Researchers hooked a flowmeter to pre-rolls at social smoke sessions in five states and found three distinct ways people pull on a joint. All three moved more smoke than the cigarette settings cannabis research has been running for decades.
Researchers found three ways people smoke a joint, and you are almost certainly one of them.
That’s the finding of a new white paper from the Network of Applied Pharmacognosy, which took a flowmeter to social consumption events across California, Massachusetts, New York, Nevada and South Carolina between September 2024 and June 2025, handed people a one-gram pre-roll, and asked them to smoke the way they normally smoke.
Sixty-two people did. Eleven recordings were thrown out for bad data. The remaining 51 sorted, almost embarrassingly cleanly, into three groups.
Which One Are You?
The intense puffer is the majority, 57% of the group. Sharp pull, peak flow almost immediately, done in a couple of seconds.
The slow puffer is the smallest group, 20%. Long, steady draw that takes nearly twice as long and never spikes.
The double puffer, 23%, takes two hits back to back, one or two seconds apart, as a single event. This is the one you can spot across a circle.
Riley Kirk, a co-author on the paper, put the difference between the first two in terms anyone who has ordered a drink can follow.
“It’s really a difference between a fast intense pull, almost like you’re sipping boba from a straw, versus a long, slow pull that’s more like sipping a thick milkshake.”
Riley Kirk, Network of Applied Pharmacognosy
And here is the counterintuitive part. The gentle one moves more smoke.
“The intense pull hits its peak strength almost immediately, under a second in, and the whole thing is over in about 2 seconds,” Kirk told High Times. “The slow pull takes almost twice as long to build up, and never hits as high a peak, but it holds there instead of spiking. Interestingly, that’s what actually pulls in more total smoke per puff than the intense style. The slower pull feels gentler, but the extra time more than makes up for the lower intensity.”
The numbers back it up. The intense profile peaks at 41.7 mL/s and averages 59.3 mL of smoke per puff. The slow profile peaks at barely half that, 22.4 mL/s, and still averages 72.5 mL. The double puff, counting both hits, comes in highest at 76.8 mL.
Science Has Been Studying You as a Cigarette Smoker
Here is why anyone outside a smoke circle should care. When a lab wants to know what is in cannabis smoke, it doesn’t ask a person to smoke. It programs a smoking machine. And for decades those machines have been running puff settings borrowed from tobacco research: roughly a two-second puff pulling 35 to 55 mL, the ISO and Health Canada standards built to model a cigarette smoker.
Every one of the three cannabis profiles is bigger than that.
| Profile | Total puff volume | Puff length | Peak flow |
|---|---|---|---|
| Health Canada (cigarette) | 55 mL | 2.0 s | 27.5 mL/s |
| Cannabis: intense | 59.3 mL | 2.3 s | 41.7 mL/s |
| Cannabis: slow | 72.5 mL | 4.7 s | 22.4 mL/s |
| Cannabis: double | 76.8 mL | 3.3 s | 40.7 and 41.8 mL/s |
“Cannabis smoke research has been borrowing cigarette puff settings for decades, and the issue is that cannabis consumers just don’t smoke like cigarette smokers do,” Kirk said. “Cannabis consumers typically inhale far more deeply and hold their breath much longer than cigarette smokers, which produces very different puff volumes.”
She added that some smokers in the dataset had puff volumes well over 100 mL, “so we’re likely underestimating smoke and compound delivery just by using the cigarette numbers as a stand-in.”
What that means in practice is that the existing body of lab work on what comes off a burning joint has largely been built around cigarette puff settings that don’t fully reflect how cannabis consumers actually smoke.
What the Chemistry Might Do
This paper measures behavior, not smoke. It does not tell you whether one style delivers more THC, more terpenes or more of the things nobody wants in their lungs. But it sets up the question, and Kirk has a hypothesis.
“Our working hypothesis is that puff length and flow rate change the temperature of the ember itself, and that a longer, slower burn could produce meaningfully different chemistry than a short, hot pull, resulting in more or less of certain cannabinoids, terpenes, or combustion byproducts,” she said. “We haven’t measured that chemistry yet, though. But that’s exactly what this data sets up, and it’s the focus of the next paper.”
That next paper will run the three profiles on smoking machines and analyze what each one produces. “This paper showed us there are clear, distinct ways people smoke cannabis,” Kirk said. “The next one will tell us what that means for the consumer, whether one profile delivers more THC, more terpenes, or more of the byproducts we’d rather people weren’t inhaling.”
Real People, Real Sessions
The thing that surprised Kirk was not the shapes. It was how tidy they were, given where the data came from.
“This wasn’t a controlled lab session. We collected data at real social gatherings across five states, with different products, papers, and filters, all factors we know introduce variability. Given all that, I expected the puff shapes to be all over the place. Instead, three clear patterns emerged anyway.”
Riley Kirk
The authors are upfront about what that setting costs them. Sixty-two people at parties is not a randomized sample, and the profiles reflect who happened to be in those rooms. Paper, filter, grind, moisture and age of the pre-roll all vary and all affect the draw. Each recording covers at least three minutes, not a whole joint, so it misses how the pull changes as the coal settles, the joint shortens and resin builds. And the device itself, with its tubing and altered mouthpiece, may change how somebody smokes.
They also note something that no cigarette protocol has ever had to account for: people pass joints around, while cigarette smokers generally smoke their own.
What the paper claims for itself is narrow and worth stating plainly. It is the first documented set of cannabis-specific puff profiles, and the first set of parameters a lab could program into a smoking machine instead of the cigarette defaults. Whether that changes what we know about cannabis smoke is a question for the next study.
In the meantime, the next time somebody in the circle takes two quick hits before passing, you know what to call them.
Disclosure: “From User Behavior to Laboratory Standards: Developing Cannabis Smoking Machine Protocols from Real-World Puff Topography Data” is an open-access white paper published by the Network of Applied Pharmacognosy in June 2026. It has not been peer reviewed. The measurements were taken with the CannaFlo, a flowmeter made by Cambustion Ltd., and two Cambustion employees are co-authors. The authors say the white paper itself received no direct funding, and that the broader Science of Smokability research program is supported by industry sponsors including Abstrax, Boveda, Cannatrol, Curaleaf, Custom Cones, Day Savers, the Ganjier Group, the Hemp Mine and Plant Life Co.
