COMMONWEALTH vs. RICHARD P. COMOLLI.

GREANEY, J. This is the Commonwealth's interlocutory appeal from the order of a District Court judge, which allowed the defendant's motion to suppress oral statements made to a State police trooper at the scene of an automobile accident. G. L. c. 278, Section 28E, inserted by St. 1979, c. 344, Section 45. Mass.R.Crim.P. 15(a) (2), 378 Mass. 882 -883 (1979). The judge filed a document entitled "Ruling of the Court" which contains only skeletal references to the facts surrounding the statements and which concludes that the statements

Page 608

were elicited contrary to the requirements of Miranda v. Arizona, 384 U.S. 436 (1966).

At the hearing on the motion, State police Trooper John Ross testified that he arrived at the scene of a two-vehicle accident on Route 3 in Bedford at approximately 5:15 A.M. on December 20, 1980. Both vehicles were heavily damaged and neither could be driven. Bedford police officers told Ross that the driver of one vehicle had been transported to a hospital and that the other driver, the defendant, was sitting in a Bedford police cruiser. The temperature was near zero. Ten to fifteen minutes after he arrived, Trooper Ross spoke to the defendant for the first time, asking him to come to his cruiser. In the cruiser, the trooper asked the defendant if he needed medical care. The defendant declined. The defendant stated that he had not seen the other car until just before his vehicle struck it in the rear. He also identified the vehicle he had been operating and said that while he was not sure exactly how fast he had been going, he normally travelled on Route 3 at a speed of sixty to sixty-five miles per hour. The conversation with the defendant was interrupted at least twice when the trooper left the cruiser to speak with Registry Inspector Robert Moreschi and a police photographer as they arrived separately. Trooper Ross testified that he never told the defendant that he could or could not leave the scene and he was unsure as to how the defendant eventually left.

The trooper stayed at the accident site for approximately two hours. Later he went to the hospital where he learned that he could not speak to the victim because of his injuries. Thereafter he went with Inspector Moreschi to the place where the vehicles had been towed. After inspecting the vehicles, the investigating officers returned to the crash site to make observations and take measurements. These inspections were necessary because earlier observations had been conducted in darkness immediately after the accident. The trooper learned of the victim's death on December 22, and testified that he made the decision to charge the defendant with vehicular homicide, operating to endanger, and

Page 609

speeding, under G. L. c. 90, Sections 17, 24, and 24G, on December 23 or 24, after further discussing the case with Inspector Moreschi.

The defendant testified that he was in the State police cruiser for approximately one and one-half hours and that there had been at least three interruptions in the questioning, the first lasting fifteen to twenty minutes. He testified that the trooper "told me to sit in the car [and] that I was to remain in the car until he told me I could leave the car." He also testified that at one point Trooper Ross left the cruiser, saying, "Stay here, I'll be right back." It is not disputed that Miranda warnings were never given. The defendant testified that he left the scene after Trooper Ross said words to the effect of "I have just about all the answers I need. You may go now."

The judge ruled that the questioning amounted to custodial interrogation, which required Miranda warnings because "the questioning took place in a closed police cruiser at the request of the [t]rooper, the questioning lasted a good bit of time, including periods of renewed questioning during which the [t]rooper left the cruiser and returned leaving the defendant inside," and because "the defendant testified that he felt he could not leave the cruiser." The judge also indicated that once the defendant was identified as the driver of the vehicle that had struck the other vehicle in the rear, he was "the only logical suspect." He ruled that, "[i]n these circumstances, while some preliminary inquiry was appropriately investigatory, once certain key facts were obtained (location of cars, injury, and that the defendant was the operator), the questions about speed and manner of operation passed beyond investigation into evidence-gathering from the only known suspect." He then suppressed "answers to the [t]rooper's questions about speed and about how the accident happened."

In Commonwealth v. Merritt, ante 601 (1982), decided today, we considered the special problems inherent in the application of Miranda to serious motor vehicle accidents. In this case, as in Merritt, initial observations by the investigating

Page 610

officer at the accident site were inconclusive as to whether a crime had been committed. It is immaterial, however, that a determination of criminal conduct had not been made at the time of the questioning if that questioning amounts to "custodial interrogation," sufficient to trigger the applicability of Miranda. It is not necessary to repeat here what we said in general terms in the Merritt decision about the meaning of "custodial interrogation." We turn to the four specific indicia of custody relied upon by the judge in his determination that the interrogation violated Miranda.

1. Location. The judge felt that questioning in a police cruiser at an officer's request was indicative of custody. The use of police premises for questioning, however, does not necessarily imply custodial interrogation. Oregon v. Mathiason, 429 U.S. 492, 495 (1977). The testimony indicates that the questioning took place on a highway in the pre-dawn hours of a bitterly cold winter day. The defendant's vehicle had been demolished. These conditions suggest that any questioning, custodial or not, might reasonably be expected to have taken place in a police vehicle, simply by reason of lack of any alternative shelter, heat, or light. It further suggests that the necessity of using the police vehicle should have been obvious to the defendant, thereby minimizing its subjective effect as an indication that he was in custody. In the absence of any explicit or implicit finding by the judge that a reasonable person would, in the circumstances, have perceived the police cruiser as a tool or token of custody, we are reluctant to assign great weight to this factor.

2. Duration. We are not able to attach much significance to the judge's conclusion that "the questioning lasted a good bit of time, including periods of renewed questioning during which the [t]rooper left the cruiser and returned leaving the defendant inside." It has been held that taking such a "quantitative approach" in analyzing whether an interrogation "was too long to constitute general on-the-scene questioning" ignores "the fundamental inquiry mandated

Page 611

by Miranda -- whether the person is placed in a coercive environment which restricts his freedom so as to render him `in custody,' Oregon v. Mathiason, supra, 429 U.S. at 495, a scenario which can develop in a matter of moments, see Orozco v. Texas, 394 U.S. 324, 325 (1969), or not materialize over a three-hour period of questioning, see Beckwith v. United States, 425 U.S. 341, 342-343 (1976)." Borodine v. Douzanis, 592 F.2d 1201, 1208 (1st Cir. 1979). In the absence of any findings as to how long the defendant was questioned, we have examined the evidence to ascertain whether the passage of time may have had a coercive effect. The defendant testified that he was in the police cruiser for more than one hour, but that the questioning was interrupted at least three times, once for fifteen to twenty minutes. He also testified that he was aware that Trooper Ross had left the cruiser to speak to and direct police personnel who had just arrived at the scene, and that the trooper did not say that he had left to check the veracity of the defendant's statements. It is thus apparent from the defendant's testimony that he was questioned for a fraction of the time he was in the cruiser and that there was no objective indication that the interruptions were related to the questioning. We are unable to discern from these facts that the questioning was so protracted as to have had a coercive effect. We note that the trooper testified that the questioning took about fifteen to twenty minutes and was interrupted at least twice. The evidence also indicates that the defendant spent some time in the cruiser after the questioning had ended while he awaited a ride home. We have no assistance in the form of findings as to the weight which the judge attached to any of the testimony on this issue, but we believe our conclusion necessary even if we limit ourselves to consideration of the defendant's testimony construed most favorably to him.

3. Defendant's subjective impressions. The judge's ruling also relied upon the fact that "the defendant testified that he felt he could not leave the cruiser." Taking this statement as a finding that the defendant actually so felt, we think it irrelevant. The issue of "custody" is not determined

Page 612

by how the defendant felt, but by whether there were objective features of the interrogation which would reasonably lead the defendant to believe that his freedom of action had been curtailed. See Borodine v. Douzanis, 592 F.2d at 1206 n.2; United States v. Pratt, 645 F.2d 89, 91 (1st Cir. 1981). To emphasize a defendant's purely subjective feelings about "custody," without considering the reasonableness of those feelings, would allow issues to turn not on the objective circumstances of the questioning, but on personal idiosyncracies which are neither within the control of, nor necessarily observable by, investigating personnel. The law imposes no burden on the police to divine a suspect's subjective impressions.

4. Focus. The judge concluded that the investigation had focused on the defendant because, having been identified as the driver of a vehicle which had struck the other vehicle from the rear, "the defendant was the only logical suspect." (Emphasis original.) Our review of the evidence indicates clear error in this conclusion. Trooper Ross testified that after the questioning he "was not quite sure whether [the other driver] had done anything to cause Mr. Comolli to have impact with him." He stated that he was interested at the time of the questioning in determining whether the victim might have caused the accident by stopping his vehicle in a traffic lane or by suddenly pulling out in front of the defendant. He further testified that the defendant was not under suspicion at the time, and that he (Ross) wanted to speak to the other operator about the accident. Later, having been informed by that operator's physicians that he was too seriously hurt to answer questions, the trooper and the Registry inspector examined the vehicles and made specific tests for equipment failure. They also returned to the accident scene, to make observations for the first time in daylight, where they looked for skid marks which would have indicated whether evasive action had been taken. The decision to bring charges was not made for some days after the accident. Moreover, that decision was made only after the trooper and the Registry inspector had

Page 613

compared notes and discussed their theories of how the accident might have occurred. Compare Commonwealth v. King, 387 Mass. 464, 474 (1982). This evidence demonstrates that while on-the-scene questioning may have focused on the defendant, simply because he was the only party able to respond, the investigating officers clearly entertained a number of alternative explanations about the cause of the accident, including equipment failure and possible fault on the victim's part. They pursued these theories even after they had finished questioning the defendant and no decision was made to charge the defendant until several days later. Commonwealth v. Podlaski, 377 Mass. 339, 343 (1979). In this context, the judge's conclusion that the investigation (as opposed to the questioning) had focused on the defendant is erroneous, in the absence of a finding that he disbelieved the trooper's uncontradicted testimony.

As previously mentioned, the judge may have disbelieved the trooper's testimony on several critical issues. He may also have credited and given considerable weight to the defendant's testimony to the effect that he was restrained in the police cruiser until the trooper had completed his questioning. The record is capable of supporting a ruling that the defendant was questioned in a custodial environment sufficient to require Miranda warnings, depending on whom and what the judge believed. However, as already indicated, we lack express findings on most of the relevant issues and we do not view the written document filed by the judge as specific enough to support a conclusion that the requisite findings have been made implicitly. We conclude that the judge's reliance on criteria which may have had minimal value in the circumstances (questioning in a police cruiser and the duration of the interrogation), no value at all (the defendant's subjective impression of custody), or which are without sufficient support in the evidence (investigative focus), together with the absence of findings, requires that the case be remanded. The order allowing the motion is vacated and the case remanded so that the judge can make (a) detailed subsidiary and ultimate findings of

Page 614

fact, and (b) relevant rulings of law, on the issues raised by the defendant's motion to suppress. The judge may, in his discretion, take such additional evidence as he deems necessary.

So ordered.