A Butler County vehicular homicide case decided today raises a question that comes up more often than you might think: when an eyewitness to a fatal crash takes the stand, can they describe what they saw — or are they limited to rattling off a miles-per-hour estimate like a speed gun?
The Pennsylvania Superior Court’s answer in Commonwealth v. Rivera, 2026 PA Super 71 (Apr. 10, 2026), is that it depends on what the prosecution is trying to prove. And in a case where the Commonwealth is charged with proving malice — the mental state required for third-degree murder — eyewitnesses get considerably more room to describe what they saw.
The Facts of the Case
On the afternoon of October 19, 2023, Joshua Rivera was driving a white Kia sedan on Dutchtown Road in Butler County when he struck a fifteen-year-old boy riding an electric bike. An eyewitness who was walking with his family at the time described the car as coming “out of nowhere” and moving “like a missile.” He estimated the vehicle’s speed at around 80 miles per hour on a road with a 35 mile-per-hour speed limit. The victim was thrown into the air and propelled roughly 75 to 80 feet from the point of impact.
Rivera got out of his car, said “Oh my God, I’m so sorry,” asked the witness to call 911 — and then got back in his car and drove away “at a high rate of speed.” The car was later found abandoned a few miles from the scene. Rivera was arrested three days later.
What the Commonwealth also found: a partially-burned marijuana cigarette lying on the ground near where Rivera’s car had stopped. A search warrant executed on the impounded vehicle three weeks later turned up a baggie of suspected marijuana behind the driver’s seat, rolling papers and unidentified pills in the glove box and center console, and an unopened alcoholic beverage in the trunk. A DNA test on the marijuana cigarette matched Rivera.
The Commonwealth charged Rivera with a range of crimes, then later sought to amend the information to add a count of third-degree murder — a charge requiring proof of malice, which under Pennsylvania law means a conscious disregard of an unjustifiably high risk of death or serious bodily harm. The trial court allowed the amendment. Then, on pretrial motions to exclude evidence, things got complicated.
What the Trial Court Did — and What the Superior Court Said About It
Rivera filed motions in limine — pretrial requests to exclude certain evidence from trial. (A motion in limine, for non-lawyers, is essentially a request to settle evidentiary disputes before the jury is seated, rather than interrupting the trial with objections.) Rivera targeted four categories of evidence: (1) eyewitness testimony describing the car’s speed and manner of travel, (2) the drugs and drug paraphernalia found in the vehicle, (3) the audio from police body camera footage captured at the scene, and (4) a Facebook video depicting someone driving at 130 miles per hour.
The trial court granted some of Rivera’s requests and denied others. The Commonwealth appealed what was excluded. The Superior Court affirmed some of the trial court’s rulings and reversed others.
The Eyewitness Testimony Issue
The most significant ruling — and the one with the broadest implications for practitioners — involves what eyewitnesses can say about how a vehicle was being driven.
The trial court relied on older Pennsylvania Supreme Court authority — specifically Catina v. Maree, 447 A.2d 228 (Pa. 1982), and Starner v. Wirth, 269 A.2d 674 (Pa. 1970) — for the proposition that lay witnesses may estimate vehicle speed only in numerical terms. Words like “fast,” “excessive,” or “like a missile” were out. A numeric estimate — “80 miles per hour” — was in, so long as the witness had an adequate opportunity to observe the vehicle.
That meant the eyewitness who described Rivera’s car as feeling “like a missile was coming” couldn’t say that. The witness who described Rivera fleeing “like a bat out of hell” couldn’t say that either. They could give a number, and nothing else.
The Superior Court reversed that ruling.
The key distinction the court drew is between cases where speed is the issue and cases where the prosecution also has to prove state of mind. The court relied primarily on its own 1974 decision in Commonwealth v. Honeycutt, 323 A.2d 775 (Pa. Super. 1974), and its more recent en banc decision in Commonwealth v. Akhmedov, 216 A.3d 307 (Pa. Super. 2019).
Honeycutt involved an involuntary manslaughter case arising from a road racing accident. The court allowed a witness to describe the defendant’s dangerous driving in the minutes before the fatal crash — not just his speed, but the manner in which he was driving — because the Commonwealth needed to prove recklessness, not just velocity. As the Honeycuttcourt put it, when a state of mind is at issue, “a great deal of latitude is appropriate when circumstantial evidence is offered to show a state of mind.” Akhmedov reaffirmed that principle in the context of a third-degree murder charge.
Here, the Rivera court applied the same reasoning: because the Commonwealth was required to prove malice — something considerably more than ordinary recklessness — eyewitnesses couldn’t be stripped down to sterile numerical outputs. They needed to be able to describe what they actually experienced. The Catina and Starner cases that the trial court relied on, the Superior Court noted, were civil cases in which no party was required to prove a mental state. That distinction matters.
The court was careful about limits, though. It upheld the exclusion of testimony from one witness — Mr. Pritschman — who had only a partial glimpse of the vehicle and testified that he “heard more than saw” the car. Because his observation was too brief and limited to provide a reliable foundation, even a numerical estimate from him would have been too speculative. The latitude afforded for state-of-mind evidence doesn’t mean every witness with a theory gets to testify.
The practical holding: in a criminal case where the prosecution must prove a state of mind — recklessness, malice, willfulness — eyewitnesses are not confined to numerical speed estimates. With proper foundation (personal experience with vehicles, opportunity to observe), they can describe the manner and character of the driving they witnessed.
The Drug Evidence
The trial court excluded the marijuana and paraphernalia found during the search of Rivera’s impounded vehicle. Its reasoning: the items weren’t within Rivera’s “reach” while driving, and the three-week gap between the incident and the warrant execution made the evidence too speculative to be relevant.
The Superior Court reversed.
The “within reach” analysis, the court explained, is a doctrine that applies to warrantless vehicle searches — it’s used to define how far police can search without a warrant when they stop someone on the road. It has nothing to do with the relevance of evidence found pursuant to a valid search warrant during an investigation. The trial court imported a doctrine from the wrong context.
As for the time gap: the vehicle was impounded immediately after the incident and its contents weren’t disturbed until the warrant was executed. There was no break in the chain of custody. A three-week gap during which the evidence sat untouched in an impounded vehicle isn’t a relevance gap — it’s just how long it took to get the warrant.
The court’s conclusion was that the marijuana and paraphernalia, combined with the other evidence of potential drug use (the joint at the scene, witness testimony about the smell of marijuana emanating from Rivera), was relevant to establishing Rivera’s state of mind at the time of the crash. Given the third-degree murder charge and the need to prove malice, excluding this evidence was an abuse of discretion.
The Body Camera Audio
Police body camera footage captured officers at the scene discussing the smell of marijuana — including a reference to “Cheech and Chong.” The trial court excluded the audio portion of the body camera video while permitting the video to be played.
The Superior Court affirmed. The audio was hearsay — out-of-court statements offered for their truth — and the Commonwealth never identified an applicable exception. Beyond that, the court agreed that a running commentary from officers at the scene, without the ability to cross-examine those officers as witnesses, would be unfairly prejudicial. The officers can testify at trial about what they smelled; the recorded narration is a different matter.
The Facebook Video
The most colorful piece of evidence was a 30-second Facebook video Rivera had posted prior to the incident, depicting someone in a car with a speedometer reading 130 miles per hour, and a voice saying, “This is how you [] drive, Banner! Like that, bitch!” The Commonwealth argued this showed Rivera’s propensity to drive recklessly, his consciousness of guilt (the video was deleted from Facebook after the Commonwealth gave notice of its intent to use it), and lack of accident.
The trial court excluded it. The Superior Court affirmed.
The problem was an authentication and specificity gap. Under Pennsylvania Rule of Evidence 404(b), prior bad act evidence is admissible for non-propensity purposes — like proving absence of accident or establishing a plan — but only if the prior act has a “close factual nexus” to the crime charged. Here, the video was undated, involved a different vehicle than the one used in the crash, and never showed the driver’s face. The Commonwealth couldn’t establish when it was recorded.
The court’s comparison to Commonwealth v. Akhmedov is instructive. In Akhmedov, a Facebook video was admitted showing the defendant drag racing a silver Audi — the same make and color as the car he used in the crash. The video was tied to a specific street in Northeast Philadelphia, and the defendant had commented on the video bragging about his speed. That’s a close factual nexus. Here, there was no identified vehicle, no identified location, no date, and no confirmed identification of the driver beyond claimed voice recognition.
What This Means for Practitioners
The core holding on eyewitness testimony is the piece worth paying attention to.
Pennsylvania courts had been applying a strict numerical-only rule for lay witness speed estimates derived from older civil cases — and several trial courts have relied on Catina and Starner to cut off descriptive testimony in criminal vehicular cases. Rivera puts a limit on that practice. When the charge requires proof of a mental state — and that’s true of third-degree murder, aggravated assault, and even some grading enhancements — the prosecution can offer testimony that goes beyond “how fast” to “what kind of driving was this.”
That’s a meaningful shift, and defense lawyers should understand it. The expanded latitude isn’t unlimited: the foundation requirements still apply, and witnesses who had only a fleeting observation don’t get to speculate regardless of how colorful their description might be. But a witness who had a clear view and adequate time to observe can now describe what they saw in terms that actually communicate the danger.
For the defense, the Facebook video ruling offers a useful counterbalance. The court’s rigorous application of the “close factual nexus” requirement under Rule 404(b) means that prior bad acts evidence doesn’t get in just because it involves a car and a high speed. The Commonwealth has to be able to connect the prior act to the specific conduct charged — same vehicle type, same location, similar circumstances, a defined timeframe. General “he likes to drive fast” evidence, even if labeled as “lack of accident,” won’t get through.
The drug evidence ruling is worth noting as a trial preparation point: where evidence is gathered through a valid warrant from a secured location, the gap between the incident and the execution of the warrant is not itself a relevance problem. The chain of custody is what matters.