DOLLIVER, J.

Allen Boyd Loucks appeals his conviction in King County Superior Court for second degree burglary. Defendant's conviction rested entirely on evidence provided by police dog Tally of the Seattle Police Department's canine unit. Loucks claims the evidence presented against him was insufficient as a matter of law to sustain his conviction. We agree with defendant that dog tracking evidence standing alone is insufficient to support a criminal conviction. Therefore, we reverse defendant's conviction.

In the early morning hours of May 13, 1979, a burglary occurred at 1147 Harvard Avenue East, Seattle, the home of Robert Cochran, Bishop of the Episcopal Dikcese of Olympia. Bishop Cochran and his wife testified they awoke to the sound of banging or hammering noises within their home. Shortly after being awakened by the noises, the Cochrans realized their residential security system indicated an unauthorized entry into their home had occurred. The Cochrans called the Seattle Police Department when they determined the intruder was still present in the house. Neither of the Cochrans saw the intruder nor did they hear any voices.

When the police arrived, they conducted a thorough search of the Cochran residence. The police discovered an open window on the main floor, as well as an open door to the basement of the house. The glass panel in the basement door was broken. Blood smears were found near the basement door. Inside the house, the police found more blood smears in the main floor library, as well as fingerprints. The police found the intruder had moved an oak prayer desk from the basement to the back patio of the house.

Soon after the police department's initial response to the Cochrans' request for help, Officer Beard and police dog Tally from the canine unit arrived on the scene. Officer Beard gave Tally the "search command", and the dog wandered through the entire residence and apparently picked up a scent in the basement. Tally exited through the exterior basement door, circled the prayer desk on the patio, then left the Cochran property, with Officer Beard in pursuit.

Tally led Officer Beard to a residence just south of the Cochran house. Tally found defendant lying at the bottom of a stairwell and bit him on the leg. The police arrested Loucks and charged him with burglary of the Cochran residence.

Defendant was not a stranger to the Seattle police officers on duty in the Capitol Hill area that evening. Approximately 1 hour before they responded to the Cochran burglary, Seattle police officers stopped and ticketed Loucks for jaywalking. At the burglary trial, the police testified they confronted Loucks at the corner of Broadway and Roy Streets after he crossed Broadway against the traffic light. Officer Brandner testified Loucks appeared to have been drinking but did not appear intoxicated. Both Officers Ayco and Brandner testified defendant Loucks showed them a piece of paper with an address on it and claimed to be looking for the address. While neither officer remembered the address, both testified it was not in the Capitol Hill area.

Laboratory analysis of the blood smears and fingerprints found at the Cochran residence showed neither belonged to the defendant. Nevertheless, Loucks was charged and convicted of second degree burglary. The prosecution theory of the case was that Loucks and an unknown accomplice committed the burglary. The prosecution's case consisted of the Cochrans' testimony describing their home and the noises they heard, the testimony of the arresting officer, and the testimony of Tally's handler. The only evidence placing Loucks within the Cochran home was that provided by Tally's handler.

[1, 2] Although there is a division of authority, we believe dog tracking evidence should be admissible where a proper foundation is made showing the qualifications of dog and handler. We concur with and adopt those conditions precedent to admissibility spelled out by the Court of Appeals in State v. Socolof, 28 Wn. App. 407, 411, 623 P.2d 733 (1981):

(1) the handler was qualified by training and experience to use the dog, (2) the dog was adequately trained in tracking humans, (3) the dog has, in actual cases, been found by experience to be reliable in pursuing human track, (4) the dog was placed on track where circumstances indicated the guilty party to have been, and (5) the trail had not become so stale or contaminated as to be beyond the dog's competency to follow.

Accord, People v. Craig, 86 Cal.App.3d 905, 915-16, 150 Cal. Rptr. 676 (1978); People v. Harper, 43 Mich. App. 500, 508, 204 N.W.2d 263 (1972). We believe, however, that dog tracking evidence must be supported by corroborating evidence.

The question whether dog tracking evidence standing alone is sufficient to support a conviction is a novel one for the Washington courts. State v. Socolof, supra. Although in Socolof the Court of Appeals intimated properly introduced tracking evidence might be sufficient evidence on which to base a conviction, in that case the victim also identified the defendant as her assailant. See Socolof, 28 Wn. App. at 409. Other courts, however, generally hold tracking evidence is not of itself sufficient to support a conviction. E.g., People v. McPherson, 85 Mich. App. 341, 271 N.W.2d 228 (1978); State v. Cheatham, 458 S.W.2d 336, 339 (Mo. 1970). See State v. Taylor, 118 N.H. 855, 858, 395 A.2d 505 (1978). See generally Annot., Evidence of Trailing by Dogs in Criminal Cases, 18 A.L.R.3d 1221, § 6 (1968).

A large majority of jurisdictions allowing introduction of dog tracking evidence do so only after other evidence has been introduced clearly connecting the accused with commission of the crime for which he is being prosecuted. See, e.g., Terrell v. State, 3 Md. App. 340, 353, 239 A.2d 128 (1968). Courts which allow dog tracking evidence regard its probative value with some suspicion. See generally Comment, Evidence — Admissibility in Criminal Prosecution of Evidence of Tracking by Bloodhounds Indicating Guilt of Accused, 9 Wn. Lee L. Rev. 248, 251 (1952); 1 J. Wigmore, Evidence § 177, at 635-37 (3d ed. 1940). Most courts allowing dog tracking evidence restrict its use to corroborative purposes only. E.g., State v. Fixley, 118 Kan. 1, 5, 233 P. 796 (1925). In People v. McPherson, supra, the Michigan Court of Appeals stated that dog tracking evidence is cumulative only, and must be supported by corroborating evidence to sustain a conviction. McPherson, 85 Mich. App. at 344-45. See also 30 Am.Jur.2d Evidence § 1146 (2d ed. 1967).

The dangers inherent in the use of dog tracking evidence can only be alleviated by the presence of corroborating evidence identifying the accused as the perpetrator of the crime. Police dogs cannot be conclusively relied on to follow the trail of one individual if other human trails cross this one, or even come near it. Craig, The Dog as a Detective, 18 Sci. Monthly 38, 39 (1924), cited in 1 J. Wigmore, Evidence, supra at 637. While a dog's trainer may be available for cross examination, he obviously will be unable to answer many questions bearing on the reliability of the dog's conclusions. Comment, 9 Wn. Lee L. Rev. at 253-54. For a pungent analysis of the capability of tracking dogs see Brott v. State, 70 Neb. 395, 397-98, 97 N.W. 593 (1903).

In the present case, evidence presented by the State failed to corroborate tracking evidence provided by Officer Beard and police dog Tally. The State's fingerprint and blood type evidence tend to exculpate defendant Loucks rather than corroborate Tally's identification of him as the intruder in the Cochran home. As previously stated, laboratory tests showed defendant's blood type did not match the blood found at the scene of the burglary. Likewise, defendant's fingerprints did not match fingerprints discovered at the Cochran residence.

The State's theory that an accomplice was involved rests on its contention the Cochrans' prayer desk was too large for one person to move and evidence of two break-in points at the Cochran residence. In its opening statement, the prosecution referred to the prayer desk as "a large oak prayer desk, which is very heavy". In contrast, Officer Ayco, who investigated the Cochran burglary, described the furniture as "a pulpit, a little brown kneeling desk". The only evidence presented regarding the size of the desk was the statement by Bishop Cochran that "[i]ts dimensions are about 33 inches high, 30 inches wide and about 2 feet deep. It has a kneeling pad. It's a pretty solid piece of furniture." The State presented no evidence at all regarding the weight of the prayer desk.

The evidence presented regarding the two break-in points is equally inconclusive. Bishop Cochran testified the interior doors to the basement were locked so an intruder entering the basement could not then enter the main floor of the house without reentering from outside. The State presented no evidence the intrusions were made by different individuals or that more than one person passed through either point of entry.

Finally, the State asserts Loucks' presence in the Capitol Hill area after telling police officers he was looking for an address in another part of the city somehow corroborates Tally's identification of Loucks. We reject the State's argument. Defendant's unexplained presence in the Capitol Hill vicinity is susceptible to too many constructions to constitute evidence of his complicity in the Cochran burglary. See State v. Cheatham, 458 S.W.2d at 339. See also In re Wilson, 91 Wn.2d 487, 491-92, 588 P.2d 1161 (1979); State v. J-R Distribs., Inc., 82 Wn.2d 584, 593, 512 P.2d 1049 (1973).

Tally's identification of defendant Loucks as the perpetrator of the Cochran burglary is unsupported by other evidence linking Loucks to the burglary. Standing alone, Tally's identification of Loucks as the burglar is insufficient to sustain his conviction.

Reversed.

BRACHTENBACH, C.J., STAFFORD, UTTER, WILLIAMS, DORE, DIMMICK, and PEARSON, JJ., and CUNNINGHAM, J. Pro Tem., concur.