From Hawai'i to the Mainland: What Tissue Culture Did Not Change in Kali Hiwa

A cultivar that cannot be bred, only kept

Kava sets no seed. Piper methysticum is a sterile cultigen propagated only by cutting, so a named cultivar is in principle a clone, and its secondary chemistry should be stable wherever it is grown. That is the assumption the whole trade rests on, and it is rarely tested. What follows is a test of it on one line: Kali Hiwa, the tissue-cultured descendant of Hawaiian Hiwa stock, against the Hawaiian material it came from, across ten laboratory assays spanning seven years.

The practical expression of that stability is the chemotype code of Lebot and Lévesque (1989) — the six major kavalactones ranked by concentration and written as a six-digit string. Two things complicate its use. Ranking is a coarse summary of a continuous composition, and adjacent analytes are often separated by margins narrower than ordinary analytical variation, so a code can change without the plant changing. And a code compresses six numbers into an order, discarding the magnitudes that tell you whether a difference matters.

That matters for American production specifically. Material entering the mainland is routinely micropropagated, both to multiply limited founder stock and to clear the pathogens that travel with field-collected cuttings. Tissue culture, phytosanitary treatment, a new climate and a new soil are four opportunities for a line to drift away from what it was. This paper asks whether the Kali Hiwa line drifted, and answers with the certificates.

Where the Hawaiian stock came from, and why Hiwa

Lebot, Merlin and Lindstrom (1992) proposed that every kava cultivar in the Pacific traces to a single ancestral plant in northern Vanuatu, most probably Piper methysticum var. wichmannii, cloned and carried outward. Because the plant cannot reproduce sexually, Hawaiian planters could not breed new varieties. They could only notice the somatic mutations an existing clone threw off and propagate those separately — selection by observation, over centuries, with no crossing.

Hiwa is one of the cultivars that came through, and the Hawaiian descriptions agree on the trait that names it: hiwa is the dark one. It is black-stemmed, with long internodes, few lenticels and an erect habit reaching eight feet or more in full sun.

It was not selected at random from the thirteen cultivars still identifiable at the end of the twentieth century. In old Hawai'i it was the 'awa of the gods, the priesthood and the chiefs — Handy (1940) recorded it as the darkest of all varieties and the ritual 'awa of the kahuna, possibly reserved for that use, with all the other varieties being common 'awa. That cultural precedence has a chemical correlate, and it is the one measured here: the published reference assay for the cultivar, air-dried lateral roots of a 5.5-year-old Hiwa plant from the Hāmākua coast analysed for Lebot et al. (1999), returned chemotype 462351 at 11.23 percent total kavalactones.

The preservation argument follows from the same biology. A sterile cultigen has no seed bank. Each named cultivar exists only as a clone, and a clone that is lost cannot be bred back.

The rank head is conserved

Kavain ranked first and methysticin second in all ten root, powder and lateral-root samples of Hiwa in the record — Hawaiian and Kali alike. That holds across two islands, two lineages separated by micropropagation and decontamination, four assay dates spanning seven years, and three material classes.

Table 1. Relative kavalactone composition, chemotype code and rank head for every Hiwa certificate in the study. Analyte values are percent of that sample's total kavalactones; Total KL is percent w/w, as-is.

• Kali Hiwa, laterals 2019 — M: 23.15 · DHM: 14.00 · K: 26.38 · DHK: 16.15 · Y: 13.46 · DMY: 6.86 · Total KL: 11.10 · CMT: 462531 · Rank head: K > M > DHK
• Kali Hiwa, crown roots 2019 — M: 24.47 · DHM: 14.42 · K: 23.87 · DHK: 16.21 · Y: 14.32 · DMY: 6.69 · Total KL: 10.10 · CMT: 642531 · Rank head: M > K > DHK
• Kali Hiwa, peeled stump 2019 — M: 21.36 · DHM: 18.23 · K: 20.93 · DHK: 22.07 · Y: 12.40 · DMY: 5.01 · Total KL: 7.02 · CMT: 264531 · Rank head: DHK > M > K
• Kali Hiwa, lateral roots 09/25 — M: 20.21 · DHM: 13.07 · K: 24.34 · DHK: 15.41 · Y: 18.99 · DMY: 7.98 · Total KL: 10.70 · CMT: 463251 · Rank head: K > M > Y
• Kali Hiwa, root and chip powder 09/25 — M: 19.20 · DHM: 12.11 · K: 24.98 · DHK: 16.94 · Y: 17.78 · DMY: 8.99 · Total KL: 7.08 · CMT: 463251 · Rank head: K > M > Y
• Hawaiian HIWA-2-15, powder — M: 19.27 · DHM: 9.17 · K: 29.06 · DHK: 17.46 · Y: 14.98 · DMY: 10.06 · Total KL: 4.41 · CMT: 462315 · Rank head: K > M > DHK
• Hawaiian HIWA-1-27, powder — M: 19.87 · DHM: 9.08 · K: 30.99 · DHK: 18.65 · Y: 11.50 · DMY: 9.92 · Total KL: 3.68 · CMT: 462315 · Rank head: K > M > DHK
• Hawaiian HIWA-2-6, powder — M: 21.49 · DHM: 9.53 · K: 29.24 · DHK: 14.40 · Y: 11.93 · DMY: 13.41 · Total KL: 4.21 · CMT: 462135 · Rank head: K > M > DHK
• Hawaiian HIWA-1-6, powder — M: 19.48 · DHM: 10.79 · K: 29.11 · DHK: 18.32 · Y: 14.58 · DMY: 7.73 · Total KL: 4.46 · CMT: 462351 · Rank head: K > M > DHK
• Hawaiian HIWA-1-15, powder — M: 20.38 · DHM: 9.44 · K: 29.03 · DHK: 18.25 · Y: 13.27 · DMY: 9.63 · Total KL: 3.29 · CMT: 462315 · Rank head: K > M > DHK
• Hawaiian Hiwa-Roots-Frozen-Molokai — M: 19.49 · DHM: 10.31 · K: 26.74 · DHK: 17.88 · Y: 16.91 · DMY: 8.67 · Total KL: 6.21 · CMT: 462351 · Rank head: K > M > DHK
• Hawaiian Hiwa-435, roots — M: 19.24 · DHM: 10.18 · K: 26.59 · DHK: 18.36 · Y: 16.89 · DMY: 8.74 · Total KL: 6.80 · CMT: 462351 · Rank head: K > M > DHK

M methysticin, DHM dihydromethysticin, K kavain, DHK dihydrokavain, Y yangonin, DMY desmethoxyyangonin. The 2019 crown roots and peeled stump are plant-part comparisons, not root-class matches.

The third position is not conserved, and the paper does not ask it to be. Dihydrokavain holds it in eight of the ten samples and yangonin in the 2025 lot. Methysticin exceeds dihydrokavain by only 0.88 to 7.08 points, under 2.2 points in five samples, and dihydrokavain exceeds yangonin by as little as 0.97 points in root material. Those are margins narrower than ordinary variation, which is exactly why the third digit should not be made load-bearing.

Relative composition across the seven Hawaiian samples was tight for the leading analytes and loose for the trailing ones — kavain 28.68 ± 1.54 percent of total kavalactones and methysticin 19.89 ± 0.81, against desmethoxyyangonin at 9.73 ± 1.82. And the one external reference point for the cultivar agrees: the Lebot et al. (1999) Hāmākua lateral roots returned the same two-position head.

Like for like, and the yardsticks that size it

The 09/25 harvest was assayed as two separate materials, lateral roots and a root and chip powder. The powder is the closest material-class match in the record to the Hawaiian references, and it sits 1.20 percentage points root-mean-square from the mean of the two Hawaiian root samples across all six relative fractions — 1.17 from the Molokaʻi roots and 1.24 from Hiwa-435.

A separation of 1.20 points means nothing without knowing what ordinary variation looks like in the same units. Three yardsticks come from inside this dataset, and the gap between the lineages is smaller than all three. The five Hawaiian powders, milled and assayed together in a single laboratory job, sit 0.92 to 3.17 points from one another, averaging 1.84. Material class alone, measured inside the Hawaiian set, moves the profile 2.01 and 2.06 points. And Kali Hiwa's own harvest-to-harvest lateral variation is 2.77 points, driven mainly by yangonin moving from 13.46 to 18.99 relative percent between 2019 and 2025.

Kali Hiwa is therefore closer to its Hawaiian source stock than it is to itself six years earlier, and closer than the Hawaiian samples are to one another. The distance between the lineages is not distinguishable from the ordinary scatter of the material.

Both 09/25 certificates return chemotype 463251, as does the separately reported assay of the same harvest used for the beverage work — three separate laboratory jobs across nine months, agreeing on the code and on the two-position head.

Plant part reorders the code

Three plant parts of Kali Hiwa, taken from the same plant on the same day and assayed in the same job, did not return the same chemotype. Lateral roots reproduced the Hawaiian rank head exactly. Crown roots exchanged the first two positions. Peeled stump reordered all three. Any account of chemotype stability that does not address this is incomplete.

Table 2. Kavalactone composition of three plant parts of Kali Hiwa taken from the same plant and assayed in the same job, September 2019. Values are percent of that sample's total kavalactones except Total KL, which is percent w/w.

• Lateral roots — Total KL: 11.10 · M: 23.15 · DHM: 14.00 · K: 26.38 · DHK: 16.15 · Y: 13.46 · DMY: 6.86 · CMT: 462531 · Rank head: K > M > DHK
• Crown roots — Total KL: 10.10 · M: 24.47 · DHM: 14.42 · K: 23.87 · DHK: 16.21 · Y: 14.32 · DMY: 6.69 · CMT: 642531 · Rank head: M > K > DHK
• Peeled stump — Total KL: 7.02 · M: 21.36 · DHM: 18.23 · K: 20.93 · DHK: 22.07 · Y: 12.40 · DMY: 5.01 · CMT: 264531 · Rank head: DHK > M > K
• Net change, laterals to stump — Total KL: -4.08 · M: -1.79 · DHM: +4.23 · K: -5.45 · DHK: +5.92 · Y: -1.06 · DMY: -1.84 · CMT: — · Rank head: —

The reordering follows a consistent gradient, not a random rearrangement.

From lateral root to peeled stump, kavain fell 5.45 percentage points while dihydrokavain rose 5.92 and dihydromethysticin rose 4.23, and total kavalactone content fell from 11.10 to 7.02 percent w/w. This is the documented kava gradient, not drift. The practical consequence is that material class has to be matched before absolute compositions are compared — which is what makes the 09/25 root and chip powder the closest comparison in this study, and the lateral, crown and stump samples a demonstration of the gradient rather than of the cultivar.

Flavokavains place the line within noble character

The flavokavain ratios offer a test of identity that is structurally independent of kavalactone ranking, since no rearrangement of the kavalactone order can alter them. Lebot et al. (2019) characterised 174 varieties on two such ratios, separating noble kava from wichmannii and two-day types.

Table 3. Flavokavains and nobility ratios for every Kali Hiwa assay carrying them, against the group means of Lebot et al. (2019). Flavokavain values are percent w/w, as-is. Low FK/KL and high K/FKB both indicate noble character.

• Kali Hiwa, laterals 2019 — FK A: 0.217 · FK B: 0.142 · FK C: 0.050 · Total FK: 0.409 · FK/KL: 0.037 · K/FKB: 20.7
• Kali Hiwa, crown roots 2019 — FK A: 0.101 · FK B: 0.059 · FK C: 0.021 · Total FK: 0.181 · FK/KL: 0.018 · K/FKB: 40.7
• Kali Hiwa, peeled stump 2019 — FK A: 0.114 · FK B: 0.071 · FK C: 0.016 · Total FK: 0.201 · FK/KL: 0.029 · K/FKB: 20.7
• Kali Hiwa, lateral roots 09/25 — FK A: 0.142 · FK B: 0.077 · FK C: 0.015 · Total FK: 0.234 · FK/KL: 0.022 · K/FKB: 33.4
• Kali Hiwa, root and chip powder 09/25 — FK A: 0.065 · FK B: 0.054 · FK C: 0.014 · Total FK: 0.133 · FK/KL: 0.019 · K/FKB: 32.6
• Noble group mean, Lebot 2019 — FK A: — · FK B: — · FK C: — · Total FK: — · FK/KL: 0.130 · K/FKB: 7.31
• Wichmannii group mean, Lebot 2019 — FK A: — · FK B: — · FK C: — · Total FK: — · FK/KL: 0.210 · K/FKB: 1.70
• Two-day group mean, Lebot 2019 — FK A: — · FK B: — · FK C: — · Total FK: — · FK/KL: 0.360 · K/FKB: 1.50

The benchmark means derive from acetonic-extract HPTLC while the values here are HPLC on an as-is basis, so the comparison is directional rather than method-matched. The margins are wide enough to carry that.

Every Kali Hiwa assay sits far inside noble territory on both axes — flavokavain-to-kavalactone ratios of 0.018 to 0.037, three and a half to seven times below the noble mean, and kavain-to-flavokavain-B ratios of 20.7 to 40.7, nearly three to five and a half times above it. Flavokavains were not resolved in the 2018–2019 Hawaiian certificates, so this axis characterises the Kali line rather than comparing the two lineages.

What reaches the cup

A chemotype says what is in the root, not what reaches the cup. Kava is drunk as an aqueous suspension, and most of the kavalactone a drinker receives travels on suspended particulate rather than in true solution, so potency has to be measured in the beverage to mean anything.

Table 4. Beverage performance of Kali Hiwa under two standardized aqueous preparations. Extraction efficiency is the percentage of starting kavalactone mass delivered to the beverage.

• Short warm-water extraction, single cycle — Material: Lateral roots, 2019 lot · Runs: n = 1 · Extraction efficiency: 78.7% · KL mg per 125 mL: 334 · Particle-associated: — · FK mg per 125 mL: —
• Short warm-water extraction, single cycle — Material: Roots and chips, 2019 lot · Runs: n = 3 · Extraction efficiency: 80.7% ± 7.3% · KL mg per 125 mL: 188–225 · Particle-associated: 97.2% · FK mg per 125 mL: —
• Bottle-shaker, first extraction (900 mL) — Material: Dried root, 2025 lot · Runs: n = 1 · Extraction efficiency: 80.1% · KL mg per 125 mL: 90 · Particle-associated: 91.2% · FK mg per 125 mL: 0.9
• Bottle-shaker, first plus second (1800 mL) — Material: Dried root, 2025 lot · Runs: n = 1 · Extraction efficiency: 89.1% · KL mg per 125 mL: 50 · Particle-associated: 85.3% · FK mg per 125 mL: 0.5

The two studies used different doses, water volumes and plant parts; they establish the order of magnitude and the delivery route, not a single figure for the cultivar.

Delivery was particle-associated in both — 97.2 percent under short warm-water extraction and 91.2 percent under the bottle-shaker single extraction, the highest particulate share of the four varieties in that comparison.

On the flavokavain specification the 09/25 harvest is comfortably within beverage grade on both limbs. Its root and chip powder carried total flavokavains of 0.133 percent w/w against a 0.4 percent ceiling, and 1.9 percent of total kavalactone content against a 5 percent relative limit.

Three axes that fail on different things

HPTLC of Kali Hiwa raw root under the British Herbal Pharmacopoeia (1996) identification returned a profile consistent with the authenticated P. methysticum root reference materials run on the same plate. That establishes genus and species independently of both the chemotype and the flavokavains.

Taken together the three axes are mutually independent in what they could fail on. Botanical identification would fail on a different plant. The rank head would fail on a different cultivar of the same species. The flavokavain ratios would fail on a two-day or wichmannii accession. No single one of them is an identity test, and that is the point of running all three.

Limitations

1. The Hawaiian set is itself split between dried powder and fresh-frozen root, and those two differ from each other by about as much as either differs from the Kali material. The comparison rests on a small number of certificates per material class.
2. The tissue-cultured line is represented by two harvests six years apart, which removes the single-lot objection but not the site one. Both were grown in the same California greenhouse, and nothing here tests the Florida field material.
3. No assay is available from shortly after micropropagation, so the record cannot separate what tissue culture did from what six years of growing did.
4. The beverage figures come from two studies with different doses, water volumes and plant parts, and neither carries preparation-level replication beyond the triplicate short warm-water runs.
5. An earlier analysis explored a standardised-distance criterion calibrated on the five October 2018 powders. It is reported only as a negative result: that criterion places a genuine Hawaiian Hiwa outside its own envelope, so it is not offered as an authentication test.
6. Most importantly for how this result should be used — the study is a continuity analysis within a single cultivar. It asks whether Kali Hiwa still resembles the stock it came from, and answers that it does. It does not ask, and cannot answer, whether the features reported here would distinguish Hiwa from another cultivar.

Conclusions

The Kali Hiwa line carries the kavalactone rank signature of the Hawaiian Hiwa stock it was developed from. Kavain leads and methysticin follows in every root and powder sample of both lineages, on margins wide enough to survive the reordering that micropropagation, phytosanitary treatment, a new climate and six years of growing might have introduced. Where the code does move, it moves at positions where the underlying margins are narrower than ordinary analytical variation, and it moves in the direction the published plant-part gradient predicts.

The line is also of practical account. Under two standardized aqueous preparations it delivered between 90 and 334 mg of kavalactone per 125 mL serving at close to four-fifths recovery in a single cycle, and its 2025 lot met the flavokavain specification for beverage-grade material on both limbs.

What the record does not yet support is a numerical identity test. The sample count per material class is too small to calibrate a tolerance that would generalise, and the study establishes continuity within the cultivar rather than separation from others. Closing that gap needs a whole-root Kali certificate matched to a Hawaiian lateral, assays from the Florida field material, and a set of non-Hiwa cultivars run under the same method.

Acknowledgements

All analyses reported here were performed by Flora Research Laboratories, LLC, Grants Pass, Oregon. The author thanks James Neal-Kababick, Laboratory Director, and analysts Reine Bravo, Pierce Prozy and Stacy Wise, whose certificates over seven years constitute the entirety of the evidence base.