Carr v. State

• Decision Date: 10 March 1997
• Docket Number: No. S96A1626,S96A1626
• Citation: 482 S.E.2d 314,267 Ga. 701
• Parties: , 65 USLW 2663, 97 FCDR 1259, 97 FCDR 819 CARR v. The STATE.
• Court: Georgia Supreme Court

Page 314

482 S.E.2d 314

267 Ga. 701, 65 USLW 2663, 97 FCDR

1259, 97 FCDR 819

CARR v. The STATE.

No. S96A1626.

Supreme Court of Georgia.

March 10, 1997.

Reconsideration Denied April 3, 1997.

Millard C. Farmer, Jr., Joseph M. Nursey, Michael A. McKenzie, Cozen & O'Connor, Donald Franklin Samuel, Garland, Samuel & Loeb, P.C., Atlanta, for Weldon Wayne Carr.

Lewis R. Slaton, District Attorney, Rebecca A. Keel, Carl P. Greenberg, Assistant District Attorneys, Michael J. Bowers, Attorney General, Paula K. Smith, Senior Assistant Attorney General, Atlanta, for State.

BENHAM, Chief Justice.

This appeal is from Weldon Wayne Carr's convictions of malice murder and first degree arson ¹. Carr's wife died in a fire in their home. He recounted to authorities that he awoke and realized there was a fire downstairs in their home; that he tried to get his wife to escape with him through a bedroom window, but she resisted and tried to go toward the fire; and that he lost her in the smoke and confusion after a struggle, but finally saved himself by jumping out of a second-story window. The State's theory of the case was that Carr set the fire, then injured his wife so that she could not escape. Although she died of smoke inhalation, Carr's wife suffered other injuries, including cerebral bleeding. Prior to the fire, the couple had been experiencing marital difficulty and had been seeing a marriage counselor. Ms. Carr was having an affair, of which Carr had learned, and she had told several persons that she intended to divorce Carr and marry her lover. In a short period before the fire, Carr engaged in conduct which appeared suspicious after the fire: checking on fire insurance; getting copies of his and his wife's will; telling their adult son, who resided elsewhere, to remove some of his belongings from the family home; putting valuables into a safe deposit box; and conducting an uncharacteristic spring cleaning of the house. To support its theory that Carr used an accelerant to start the fire, the State presented evidence that burn patterns at the site of the fire's origin were consistent with the use of an accelerant, and presented testimony that a dog trained to indicate the presence of chemical accelerants had alerted at the same site.

On appeal, Carr contends the trial court erred in the following ways: permitting evidence of a dog alerting to the possible presence of an accelerant as substantive evidence of the presence of an accelerant; ordering certain pre-trial discovery; admitting into evidence inadmissible hearsay testimony; improperly admitting bad character evidence; excluding defense evidence regarding the reliability of dog alert evidence; erroneously qualifying witnesses as experts; permitting testimony tainted by a warrantless search; refusing to record the entire proceeding; and improperly limiting the evidence presented on motion for new trial. He also contends his convictions must be reversed because of prosecutorial misconduct and because the evidence presented at trial was insufficient to support the convictions. After a review of the voluminous record and the extensive argument of counsel, we conclude that the convictions must be reversed for the reasons which follow.

1. The question of whether an accelerant had been used in the Carr house was an important point in the State's case. Testimony from fire fighters was used to show that the progress of the fire and burn patterns in the house were consistent with the presence of an accelerant, but the State Crime Lab report on fire-debris sent to it was negative for the presence of accelerants. However, the State proffered evidence that Blaze, a dog trained to give an alert when he smelled certain hydrocarbons, had given such an alert at the site in the house where investigators had already come to believe an accelerant was used. Carr argues on appeal that the evidence was inadmissible because there was no evidence at trial that dog-alerts have reached a state of verifiable certainty as required by Harper v. State, 249 Ga. 519(1), 292 S.E.2d 389 (1982), and that the error was harmful because it was the only substantive evidence purporting to show the presence of an accelerant. We agree.

Prior to permitting the testimony about the dog's alerts, the trial court conducted a hearing at which the State put on witnesses vouching for the reliability of the dog's alerts. The trial court concluded finally that the evidence should be admitted, ruling that the test in Harper, supra, was not applicable, but that if it were, the evidence submitted at the hearing was sufficient to meet that test.

The question of what type of evidence constitutes scientific test evidence which should be subjected to the Harper test has not been directly decided by the appellate courts of Georgia, but the test has been applied to a wide range of evidence: a possible genetic basis for violent and impulsive behavior in certain individuals (Mobley v. State, 265 Ga. 292(2), 455 S.E.2d 61 (1995)); breathalyzer tests (Lattarulo v. State, 261 Ga. 124(3), 401 S.E.2d 516 (1991)); the phenomenon of sleep talk (Godfrey v. State, 258 Ga. 28, 365 S.E.2d 93 (1988)); analysis of fibers (Williams v. State, 251 Ga. 749(1), 312 S.E.2d 40 (1983)); urinalysis for the presence of cannabinoids, using the SYVA EMIT 1000 test (Smith v. State, 250 Ga. 438, 298 S.E.2d 482 (1983)); penile plethysmograph evidence (Gentry v. State, 213 Ga.App. 24(2), 443 S.E.2d 667 (1994)); the "ontrack system" for detecting alcohol and controlled substances (Hubbard v. State, 207 Ga.App. 703, 429 S.E.2d 123 (1993)); horizontal gaze nystagmus test for intoxication (Manley v. State, 206 Ga.App. 281, 424 S.E.2d 818 (1992)); and the "child sexual abuse accommodation syndrome." Rolader v. State, 202 Ga.App. 134(2), 413 S.E.2d 752 (1991).

What these cases have in common is expert opinion based on an analysis of data, an opinion that could only be based on something more than mere observation. Of course, if the conclusion at issue could be drawn by anyone based on observation, there would be no need for expert testimony. "Expert opinion testimony ... is admissible where the conclusion of the expert is one which jurors would not ordinarily be able to draw for themselves; i.e., the conclusion is beyond the ken of the average layman. [Cits.]" Smith v. State, 247 Ga. 612, 619, 277 S.E.2d 678, 683 (1981). Applying that rationale to the evidence involved in this case, it is plain that the dog alert testimony was expert testimony in that the average layperson would not be able to determine from watching the dog lie down, point with his nose, or paw the ground (the methods used by the dog to indicate the presence of hydrocarbons) that chemicals which could accelerate a fire were present. It is only by the dog handler's analysis of the dog's behavior that the conclusion could be reached that an accelerant was present.

The dog handler's testimony concerning the presence of accelerants, based on the dog's alert, is conceptually indistinguishable from the other types of tests and analyses mentioned above. Therefore, the analysis and data gathering leading to the testimony should be subject to the same requirements of scientific verifiability applicable to the other procedures. See Harper, supra. To hold otherwise would make the dog alert procedure analogous to the earliest recorded lie detector, which involved touching a donkey's tail. ²

The trial court, in ruling the testimony admissible, concluded that if Harper did apply to this testimony, the requirements of that case were met by the State's evidence at the hearing held to consider admission of the testimony. The foundational evidence on the subject consisted of a recital of the dog's and the handler's training, background information on the use of dogs to detect chemical accelerants, and anecdotal evidence of the dog's success in tests. What was not presented was a showing "with verifiable certainty that these tests are an accurate and reliable means of ascertaining whether [an accelerant is actually present]." Harper v State, supra at 526, 292 S.E.2d 389. The State also offered the trial court and has offered this court case authority permitting the use of dog alert evidence, attempting to satisfy the provision in Harper that "[o]nce a procedure has been recognized in a substantial number of courts, a trial judge may judicially notice, without receiving evidence, that the procedure has been established with verifiable certainty, or that it rests upon the laws of nature." Id. at 526, 292 S.E.2d 389. The proffered authority, however, did not involve the use of dog alert testimony to show the actual presence of a substance. The State has not cited, and we have not found, any case in Georgia in which there has been such holding. ³ While the use of trained dogs can be a valuable part of investigative procedures and can provide important elements of probable cause to search (Bothwell v. State, 250 Ga. 573(7), 300 S.E.2d 126 (1983); Carter v. State, 222 Ga.App. 345(1), 474 S.E.2d 240 (1996)), dog alerts to accelerants have not been shown, neither at the trial of this case nor in any Georgia appellate decision, to have the scientific reliability necessary to permit their use as substantive evidence of the presence of accelerants. The trial court's ruling to the contrary was error. The State argues that the admission of the evidence, if error, was harmless in light of other evidence of the presence of an accelerant. However, there was no other direct evidence of the presence of an accelerant, and thus, no direct evidence of arson. Notwithstanding the trial court's instruction that the evidence of the dog alert must be considered along with other evidence, we conclude that the potential impact of the evidence admitted was too great for us to conclude that no harm to Carr's right to a fair trial flowed from it. The erroneous admission of that evidence requires a new trial.

2. In...