HPTLC chromatogram of Kali-ISA, Kali-Rogu and Kali-Hiwa kava roots under derivatized white light, showing the characteristic kavalactone band at Rf 0.4
american kava hplc hptlc kava science kavalytics nirs research white paper
2 min read

Three Ways of Asking the Same Question: Is American-Grown Kava Really Kava?

American kava is young enough that scepticism is fair. A grower can call a plant Borogu, but a name travels on a cutting, not on a certificate, and cuttings get mislabelled, mixed and renamed across decades of hobbyist exchange. So for American Kava Association Technical White Paper No. 2 we did not ask anyone to take the names on trust.

Three American-grown cultivars, Kali-ISA, Kali-Rogu and Kali-Hiwa, grown at the AKA Research Greenhouse in Sacramento, were run through three independent identity checks. Flora Research Laboratories of Grants Pass, Oregon performed HPTLC fingerprinting against two authenticated Piper methysticum reference materials and HPLC-UV quantification of all six major kavalactones and the flavokavains. Separately, our in-house KavaLytics programme classified each lot from its near-infrared spectrum. All three methods agreed.

The material assayed between 7.09% and 14.29% total kavalactones, with flavokavains below 0.4% and at 1.6–2.5% of total kavalactones across every lot,  the signature of beverage-grade kava rather than tudei-type material. Kali-Hiwa (chemotype 463251) and Kali-Rogu (423561) classify as noble; Kali-ISA reads 254631, an ISA-type chemotype. What follows is a walkthrough of each method, the lab numbers, the TLC plates, and why a 19.9% figure out of a Florida field is not the anomaly it might be viewed as.

American Kava Association · From Technical White Paper No. 2

Three ways of asking the same question: is American-grown kava really kava?

Before you can say anything useful about how strong a kava is, you have to prove what it is. The three American-grown cultivars behind White Paper No. 2 were put through three independent identity checks — HPTLC fingerprinting, HPLC-UV quantification and near-infrared classification — and all three agreed. Every lot is authentic Piper methysticum, and every lot is beverage-grade, with flavokavain content between 1.6% and 2.5% of total kavalactones.

Analytical work by Flora Research Laboratories, Grants Pass, Oregon (job J25-0930-H, 9 October 2025). Material grown at the AKA Research Greenhouse in Sacramento, California, and served fresh at Root of Happiness kava bars.

  • HPTLC identity
  • HPLC-UV potency
  • NIRS classification
  • 3 American cultivars
Published by
American Kava Association · Las Vegas, NV
Source
Technical White Paper No. 2 — Where the Kavalactones Go
Authors
Tyler Blythe · Jeffrey Bowman · Matthew Masifilo
Analytical laboratory
Flora Research Laboratories, LLC — Grants Pass, Oregon (job J25-0930-H)
Analysis date
9 October 2025
Cultivars
Kali-ISA · Kali-Rogu · Kali-Hiwa
Grown at
AKA Research Greenhouse, Sacramento, California
Identity methods
HPTLC · HPLC-UV · NIRS (KavaLytics)
Drink it fresh at our kava barsJump to the lab numbers
HPTLC chromatogram of Kali-ISA, Kali-Rogu and Kali-Hiwa kava roots under derivatized white light, showing the characteristic kavalactone band at Rf 0.4

All three cultivars, run on a single HPTLC plate alongside two authenticated Piper methysticum reference materials and a kavain standard. The dark band at Rf ~0.4 is the principal kavalactone zone.

The premise

Naming a cultivar is not the same as proving one

American kava is young enough that scepticism is fair. A grower can call a plant Borogu, but a name travels on a cutting, not on a certificate — and cuttings get mislabelled, mixed and renamed across decades of hobbyist exchange. So for White Paper No. 2 we did not ask anyone to take the names on trust.

Three separate analytical methods were used, each answering a different question:

  1. HPTLC — is it Piper methysticum at all?

    High-performance thin-layer chromatography compares the whole chemical fingerprint of a sample against authenticated reference root. It is the identity test written into the British Herbal Pharmacopoeia, and it is pass/fail: either the band pattern matches the reference material or it does not.

  2. HPLC-UV — what exactly is in it, and how much?

    High-performance liquid chromatography with UV detection puts a number on each of the six major kavalactones and on the flavokavains. That gives you the chemotype code, the total potency, and the flavokavain ratio that separates beverage-grade kava from tudei-type material.

  3. NIRS — can it be verified again, cheaply, at scale?

    Near-infrared spectroscopy paired with machine learning classifies a sample in seconds without solvents or a chromatography bench. It cannot replace a reference-lab assay, but it can be run on every incoming lot, which a reference-lab assay cannot.

What we are and are not claiming

HPTLC and HPLC-UV were performed by an independent accredited laboratory as part of White Paper No. 2. The NIRS work described further down is our own in-house KavaLytics programme and is not part of that paper. We have kept the two clearly separated throughout.

Method one · HPTLC

The fingerprint test: all three cultivars matched authenticated reference root

Flora Research Laboratories ran the British Herbal Pharmacopoeia TLC-ID-Kava identity test, modified so that all three cultivars sat on one silica plate with shared reference and standard lanes — which means the comparison is direct, not inferred across separate runs. Extraction was by chloroform sonication; application was by CAMAG ATS4 autosampler at 1 µL and 2 µL per cultivar.

Lane assignments — FRL job J25-0930-H

All three American-grown lots and both authenticated reference materials were developed on the same plate.

LanesSampleApplication
5–6Kali-ISA (LOT #KI-R)1 µL, 2 µL
7–8Kali-Rogu (LOT #KR-R)1 µL, 2 µL
11–12Kali-Hiwa (LOT #KH-R)1 µL, 2 µL
13Authenticated Piper methysticum root, FRL 10-032-0222 µL
14Authenticated Piper methysticum root, FRL 22-62-2752 µL
15Kavain reference standard, FRL 29-94-0012 µL

Analysis performed at Flora Research Laboratories, Grants Pass, Oregon; 9 October 2025.

HPTLC chromatogram under derivatized white light after vanillin-sulfuric acid treatment, showing an intense band at Rf 0.4 in the Kali-ISA, Kali-Rogu and Kali-Hiwa lanes matching the reference lanes

Plate under white light, after vanillin–sulfuric acid derivatisation. All three cultivars produced the characteristic intense band at Rf ~0.4 corresponding to the principal kavalactone zone of the reference materials, with co-migration of the kavain standard in lane 15.

The same HPTLC plate imaged under UV 366 nm, showing blue-fluorescing bands at Rf 0.4 to 0.5 and red-fluorescing bands higher up the plate

The same plate under UV 366 nm. All three cultivars show the diagnostic blue-fluorescing bands in the Rf 0.4–0.5 region and red-fluorescing bands in the upper Rf range, matching the fluorescence pattern of the authenticated reference materials.

Two illuminations, one plate, two independent reference lots, and a pure kavain standard for good measure. The band pattern is not similar to authenticated Piper methysticum — it is the same pattern.

Method two · HPLC-UV

The numbers: potency, chemotype and flavokavain content

Quantification was by HPLC-UV on an Agilent 1200 RRLC with a Poroshell 120 SB-C18 column at 55 °C, eight-point calibration from 5 to 250 µg/mL with R² ≥ 0.999, detection at 240 nm for methysticin, dihydromethysticin, kavain and dihydrokavain and at 355 nm for yangonin and desmethoxyyangonin. Chemotype codes read in the conventional order 1 = desmethoxyyangonin, 2 = dihydrokavain, 3 = yangonin, 4 = kavain, 5 = dihydromethysticin, 6 = methysticin.

  • 14.29%total kavalactones in the Kali-ISA whole-root lot

    At the very top of the range reported for noble Fijian cultivars, and grown in a Sacramento greenhouse.

  • 1.6–2.5%flavokavains as a share of total kavalactones

    Low across every lot — the signature of beverage-grade material rather than tudei-type kava.

  • 2 of 3lots classified noble by chemotype

    Kali-Hiwa (463251) and Kali-Rogu (423561). Kali-ISA reads 254631, an Isa-type chemotype.

  • 188–452 mgkavalactones delivered per 125 mL serving

    Measured in the finished beverages prepared from these lots by traditional squeeze-and-strain.

Raw root composition by lot

Table 1 — Kavalactone and flavokavain content of the raw material

All values determined by HPLC-UV at Flora Research Laboratories, job J25-0930-H, 9 October 2025.

LotTotal kavalactonesTotal flavokavainsFK as % of KLChemotypeClassification
Kali-ISA (whole root)14.29%0.354%2.5%254631Isa-type
Kali-Hiwa (whole root)7.09%0.133%1.9%463251Noble
Kali-Hiwa (laterals only)10.64%0.234%2.2%463251Noble
Kali-Rogu (whole root)9.98%0.162%1.6%423561Noble

Percentages are of dry root mass. Chemotype digits: 1 = desmethoxyyangonin, 2 = dihydrokavain, 3 = yangonin, 4 = kavain, 5 = dihydromethysticin, 6 = methysticin, read in descending order of abundance.

What the chemotype codes actually say

Kali-Rogu reads 423561 — kavain first, then dihydrokavain, yangonin, dihydromethysticin, methysticin and desmethoxyyangonin. That is the canonical Borogu profile described by Lasme and colleagues in 2008 and by Teschke, Sarris and Lebot in 2011. The plant is doing in Sacramento what its namesake does in Vanuatu.

Kali-Hiwa reads 463251, a kavain-led noble profile, in both the whole-root and laterals-only preparations. Note that the laterals alone came in at 10.64% total kavalactones against 7.09% for the whole-root mix — the roots concentrate the lactones, exactly as they should.

Kali-ISA reads 254631, led by dihydrokavain and dihydromethysticin. That is an Isa-type chemotype, not a noble one, and we are not going to call it noble because it isn't. What it is is beverage-grade: its flavokavain load sits at 2.5% of total kavalactones, nowhere near the flavokavain-heavy signature of tudei material.

Beverage-grade is the claim; noble is a subset of it

Low flavokavain content demonstrates that a lot is beverage-grade rather than tudei-type. It does not by itself establish noble status — that comes from the chemotype code. Two of these three lots are noble by chemotype; all three are beverage-grade by flavokavain content. We keep those two statements separate on purpose.

How this compares internationally

A 2025 survey of 77 noble Fijian cultivars from Taveuni, Qamea and Rabi reported total kavalactones between 7.2 and 14.1 wt.%, with a mean of 10.1 ± 1.8 and a kavain-dominant root code of 4[632][51]. Every lot in this study falls inside that window, and the Kali-ISA lot at 14.29% sits marginally above its top end.

Whatever else is true about American kava, potency is not the weak point.

Reference range from Devi & Pasinszki, Agriculture 2025, 15(5), 478 — doi.org/10.3390/agriculture15050478. Chromatographic method adapted from Liu et al. 2018.

Method three · NIRS

KavaLytics: near-infrared spectra plus machine learning, running since 2017

The third check is ours. KavaLytics is an in-house programme we began in 2017 to classify kava lots from near-infrared spectra using machine-learning models, so that identity and grade can be screened on every incoming lot rather than on the handful you can afford to send to a reference lab. Against our internal validation set it reaches a 94% correct-classification rate. Each of the three cultivars in this study read beverage-grade.

Scope note

The KavaLytics NIRS results are internal quality-control work and are not part of Technical White Paper No. 2, which reports the HPTLC and HPLC-UV analyses only. The 94% figure comes from our own validation and has not been peer reviewed. We report it here because it is the third leg of how we actually verify these cultivars in practice — not as a published finding.

An independent team has now built something very similar

In October 2025 a team led by Ronick Spenly Shadrack at the Vanuatu Bureau of Standards Laboratory, with the eminent kava scientist Vincent Lebot as senior author, posted a preprint applying an artificial neural network to infrared spectra of kava for exactly this purpose: telling cultivars and geographic origins apart, and separating noble from tudei.

Their method is not identical to ours. They use ATR-FTIR — mid-infrared spectroscopy on acetone extracts — where KavaLytics works in the near-infrared. But the shape of the idea is the same: capture a vibrational spectrum, let a neural network learn the class boundaries, and get an answer in seconds instead of hours.

What the Vanuatu Bureau of Standards / Lebot preprint reports

Selected results from the v2 preprint, posted 14 October 2025.

TaskReported performance
Noble vs tudei discrimination100% class accuracy
Geographic origin — Vanuatu, PNG, Hawaii100% class accuracy
Geographic origin — Fiji33.3% (limited samples; no blind test)
Malo / Santo micro-region100% class accuracy
Adulteration detectionVisible at 1% tudei or wichmannii substitution (1585 cm⁻¹)
Model fit (geographic origin)R² 0.99 train / 0.84 validation / 0.95 test

The authors state plainly that, due to limited independent Fiji samples, no blind-test validation was performed for that region, and that robustness claims apply only to regions with sufficient validation data. We think that is the right way to report a model.

Two teams on opposite sides of the Pacific, working independently, arrived at the same conclusion: an infrared spectrum plus a trained model is enough to tell one kava from another.

The useful takeaway

We are not claiming they replicated us, and we would not want anyone to read it that way. We started in 2017 and they published in 2025; the two efforts developed independently, in different spectral regions, on different sample preparations. What we are saying is that convergent development is the strongest kind of validation a method can get. When a team of that calibre — a national standards laboratory, with Lebot's name on the paper — builds the same tool and finds it works, it tells us the approach is sound. We are genuinely glad to see it.

It also matters commercially. Spectroscopic screening is how kava authentication scales beyond the small number of lots anyone can afford to send for chromatography, and the more independent groups validating it, the sooner it becomes normal practice across the industry.

Shadrack, R. S.; Daniel, T.; Tabi, H.; Botleng, J.; Kelep, R.; Regenvanu, L.; Pakoasongi, M.; Butjukabwaelep, E.; Siro, G.; Pipite, A.; Lebot, V. Geographical Origin and Cultivar Differentiation of Kava (Piper methysticum) using Artificial Neural Network with FTIR Spectroscopy: A Novel Method. bioRxiv preprint, v2 posted 14 October 2025 — doi.org/10.1101/2025.10.05.679113. Preprint; not yet peer reviewed. Related earlier work: Lebot et al., Food Control 2024 — doi.org/10.1016/j.foodcont.2024.110598.

Greenhouse to field

Why 19.9% kava out of Florida is not an anomaly

Everything above was grown under glass. The three-year plants behind White Paper No. 2 came out of the AKA Research Greenhouse in Sacramento — a controlled environment, good for a study, but not the environment American kava will actually be produced in. That is happening in Florida.

Matt Masifilo standing behind a field-grown kava plant at his Florida grow

Matt Masifilo — founder of Kavafied, inventor of the AluBall and a co-author of White Paper No. 2 — with home-grown kava at his Florida planting.

Matt has grown kava in ground in Florida and his material assaying at 19.9% total kavalactones. That number gets treated as suspicious, and we understand why as it sits well above the 7.2–14.1% band reported for noble Fijian cultivars. But look at what the whole plant greenhouse material did.

The arithmetic that makes it plausible

The Kali-ISA lot that reached 14.29% was a whole-plant mix — not laterals only. Lateral roots concentrate kavalactones substantially above a whole-plant average: in the Kali-Hiwa lot on the same plate, laterals came in at 10.64% against 7.09% for the whole-root mix, a lift of roughly 50%. If a whole-plant Florida lot of comparable genetics were selected down to laterals, a figure approaching 19.9% is not an outlier. It is what the same ratio predicts.

To be explicit about what this is: an argument from plausibility, not a finding of this study. We did not assay Matt's Florida material as part of White Paper No. 2, and nothing here should be read as independent verification of the 19.9% figure. What the greenhouse data does establish is that the number is well inside the range American genetics can reach, and that dismissing it out of hand is no longer a reasonable default.

The growing conditions are worth stating plainly, because they cut against nearly every assumption about what kava requires. Matt did not plant into prepared beds of rich topsoil. He planted into construction debris — broken concrete, aggregate and rubble — with mulch laid over the top. That is what produced the material in question. The same planting also came through 28 °F frosts in Florida, well below the temperature at which kava is usually assumed to give up.

Rubble below, organic litter above. That construction is stranger to modern horticulture than it is to the historical record. Samuel Kamakau’s account of how Hawaiians planted ‘awa describes very nearly the same thing:

To plant awa, the planter first went to fetch stems (hakai) and broke off (ha’iha’i) a quantity of them. He carried them on his shoulder to a suitable place, where he broke them into sections (poke), being careful not to break off the nodes on the joints, and laid them in a trench in a compost of muck and trash (ho’owa’a a kipulu i ka lepo ’opulupulu a me ka ’opala). He cut a lot of greenery (pulu) for mulch, and left these grasses until they had dried, then mixed them with soil. Then he fetched the stems which had been laid in the trench. Some of them had sprouted ‘ears’ (pepeiao; stipules), some had sprouted leaves, and some, fine roots. The farmer took them and planted them in the mulch, along with some plants that had been ground-layered before (kakiwi mua ’ia).

Samuel M. Kamakau, The Works of the People of Old (Na Hana a ka Po’e Kahiko), Bishop Museum Press, 1976

A trench of muck and trash, dried greenery mixed back into the soil, and the sprouted stems set into the mulch rather than into fine tilth. Kava has been carried and planted by people for millennia, and for most of that history it went into disturbed ground and organic litter, not into a nursery bed. What Matt did in Florida is not a horticultural stunt that happened to work. It is much closer to the original method, built out of what a construction site leaves behind.

Florida field-grown kava is formidably strong, and it is now in production on multiple farms. We cannot wait to be able to offer it.

Where to drink it

This is the same fresh material we serve

The cultivars in this study are not a research curiosity sitting in a freezer. Fresh kava from these plants is served regularly at Root of Happiness kava bars — prepared traditionally, by hand, from the same American-grown material that went through the plate and the column.

Interior of a Root of Happiness kava bar with a large wooden tanoa bowl of kava, coconut shells and communal tables

Fresh American-grown kava, prepared and served the traditional way.

Find a Root of Happiness kava bar

Can’t make it to a kava bar? Start here.

Our American-grown kava does not leave our six kava bar locations. It became a fan favorite almost immediately and we can barely keep up with the demand in-house, so there is simply none left to box up and ship.

What we can send you is our Fresh Frozen Fijian, and it is genuinely on par with the American-grown material. It goes through the same rigorous cleaning and processing — washed, trimmed, hand-sorted and kept in cold chain from the ground to your freezer — and it is held to the same standard for what makes it into a box. It is an excellent choice, and it is what we reach for ourselves when an American lot runs dry.

Credits and citation

Sources, methods and disclosures

This article draws on American Kava Association Technical White Paper No. 2. The HPTLC and HPLC-UV data reported here are from that paper; the NIRS classification work is separate in-house KavaLytics work and is identified as such above.

Suggested citation. Blythe, T.; Bowman, J.; Masifilo, M. Where the Kavalactones Go: a quantitative mass-balance analysis of traditionally prepared aqueous kava beverages. American Kava Association Technical White Paper No. 2; American Kava Association: Las Vegas, NV, USA, 2026.

Analytical work. Flora Research Laboratories, LLC, Grants Pass, Oregon (job J25-0930-H, 9 October 2025). Identity by HPTLC following the British Herbal Pharmacopoeia TLC-ID-Kava test, modified. Quantification by HPLC-UV, eight-point calibration 5–250 µg/mL, R² ≥ 0.999.

Cultivation. Three-year plants grown at the AKA Research Greenhouse, Sacramento, California.

Disclosures. The authors are commercially engaged in the kava trade. Tyler Blythe is affiliated with Root of Happiness and Kali-Kava, which supplied the material; Matthew Masifilo is affiliated with Kavafied; Jeffrey Bowman is affiliated with Nakava. KavaLytics is an American Kava Association programme. The independent laboratory generated the analytical data; the authors performed the interpretation.

References. Devi & Pasinszki, Agriculture 2025, 15(5), 478. Lasme et al. 2008. Teschke, Sarris & Lebot 2011. Liu et al. 2018. Shadrack et al. (incl. Lebot), bioRxiv 2025.10.05.679113 (preprint).