I dissent.1 First, as a matter of law, I do not believe that the evidence is sufficient to prove beyond a reasonable doubt that appellant committed the burglary; second, even assuming the sufficiency of the evidence as a matter of law, I believe the three instructional errors (two of which are conceded by the majority) were highly prejudicial in the light of the whole record.
*243The only direct evidence connecting the appellant to the burglary was the fact that Sarge tracked to him. The evidence that corroborated Sarge was: (1) the unlikely location of the appellant late at night; (2) appellant’s trousers were wet at the leg bottoms and his shoes were muddy and grass-stained, which could lead to an inference that he ran through the high grass of the game preserve; (3) appellant was perspiring and out of breath; and (4) a pair of pliers and a penlight were found in the same general direction from the Cox residence as the appellant.
We start with the proposition that the identity of the appellant must be proven beyond a reasonable doubt. (People v. Law (1974) 40 Cal.App.3d 69, 85 [114 Cal.Rptr. 708].) The reasons for which the minority of states2 exclude dog-tracking evidence per se provide ample ground for requiring human corroboration and cautionary instructions: (1) the motivation actions of the dog are unreliable; (2) the evidence constitutes hearsay; (3) the defendant is deprived of the constitutional right to be confronted by witnesses against him; (4) the defendant should not be placed in jeopardy by the actions of an animal; (5) a jury might give much greater weight and importance to the evidence than it warrants because of the tracking ability which people ascribe to dogs and the affection with which dogs are regarded.
Where, as here, dog-tracking evidence is the only evidence upon which a conviction could be based, great care must be exercised in its admission and the evidence excluded if an adequate foundation is not laid. (1 Wigmore, Evidence (3d ed. 1940) § 177, p. 633.) I agree that “[h]uman life and liberty are much too sacred and too highly regarded in this country to be jeopardized or taken from the citizen, except upon testimony carrying conviction beyond a reasonable doubt.” (Meyers v. Commonwealth (1922) 194 Ky. 523 [240 S.W. 71, 74].) I believe that “there must be other and human testimony. ” (Carter v. State (1914) 106 Miss. 507 [64 So. 215], italics added.)
*244As the majority concedes (ante, p. 239), the states that follow the majority rule of admissibility3 agree with People v. Craig, supra, 86 Cal.App.3d at page 918, that “[d]og trailing, by itself, is not sufficient to warrant conviction .... [T]he evidence could only be sufficient to warrant conviction if supported by other evidence.” (Ibid., italics in original; State v. Fixley (1925) 118 Kan.l [233 P. 796]; Meyers v. Commonwealth, supra, [240 S.W. 71]; Daugherty v. Commonwealth (1943) 293 Ky. 147 [168 S.W.2d 564]; State v. Freyer (1932) 330 Mo. 62 [48 S.W.2d 894]; Carter v. State, supra, 64 So. 215; People v. Centolella, supra, 305 N.Y.S.2d 279, 282; People v. McPherson (1978) 85 Mich.App. 341 [271 N.W.2d 228].)
Appellant’s location, perspiration, lack of breath, and the leaves on his jacket as easily corroborate his alibi as the prosecution’s case (i.e., if the appellant was fleeing the scene of the crime, why was he only 7/10 of a mile away 45 minutes later?). His trouser legs could have become damp and his feet muddy from trying to avoid Sarge, not because of the burglary but because the dog had come on his scent in the game reserve. The fact that the unidentified pair of *245pliers and penlight were found along the general route that Sarge took merely indicates that the burglar headed in the same general direction as the appellant.
The majority also relies, for corroboration, on the proposition that the jury, obviously disbelieving appellant’s version as to how he came to be in the bushes, could have inferred that appellant, by testifying falsely, showed a consciousness of guilt, and cites People v. Wayne (1953) 41 Cal.2d 814, 823 [264 P.2d 547], in support thereof.
Wayne is a case where the defendant was convicted of soliciting a bribe from a bookmaker for the purpose of “buying protection” for the bookmaker from the local police. The bookmaker paid the money through a third party to the defendant. At trial, the bookmaker and the third party testified to the payment of the money to defendant and the purpose for which it was paid. The defendant testified, admitting the receipt of the money, denied it was in any way connected with bookmaking or bribery, and gave an explanation as to why he was being paid by the bookmaker which the Supreme Court characterized as “rather peculiar” (41 Cal.2d at p. 821) and “so unusual” (41 Cal.2d at p. 823).
Because the bookmaker and the go-between were accomplices, as a matter of law, their testimony had to be corroborated before a conviction could be obtained. It is in this context that the Wayne court made the observation concerning consciousness of guilt quoted by the majority (ante, p. 240).
The difference between the case at bar and Wayne is apparent. In Wayne, there was direct prosecution testimony that money had been paid to defendant for the purpose of bribing the police. The defendant admitted that he had received the money but gave what appeared to be a contrived answer as to why the money was given to him. A jury disbelieving the defendant’s testimony could well believe the prosecution testimony and infer that defendant, by lying about how the money came into his possession, showed a consciousness of guilt. In the instant case, even if the jury believed that appellant deliberately lied about how he came to be in the bushes, does that show a consciousness of guilt of a burglary committed about an hour earlier at a residence almost a mile away? Suppose that the police in searching the area had found appellant in those same bushes without the tracking dog. Suppose further that at trial, appellant had given the identical lame explanation for his presence there. Would that show a consciousness of guilt that would support a conviction? (See People v. Bamber (1968) 264 Cal.App.2d 625 [70 Cal.Rptr. 662], and cases there cited.)
Considering the record as a whole, I am of the view that the evidence corroborative of the dog’s tracking is, at best, fragmentary and insubstantial. On the other hand, I would agree that a reasonable suspicion that appellant was the burglar does arise from the evidence, but suspicion is one thing—proof beyond *246a reasonable doubt is another. The means do not justify the ends. In all the cases relied upon by the majority, there was much stronger evidence as to the identity of the perpetrator in addition to the dog-tracking.
Assuming that there was sufficient evidence of the identity of the appellant to meet the burden of proof, I cannot agree with the majority that the court was not required to instruct, sua sponte, that the dog-tracking evidence is of little probative value and must be viewed with caution. While I agree with the majority that a sua sponte instruction is required as to the necessity for other direct or circumstantial evidence of the identity of the defendant (ante, p. 242), I would hold further that whenever the evidence is admitted after a proper foundation has been laid, the jury must be instructed to view it with caution. (Cf. State v. Taylor, supra, 395 A.2d 505, 507.) I do not believe that the scientific validity of dog-tracking evidence has been demonstrated even as well as voiceprinters; exercise of restraint is therefore warranted. (See People v. Kelly (1976) 17 Cal.3d 24, 32 [130 Cal.Rptr. 144, 549 P.2d 1240].) I am concerned with the matter of undue weight as evidence gleaned from the efforts of dogs has been part of our folklore for centuries.4
It comes down to this: Appellant has been convicted on the basis of a dog’s ability to scent5 corroborated by: (1) a pair of pliers linked to the burglary but not connected to appellant; (2) a penlight connected neither to the burglary nor to appellant; (3) a dubious “consciousness of guilt”; and (4) appellant’s presence in the bushes one hour after the burglary and over seven-tenths of a mile from the burglarized house.
*247As against that, as conceded by the majority herein: (1) a sua sponte instruction should have been given but was not; (2) a “flight” instruction was given but should not have been given; and (3) in my view, an instruction should have been given that dog-tracking evidence should be viewed with caution.
The judgment should be reversed and the case dismissed because the prosecution did not sustain its burden of proof, as a matter of law. At the least, the error in instructions requires a reversal because it was prejudicial.
Appellant’s petition for a hearing by the Supreme Court was denied March 23, 1983. Bird, C. J., and Grodin, J., were of the opinion that the petition should be granted.
I do agree with the majority’s adoption of specific criteria that must be shown before dog-trailing evidence is admissible (ante, at p. 238) since this matter was not elucidated in People v. Craig (1978) 86 Cal.App.3d 905, 917-918 [150 Cal.Rptr. 676].
The following jurisdictions hold that dog-tracking evidence is inadmissible per se:
Illinois: People v. Pfanschmidt (1914) 262 Ill. 411 [104 N.E. 804] (bloodhounds held to be unreliable).
Indiana: Ruse v. State (1917) 186 Ind. 237 [115 N.E. 778] (evidence too uncertain).
Iowa: State v. Grba (1923) 196 Iowa 241 [194 N.W. 250] (evidence weak and uncertain).
Montana: State v. Storm (1951) 125 Mont. 346 [238 P.2d 1161] (evidence incompetent).
Nebraska: Brott v. State (1903) 70 Neb. 395 [97 N.W. 593] (evidence unsafe).
New York: People v. Centolella (1969) 61 Misc.2d 726 [305 N.Y.S.2d 279.]
Minnesota declined to pass upon the question of admissibility but held that in Crosby v. Moriarty (1921) 148 Minn. 201 [181 N.W. 199], no proper foundation had been laid for the evidence.
See Annot., Evidence of Trailing by Dogs in Criminal Cases (1968) 18 A.L.R.3d 1221 et seq.
The majority of jurisdictions admit dog-tracking evidence with the limitations indicated:
Alabama: Burks v. State (1941) 240 Ala. 587 [200 So. 418].
Arizona: State v. Coleman (1978) 122 Ariz. 130 [593 P.2d 684].
Arkansas: Rolen v. State (1936) 191 Ark. 1120 [89 S.W.2d 614].
California: People v. Craig (1978) 86 Cal.App.3d 905 [150 Cal.Rptr. 676].
Delaware: Cook v. State (Del.Sup. 1977) 374 A.2d 264.
Florida: Tomlinson v. State (1937) 129 Fla. 658 [176 So. 543].
Georgia: Mitchell v. State (1947) 202 Ga. 247 [42 S.E.2d 767].
Kansas: State v. Netherton (1931) 133 Kan. 685 [3 P.2d 495],
Kentucky: Daugherty v. Commonwealth (1943) 293 Ky. 147 [168 S.W.2d 564].
Louisiana: State v. Green (1946) 210 La. 157 [26 So.2d 487].
Maryland: Terrell v. State (1968) 3 Md.App. 340 [239 A.2d 128].
Massachusetts: Commonwealth v. LePage (1967) 352 Mass. 403 [226 N.E.2d 200].
Michigan: People v. Harper (1972) 43 Mich.App. 500 [204 N.W.2d 263]; People v. Perryman (1979) 89 Mich.App.516, 524 [280 N.W.2d 579] (cautionary sua sponte instructions required).
Mississippi: Hinton v. State (1936) 175 Miss. 308 [166 So. 762].
Missouri: State v. Fields (Mo. 1968) 434 S.W.2d 507.
New Hampshire: State v. Taylor (1978) 118 N.H. 855, [395 A.2d 505] (admissible where foundational elements met; jury must be instructed to view it with caution, etc.).
North Carolina: State v. Rowland (1965) 263 N.C. 353 [139 S.E.2d 661] (dog’s pedigree required in addition to other foundational elements).
Ohio: State v. Dickerson (1907) 77 Ohio St. 34 [82 N.E. 969].
Oklahoma: Buck v. State (1943) 77 Okla.Crim. 17 [138 P.2d 115].
Pennsylvania: Commonwealth v. Hoffman (1913) 52 Pa.Super. 272.
South Carolina: State v. Brown (1916) 103 S.C. 437 [88 S.E. 21],
Tennessee: Copley v. State (1926) 153 Tenn. 189 [281 S.W. 460]; State v. Barger (Tenn.Crim. 1980) 612 S.W.2d 485.
Texas: Parker v. State (1904) 46 Tex.Crim 461 [80 S.W. 1008],
Vermont: State v. Bourassa (1979) 137 Vt. 62 [399 A.2d 507].
Washington: State v. Socolof (1981) 28 Wn.App. 407 [623 P.2d 733] (admissible where all five foundational elements met).
West Virginia: State v. McKinney (1921) 88 W.Va. 400 [106 S.E. 894].
As noted in Blair v. Commonwealth (1918) 181 Ky. 218, 220-221 [204 S.W. 67, 68]: “If we may credit Sir Walter Scott, such evidence was looked upon with favor as early as the twelfth century. In ‘The Talisman’ it is related that in the joint crusade of Richard I of England and Philip H of France, Roswell, the hound, pulled from the saddle Conrade, Marquis of Montserrat, thus mutely accusing him of the theft of the banner of England. Philip defended the marquis with the remark, ‘Surely the word of a knight and a prince should bear him out against the barking of a cur. ’ To which Richard replied, ‘Royal brother, recollect that the Almighty, who gave the dog to be companion of our pleasures and our toils, both invested him with a nature noble and incapable of deceit. He forgets neither friend nor foe—remembers, and with accuracy, both benefit and injury. He has a share of man’s intelligence, but no share of man’s falsehood. You may bribe a soldier to slay a man with his sword, or a witness to take life by false accusation; but you cannot make a hound tear his benefactor; he is the friend of man save when man justly incurs his enmity. Dress yonder Marquis in what peacock robes you will, disguise his appearance, alter his complexion with drugs, and washes, and hide himself amidst a hundred men, I will yet pawn my sceptre that the hound detects him, and expresses his resentment, as you have tins day beheld.’” (Italics added.)
While the dog may have remained man’s best friend through the ages, I would like to think that the law has made some progress since the reign of Richard I of England from 1189-1199.
We observe that it took the dog about 35 to 40 minutes to track down appellant, a distance of about seven-tenths of a mile. The explanation seems to be that the wind blew the scent around and the dog, apparently, would lose it and then have to nose around to pick up the scent again. Obviously, this would seem to detract from the reliability of the dog-tracking here.