise in the tracking of human beings, and these facts must appear from the testimony of some person who has personal knowledge thereof. We think it must appear also that the dog so trained and tested was laid on the trail, whether visible or not, concerning which testimony has been admitted, at a point where the circumstances tend clearly to show that the guilty party had been; or upon a track which such circumstances indicated to have been made by him. When so indicated, testimony as to trailing by a bloodhound may be permitted to go to the jury for what it is worth, as one of the circumstances which may tend to connect the defendant with the crime of which he is accused. When not so indicated, the trial court should exclude the entire testimony in that regard from the jury.”
Bearing in mind, as we have already intimated, that the attention of the jury will be called to the fact that the proof of trailing is not proof of guilt, but only a circumstance tending to show whether the defendant had been at the scene of the crime, — a circumstance of no value except as corroborative of evidence of guilt,— we adopt the rule laid down in Pedigo’s case as quoted above. Measured by this rule, we find no error in the admission of the testimony of which complaint is made, and the instructions of the *854able trial judge substantially conform to its requirements. Though the question has not heretofore been ruled in this State, evidence of the location of a criminal by bloodhounds has been held to be admissible, when properly introduced by other testimony, in Richardson v. State, 145 Ala. 46 (41 Sou. 82); Hargrove v. State, 147 Ala. 97 (41 Sou. 972); Davis v. State, 46 Fla. 137 (35 Sou. 76); Denham v. Commonwealth, 119 Ky. 508 (84 S. W. 538); Speers v. State, 92 Miss. 613 (46 Sou. 166); State v. Dickerson, 77 Ohio St. 34 (82 N. E. 969); Parker v. State, 46 Texas Cr. Rep. 461 (88 S. W. 1008); State v. Freeman, 146 N. C. 615 (60 S. E. 986); State v. Spivey, 151 N. C. 676 (65 S. E. 995).
While the apparent exercise of a mysterious power not possessed by human beings is likely to beget in the minds of many people a superstitious awe, and therefore, unless the greatest care be used in laying a proper foundation, the admission of evidence in a criminal case as to the conduct of a bloodhound in following tracks would be dangerous in the extreme, still such evidence is, after all, to be considered, though merely as a circumstance tending to connect the accused with the crime. When it is shown that the bloodhound in question was trained to follow human beings by their tracks, and was previously tested as to its accuracy in trailing, and the jury are instructed that the evidence as to the acts of the bloodhounds is to be received merely as cumulative or corroborative evidence against the person to whom other cireumstances in proof point as being guilty, the admission of this class of evidence is hedged about with such safeguards as that bloodhound evidence is no more dangerous than any other class of circumstantial evidence. Baum v. State, 6 Ohio C. C. (N. S.) 515, 27 Ohio C. C. 569.
When it is considered that in the present case the tracks followed a public street and a public road for a considerable distance, and that the dog did not enter the house in which the defendant was found in bed, or indicate in any way that the trail which she followed from the window of the burglarized house was made by the accused rather than by another negro man who was found in the same room, it is not apparent that the jury based their verdict of guilty to any great extent upon the circumstance of tracking as testified to by the witness upon that point. It is more probable that the result was reached by reason of other testimony in the case, which, in our opinion, was erroneously admitted.
*8552-3. The facts, as well as the complaint presented by tbe remaining exceptions, are stated with such fullness in the second and third headnotes as to require no further elaboration. The point involved in the second headnote is substantially similar to that ruled in Day v. State, 63 Ga. 668. Proof that the lady who occupied the room which was entered was ill and in a delicate condition could only serve to stir the feelings of the jury, since it was alleged that the burglary was committed with the intent to steal, and there is no suggestion of violence in the record. For the reasons stated in the second and third headnotes, the court erred in overruling the motion for a new trial. Judgment reversed.
See footnote on page 817, ante.